Fix handling of embeddings_key in api_search() function

Fix application + script requirements.txt

.env-template

@@ -1,7 +1,8 @@
-OPENAI_API_KEY=<LLM api key (for example, open ai key)>
-EMBEDDINGS_KEY=<LLM embeddings api key (for example, open ai key)>
+API_KEY=<LLM api key (for example, open ai key)>
+LLM_NAME=docsgpt
+VITE_API_STREAMING=true
 
-#For Azure
+#For Azure (you can delete it if you don't use Azure)
 OPENAI_API_BASE=
 OPENAI_API_VERSION=
 AZURE_DEPLOYMENT_NAME=

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -0,0 +1,138 @@
+name: "🐛 Bug Report"
+description: "Submit a bug report to help us improve"
+title: "🐛 Bug Report: "
+labels: ["type: bug"]
+body:
+  - type: markdown
+    attributes:
+      value: We value your time and your efforts to submit this bug report is appreciated. 🙏
+
+  - type: textarea
+    id: description
+    validations:
+      required: true
+    attributes:
+      label: "📜 Description"
+      description: "A clear and concise description of what the bug is."
+      placeholder: "It bugs out when ..."
+
+  - type: textarea
+    id: steps-to-reproduce
+    validations:
+      required: true
+    attributes:
+      label: "👟 Reproduction steps"
+      description: "How do you trigger this bug? Please walk us through it step by step."
+      placeholder: "1. Go to '...'
+        2. Click on '....'
+        3. Scroll down to '....'
+        4. See error"
+
+  - type: textarea
+    id: expected-behavior
+    validations:
+      required: true
+    attributes:
+      label: "👍 Expected behavior"
+      description: "What did you think should happen?"
+      placeholder: "It should ..."
+
+  - type: textarea
+    id: actual-behavior
+    validations:
+      required: true
+    attributes:
+      label: "👎 Actual Behavior with Screenshots"
+      description: "What did actually happen? Add screenshots, if applicable."
+      placeholder: "It actually ..."
+
+  - type: dropdown
+    id: operating-system
+    attributes:
+      label: "💻 Operating system"
+      description: "What OS is your app running on?"
+      options:
+        - Linux
+        - MacOS
+        - Windows
+        - Something else
+    validations:
+      required: true
+
+  - type: dropdown
+    id: browsers
+    attributes:
+      label: What browsers are you seeing the problem on?
+      multiple: true
+      options:
+        - Firefox
+        - Chrome
+        - Safari
+        - Microsoft Edge
+        - Something else
+
+  - type: dropdown
+    id: dev-environment
+    validations:
+      required: true
+    attributes:
+      label: "🤖 What development environment are you experiencing this bug on?"
+      options:
+        - Docker
+        - Local dev server
+
+  - type: textarea
+    id: env-vars
+    validations:
+      required: false
+    attributes:
+      label: "🔒 Did you set the correct environment variables in the right path? List the environment variable names (not values please!)"
+      description: "Please refer to the [Project setup instructions](https://github.com/arc53/DocsGPT#quickstart) if you are unsure."
+      placeholder: "It actually ..."
+
+  - type: textarea
+    id: additional-context
+    validations:
+      required: false
+    attributes:
+      label: "📃 Provide any additional context for the Bug."
+      description: "Add any other context about the problem here."
+      placeholder: "It actually ..."
+
+  - type: textarea
+    id: logs
+    validations:
+      required: false
+    attributes:
+      label: 📖 Relevant log output
+      description: Please copy and paste any relevant log output. This will be automatically formatted into code, so no need for backticks.
+      render: shell
+
+  - type: checkboxes
+    id: no-duplicate-issues
+    attributes:
+      label: "👀 Have you spent some time to check if this bug has been raised before?"
+      options:
+        - label: "I checked and didn't find similar issue"
+          required: true
+
+  - type: dropdown
+    id: willing-to-submit-pr
+    attributes:
+      label: 🔗 Are you willing to submit PR?
+      description: This is absolutely not required, but we are happy to guide you in the contribution process.
+      options: # Added options key
+        - "Yes, I am willing to submit a PR!"
+        - "No"
+    validations:
+      required: false
+
+
+  - type: checkboxes
+    id: terms
+    attributes:
+      label: 🧑‍⚖️ Code of Conduct
+      description: By submitting this issue, you agree to follow our [Code of Conduct](https://github.com/arc53/DocsGPT/blob/main/CODE_OF_CONDUCT.md)
+      options:
+        - label: I agree to follow this project's Code of Conduct
+          required: true

.github/ISSUE_TEMPLATE/feature_request.yml

@@ -0,0 +1,54 @@
+name: 🚀 Feature
+description: "Submit a proposal for a new feature"
+title: "🚀 Feature: "
+labels: [feature]
+body:
+  - type: markdown
+    attributes:
+      value: We value your time and your efforts to submit this bug report is appreciated. 🙏
+  - type: textarea
+    id: feature-description
+    validations:
+      required: true
+    attributes:
+      label: "🔖 Feature description"
+      description: "A clear and concise description of what the feature is."
+      placeholder: "You should add ..."
+  - type: textarea
+    id: pitch
+    validations:
+      required: true
+    attributes:
+      label: "🎤 Why is this feature needed ?"
+      description: "Please explain why this feature should be implemented and how it would be used. Add examples, if applicable."
+      placeholder: "In my use-case, ..."
+  - type: textarea
+    id: solution
+    validations:
+      required: true
+    attributes:
+      label: "✌️ How do you aim to achieve this?"
+      description: "A clear and concise description of what you want to happen."
+      placeholder: "I want this feature to, ..."
+  - type: textarea
+    id: alternative
+    validations:
+      required: false
+    attributes:
+      label: "🔄️ Additional Information"
+      description: "A clear and concise description of any alternative solutions or additional solutions you've considered."
+      placeholder: "I tried, ..."
+  - type: checkboxes
+    id: no-duplicate-issues
+    attributes:
+      label: "👀 Have you spent some time to check if this feature request has been raised before?"
+      options:
+        - label: "I checked and didn't find similar issue"
+          required: true
+  - type: dropdown
+    id: willing-to-submit-pr
+    attributes:
+      label: Are you willing to submit PR?
+      description: This is absolutely not required, but we are happy to guide you in the contribution process.
+      options:
+        - "Yes I am willing to submit a PR!"

.github/PULL_REQUEST_TEMPLATE.md

@@ -0,0 +1,5 @@
+- **What kind of change does this PR introduce?** (Bug fix, feature, docs update, ...)
+
+- **Why was this change needed?** (You can also link to an open issue here)
+
+- **Other information**:

.github/holopin.yml

@@ -0,0 +1,5 @@
+organization: arc53
+defaultSticker: clqmdf0ed34290glbvqh0kzxd
+stickers:
+  - id: clqmdf0ed34290glbvqh0kzxd
+    alias: festive

.github/labeler.yml

@@ -0,0 +1,23 @@
+repo:
+  - '*'
+
+github:
+  - .github/**/*
+
+application:
+  - application/**/*
+
+docs:
+  - docs/**/*
+
+extensions:
+  - extensions/**/*
+
+frontend:
+  - frontend/**/*
+
+scripts:
+  - scripts/**/*
+
+tests:
+  - tests/**/*

.github/workflows/labeler.yml

@@ -0,0 +1,15 @@
+# https://github.com/actions/labeler
+name: Pull Request Labeler
+on:
+  - pull_request_target
+jobs:
+  triage:
+    permissions:
+      contents: read
+      pull-requests: write
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/labeler@v4
+        with:
+          repo-token: "${{ secrets.GITHUB_TOKEN }}"
+          sync-labels: true

.github/workflows/pytest.yml

@@ -1,15 +1,12 @@
 name: Run python tests with pytest
-
 on: [push, pull_request]
-
 jobs:
-  build:
-
+  pytest_and_coverage:
+    name: Run tests and count coverage
     runs-on: ubuntu-latest
     strategy:
       matrix:
         python-version: ["3.9", "3.10", "3.11"]
-
     steps:
       - uses: actions/checkout@v3
       - name: Set up Python ${{ matrix.python-version }}
@@ -19,9 +16,15 @@ jobs:
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip
-          pip install pytest
+          pip install pytest pytest-cov
           cd application
           if [ -f requirements.txt ]; then pip install -r requirements.txt; fi
-      - name: Test with pytest
+      - name: Test with pytest and generate coverage report
         run: |
-          python -m pytest
+          python -m pytest --cov=application --cov=scripts --cov=extensions --cov-report=xml
+      - name: Upload coverage reports to Codecov
+        if: github.event_name == 'pull_request' && matrix.python-version == '3.11'
+        uses: codecov/codecov-action@v3
+        env:
+          CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
+

.gitignore

@@ -5,7 +5,7 @@ __pycache__/
 
 # C extensions
 *.so
-
+*.next
 # Distribution / packaging
 .Python
 build/
@@ -169,4 +169,6 @@ application/vectors/
 
 **/yarn.lock
 
-node_modules/
+node_modules/
+.vscode/settings.json
+models/

CODE_OF_CONDUCT.md

@@ -2,58 +2,58 @@
 
 ## Our Pledge
 
-We as members, contributors, and leaders pledge to make participation in our
-community a harassment-free experience for everyone, regardless of age, body
+We as members, contributors and leaders pledge to make participation in our
+community, a harassment-free experience for everyone, regardless of age, body
 size, visible or invisible disability, ethnicity, sex characteristics, gender
 identity and expression, level of experience, education, socio-economic status,
-nationality, personal appearance, race, religion, or sexual identity
+nationality, personal appearance, race, religion or sexual identity
 and orientation.
 
 We pledge to act and interact in ways that contribute to an open, welcoming,
-diverse, inclusive, and healthy community.
+diverse, inclusive and a healthy community.
 
 ## Our Standards
 
-Examples of behavior that contributes to a positive environment for our
+Examples of behavior that contribute to a positive environment for our
 community include:
 
-* Demonstrating empathy and kindness toward other people
-* Being respectful of differing opinions, viewpoints, and experiences
-* Giving and gracefully accepting constructive feedback
-* Accepting responsibility and apologizing to those affected by our mistakes,
-  and learning from the experience
-* Focusing on what is best not just for us as individuals, but for the
-  overall community
+## Demonstrating empathy and kindness towards other people
+1. Being respectful and open to differing opinions, viewpoints, and experiences
+2. Giving and gracefully accepting constructive feedback
+3. Taking accountability and offering apologies to those who have been impacted by our errors,
+  while also gaining insights from the situation
+4. Focusing on what is best not just for us as individuals but for the
+  community as a whole
 
 Examples of unacceptable behavior include:
 
-* The use of sexualized language or imagery, and sexual attention or
+1. The use of sexualized language or imagery, and sexual attention or
   advances of any kind
-* Trolling, insulting or derogatory comments, and personal or political attacks
-* Public or private harassment
-* Publishing others' private information, such as a physical or email
+2. Trolling, insulting or derogatory comments, and personal or political attacks
+3. Public or private harassment
+4. Publishing other's private information, such as a physical or email
   address, without their explicit permission
-* Other conduct which could reasonably be considered inappropriate in a
+5. Other conduct which could reasonably be considered inappropriate in a
   professional setting
 
 ## Enforcement Responsibilities
 
 Community leaders are responsible for clarifying and enforcing our standards of
 acceptable behavior and will take appropriate and fair corrective action in
-response to any behavior that they deem inappropriate, threatening, offensive,
+response to any behavior that they deem inappropriate, threatening, offensive
 or harmful.
 
 Community leaders have the right and responsibility to remove, edit, or reject
 comments, commits, code, wiki edits, issues, and other contributions that are
-not aligned to this Code of Conduct, and will communicate reasons for moderation
+not aligned to this Code of Conduct and will communicate reasons for moderation
 decisions when appropriate.
 
 ## Scope
 
-This Code of Conduct applies within all community spaces, and also applies when
+This Code of Conduct applies within all community spaces and also applies when
 an individual is officially representing the community in public spaces.
 Examples of representing our community include using an official e-mail address,
-posting via an official social media account, or acting as an appointed
+posting via an official social media account or acting as an appointed
 representative at an online or offline event.
 
 ## Enforcement
@@ -63,29 +63,27 @@ reported to the community leaders responsible for enforcement at
 contact@arc53.com.
 All complaints will be reviewed and investigated promptly and fairly.
 
-All community leaders are obligated to respect the privacy and security of the
+All community leaders are obligated to be respectful towards the privacy and security of the
 reporter of any incident.
 
 ## Enforcement Guidelines
 
 Community leaders will follow these Community Impact Guidelines in determining
-the consequences for any action they deem in violation of this Code of Conduct:
+the consequences for any action that they deem in violation of this Code of Conduct:
 
 ### 1. Correction
+* **Community Impact**: Use of inappropriate language or other behavior deemed
+unprofessional or unwelcome in the community space.
 
-**Community Impact**: Use of inappropriate language or other behavior deemed
-unprofessional or unwelcome in the community.
-
-**Consequence**: A private, written warning from community leaders, providing
+* **Consequence**: A private, written warning from community leaders, providing
 clarity around the nature of the violation and an explanation of why the
 behavior was inappropriate. A public apology may be requested.
 
 ### 2. Warning
-
-**Community Impact**: A violation through a single incident or series
+* **Community Impact**: A violation through a single incident or series
 of actions.
 
-**Consequence**: A warning with consequences for continued behavior. No
+* **Consequence**: A warning with consequences for continued behavior. No
 interaction with the people involved, including unsolicited interaction with
 those enforcing the Code of Conduct, for a specified period of time. This
 includes avoiding interactions in community spaces as well as external channels
@@ -93,23 +91,21 @@ like social media. Violating these terms may lead to a temporary or
 permanent ban.
 
 ### 3. Temporary Ban
-
-**Community Impact**: A serious violation of community standards, including
+* **Community Impact**: A serious violation of community standards, including
 sustained inappropriate behavior.
 
-**Consequence**: A temporary ban from any sort of interaction or public
+* **Consequence**: A temporary ban from any sort of interaction or public
 communication with the community for a specified period of time. No public or
 private interaction with the people involved, including unsolicited interaction
 with those enforcing the Code of Conduct, is allowed during this period.
 Violating these terms may lead to a permanent ban.
 
 ### 4. Permanent Ban
+* **Community Impact**: Demonstrating a pattern of violation of community
+standards, including sustained inappropriate behavior,harassment of an
+individual or aggression towards or disparagement of classes of individuals.
 
-**Community Impact**: Demonstrating a pattern of violation of community
-standards, including sustained inappropriate behavior,  harassment of an
-individual, or aggression toward or disparagement of classes of individuals.
-
-**Consequence**: A permanent ban from any sort of public interaction within
+* **Consequence**: A permanent ban from any sort of public interaction within
 the community.
 
 ## Attribution

CONTRIBUTING.md

@@ -1,38 +1,128 @@
-# Welcome to DocsGPT Contributing guideline
+# Welcome to DocsGPT Contributing Guidelines
 
-Thank you for choosing this project to contribute to, we are all very grateful!
+Thank you for choosing to contribute to DocsGPT! We are all very grateful! 
 
 # We accept different types of contributions
 
-📣 Discussions - where you can start a new topic or answer some questions
+📣 **Discussions** - Engage in conversations, start new topics, or help answer questions.
 
-🐞 Issues - Is how we track tasks, sometimes its bugs that need fixing, sometimes its new features
+🐞 **Issues** - This is where we keep track of tasks. It could be bugs,fixes or suggestions for new features.
 
-🛠️ Pull requests - Is how you can suggest changes to our repository, to work on existing issue or to add new features
+🛠️ **Pull requests** - Suggest changes to our repository, either by working on existing issues or adding new features.
 
-📚 Wiki - where we have our documentation
+📚 **Wiki** - This is where our documentation resides.
 
 
 ## 🐞 Issues and Pull requests
 
-We value contributions to our issues in form of discussion or suggestion, we recommend that you check out existing issues and our [Roadmap](https://github.com/orgs/arc53/projects/2)
-
-If you want to contribute by writing code there are few things that you should know before doing it:
-We have frontend (React, Vite) and Backend (python)
-
-### If you are looking to contribute to Frontend (⚛️React, Vite):
-Current frontend is being migrated from /application to /frontend with a new design, so please contribute to the new on. Check out this [Milestone](https://github.com/arc53/DocsGPT/milestone/1) and its issues also [Figma](https://www.figma.com/file/OXLtrl1EAy885to6S69554/DocsGPT?node-id=0%3A1&t=hjWVuxRg9yi5YkJ9-1)
-Please try to follow guidelines
+- We value contributions in the form of discussions or suggestions. We recommend taking a look at existing issues and our [roadmap](https://github.com/orgs/arc53/projects/2).
 
 
-### If you are looking to contribute to Backend (🐍Python):
-Check out our issues, and contribute to /application or /scripts (ignore old  ingest_rst.py ingest_rst_sphinx.py files, they will be deprecated soon)
-Currently we don't have any tests(which would be useful😉) but before submitting you PR make sure that after you ingested some test data its queryable
+- If you're interested in contributing code, here are some important things to know:
 
-### Workflow:
-Create a fork, make changes on your forked repository, submit changes in a form of pull request
+- We have a frontend built on React (Vite) and a backend in Python.
+=======
+Before creating issues, please check out how the latest version of our app looks and works by launching it via [Quickstart](https://github.com/arc53/DocsGPT#quickstart) the version on our live demo is slightly modified with login. Your issues should relate to the version that you can launch via [Quickstart](https://github.com/arc53/DocsGPT#quickstart).
 
-## Questions / collaboration
-Please join our [Discord](https://discord.gg/n5BX8dh8rU) don't hesitate, we are very friendly and welcoming to new contributors.
+### 👨‍💻 If you're interested in contributing code, here are some important things to know:
 
-# Thank you so much for considering to contribute to DocsGPT!🙏
+
+Tech Stack Overview:
+
+- 🌐 Frontend: Built with React (Vite) ⚛️,
+
+- 🖥 Backend: Developed in Python 🐍
+
+### 🌐 If you are looking to contribute to frontend (⚛️React, Vite):
+
+- The current frontend is being migrated from [`/application`](https://github.com/arc53/DocsGPT/tree/main/application) to [`/frontend`](https://github.com/arc53/DocsGPT/tree/main/frontend) with a new design, so please contribute to the new one.
+- Check out this [milestone](https://github.com/arc53/DocsGPT/milestone/1) and its issues.
+- The updated Figma design can be found [here](https://www.figma.com/file/OXLtrl1EAy885to6S69554/DocsGPT?node-id=0%3A1&t=hjWVuxRg9yi5YkJ9-1).
+
+Please try to follow the guidelines.
+
+### 🖥 If you are looking to contribute to Backend (🐍 Python):
+
+- Review our issues and contribute to [`/application`](https://github.com/arc53/DocsGPT/tree/main/application) or [`/scripts`](https://github.com/arc53/DocsGPT/tree/main/scripts) (please disregard old [`ingest_rst.py`](https://github.com/arc53/DocsGPT/blob/main/scripts/old/ingest_rst.py) [`ingest_rst_sphinx.py`](https://github.com/arc53/DocsGPT/blob/main/scripts/old/ingest_rst_sphinx.py) files; they will be deprecated soon).
+- All new code should be covered with unit tests ([pytest](https://github.com/pytest-dev/pytest)). Please find tests under [`/tests`](https://github.com/arc53/DocsGPT/tree/main/tests) folder.
+- Before submitting your Pull Request, ensure it can be queried after ingesting some test data.
+  
+### Testing
+
+To run unit tests from the root of the repository, execute:
+```
+python -m pytest
+```
+
+## Workflow 📈
+
+Here's a step-by-step guide on how to contribute to DocsGPT:
+
+1. **Fork the Repository:**
+   - Click the "Fork" button at the top-right of this repository to create your fork.
+
+2. **Clone the Forked Repository:**
+   - Clone the repository using:
+      ``` shell
+      git clone https://github.com/<your-github-username>/DocsGPT.git
+      ```
+
+3. **Keep your Fork in Sync:**
+   - Before you make any changes, make sure that your fork is in sync to avoid merge conflicts using:
+     ```shell
+     git remote add upstream https://github.com/arc53/DocsGPT.git
+     git pull upstream main
+     ```
+
+4. **Create and Switch to a New Branch:**
+   - Create a new branch for your contribution using:
+     ```shell
+     git checkout -b your-branch-name
+     ```
+
+5. **Make Changes:**
+   - Make the required changes in your branch.
+
+6. **Add Changes to the Staging Area:**
+   - Add your changes to the staging area using:
+     ```shell
+     git add .
+     ```
+
+7. **Commit Your Changes:**
+   - Commit your changes with a descriptive commit message using:
+     ```shell
+     git commit -m "Your descriptive commit message"
+     ```
+
+8. **Push Your Changes to the Remote Repository:**
+   - Push your branch with changes to your fork on GitHub using:
+     ```shell
+     git push origin your-branch-name
+     ```
+
+9. **Submit a Pull Request (PR):**
+   - Create a Pull Request from your branch to the main repository. Make sure to include a detailed description of your changes and reference any related issues.
+
+10. **Collaborate:**
+   - Be responsive to comments and feedback on your PR.
+   - Make necessary updates as suggested.
+   - Once your PR is approved, it will be merged into the main repository.
+
+11. **Testing:**
+   - Before submitting a Pull Request, ensure your code passes all unit tests.
+   - To run unit tests from the root of the repository, execute:
+     ```shell
+     python -m pytest
+     ```
+
+*Note: You should run the unit test only after making the changes to the backend code.*
+
+12. **Questions and Collaboration:**
+    - Feel free to join our Discord. We're very friendly and welcoming to new contributors, so don't hesitate to reach out.
+
+Thank you for considering contributing to DocsGPT! 🙏
+
+## Questions/collaboration
+Feel free to join our [Discord](https://discord.gg/n5BX8dh8rU). We're very friendly and welcoming to new contributors, so don't hesitate to reach out.
+# Thank you so much for considering to contribute DocsGPT!🙏

README.md

@@ -7,130 +7,193 @@
 </p>
 
 <p align="left">
-  <strong>DocsGPT</strong> is a cutting-edge open-source solution that streamlines the process of finding information in project documentation. With its integration of the powerful <strong>GPT</strong> models, developers can easily ask questions about a project and receive accurate answers.
+  <strong><a href="https://docsgpt.arc53.com/">DocsGPT</a></strong> is a cutting-edge open-source solution that streamlines the process of finding information in the project documentation. With its integration of the powerful <strong>GPT</strong> models, developers can easily ask questions about a project and receive accurate answers.
   
-Say goodbye to time-consuming manual searches, and let <strong>DocsGPT</strong> help you quickly find the information you need. Try it out and see how it revolutionizes your project documentation experience. Contribute to its development and be a part of the future of AI-powered assistance.
+Say goodbye to time-consuming manual searches, and let <strong><a href="https://docsgpt.arc53.com/">DocsGPT</a></strong> help you quickly find the information you need. Try it out and see how it revolutionizes your project documentation experience. Contribute to its development and be a part of the future of AI-powered assistance.
 </p>
 
 <div align="center">
   
-  <a href="https://discord.gg/n5BX8dh8rU">![example1](https://img.shields.io/github/stars/arc53/docsgpt?style=social)</a>
-  <a href="https://discord.gg/n5BX8dh8rU">![example2](https://img.shields.io/github/forks/arc53/docsgpt?style=social)</a>
-  <a href="https://discord.gg/n5BX8dh8rU">![example3](https://img.shields.io/github/license/arc53/docsgpt)</a>
-  <a href="https://discord.gg/n5BX8dh8rU">![example3](https://img.shields.io/discord/1070046503302877216)</a>
-  
+  <a href="https://github.com/arc53/DocsGPT">![link to main GitHub showing Stars number](https://img.shields.io/github/stars/arc53/docsgpt?style=social)</a>
+  <a href="https://github.com/arc53/DocsGPT">![link to main GitHub showing Forks number](https://img.shields.io/github/forks/arc53/docsgpt?style=social)</a>
+  <a href="https://github.com/arc53/DocsGPT/blob/main/LICENSE">![link to license file](https://img.shields.io/github/license/arc53/docsgpt)</a>
+  <a href="https://discord.gg/n5BX8dh8rU">![link to discord](https://img.shields.io/discord/1070046503302877216)</a>
+  <a href="https://twitter.com/docsgptai">![X (formerly Twitter) URL](https://img.shields.io/twitter/follow/docsgptai)</a>
+
+ 
 </div>
 
+### Production Support / Help for Companies:
+
+We're eager to provide personalized assistance when deploying your DocsGPT to a live environment.
+
+- [Book Demo :wave:](https://airtable.com/appdeaL0F1qV8Bl2C/shrrJF1Ll7btCJRbP)
+- [Send Email :email:](mailto:contact@arc53.com?subject=DocsGPT%20support%2Fsolutions)
+
 ![video-example-of-docs-gpt](https://d3dg1063dc54p9.cloudfront.net/videos/demov3.gif)
 
 ## Roadmap
 
-You can find our [Roadmap](https://github.com/orgs/arc53/projects/2) here, please don't hesitate contributing or creating issues, it helps us make DocsGPT better!
+You can find our roadmap [here](https://github.com/orgs/arc53/projects/2). Please don't hesitate to contribute or create issues, it helps us improve DocsGPT!
 
-## Our open source models optimised for DocsGPT:
+## Our Open-Source Models Optimized for DocsGPT:
 
-| Name              | Base Model | Requirements (or similar)                        |
-|-------------------|------------|----------------------------------------------------------|
-| [Docsgpt-7b-falcon](https://huggingface.co/Arc53/docsgpt-7b-falcon)  | Falcon-7b  |  1xA10G gpu   |
-| [Docsgpt-14b](https://huggingface.co/Arc53/docsgpt-14b)              | llama-2-14b    | 2xA10 gpu's   |
-| [Docsgpt-40b](https://huggingface.co/Arc53/docsgpt-40b-falcon)       | falcon-40b     | 8xA10G gpu's  |
-
-
-If you don't have enough resources to run it you can use bitsnbytes to quantize
+| Name                                                                  | Base Model  | Requirements (or similar) |
+| --------------------------------------------------------------------- | ----------- | ------------------------- |
+| [Docsgpt-7b-falcon](https://huggingface.co/Arc53/docsgpt-7b-falcon)   | Falcon-7b   | 1xA10G gpu                |
+| [Docsgpt-14b](https://huggingface.co/Arc53/docsgpt-14b)               | llama-2-14b | 2xA10 gpu's               |
+| [Docsgpt-40b-falcon](https://huggingface.co/Arc53/docsgpt-40b-falcon) | falcon-40b  | 8xA10G gpu's              |
 
+If you don't have enough resources to run it, you can use bitsnbytes to quantize.
 
 ## Features
 
-![Group 9](https://user-images.githubusercontent.com/17906039/220427472-2644cff4-7666-46a5-819f-fc4a521f63c7.png)
+![Main features of DocsGPT showcasing six main features](https://user-images.githubusercontent.com/17906039/220427472-2644cff4-7666-46a5-819f-fc4a521f63c7.png)
 
+## Useful Links
 
-## Useful links
- [Live preview](https://docsgpt.arc53.com/)
- 
- [Join Our Discord](https://discord.gg/n5BX8dh8rU)
- 
- [Guides](https://github.com/arc53/docsgpt/wiki)
+- :mag: :fire: [Live preview](https://docsgpt.arc53.com/)
 
- [Interested in contributing?](https://github.com/arc53/DocsGPT/blob/main/CONTRIBUTING.md)
+- :speech_balloon: :tada: [Join our Discord](https://discord.gg/n5BX8dh8rU)
 
- [How to use any other documentation](https://github.com/arc53/docsgpt/wiki/How-to-train-on-other-documentation)
+- :books: :sunglasses: [Guides](https://docs.docsgpt.co.uk/)
 
- [How to host it locally (so all data will stay on-premises)](https://github.com/arc53/DocsGPT/wiki/How-to-use-different-LLM's#hosting-everything-locally)
+- :couple: [Interested in contributing?](https://github.com/arc53/DocsGPT/blob/main/CONTRIBUTING.md)
 
+- :file_folder: :rocket: [How to use any other documentation](https://docs.docsgpt.co.uk/Guides/How-to-train-on-other-documentation)
 
-## Project structure
-- Application - Flask app (main application)
+- :house: :closed_lock_with_key: [How to host it locally (so all data will stay on-premises)](https://docs.docsgpt.co.uk/Guides/How-to-use-different-LLM)
 
-- Extensions - Chrome extension
+## Project Structure
 
-- Scripts - Script that creates similarity search index and store for other libraries. 
+- Application - Flask app (main application).
 
-- Frontend - Frontend uses Vite and React
+- Extensions - Chrome extension.
+
+- Scripts - Script that creates similarity search index for other libraries.
+
+- Frontend - Frontend uses <a href="https://vitejs.dev/">Vite</a> and <a href="https://react.dev/">React</a>.
 
 ## QuickStart
 
-Note: Make sure you have docker installed
+> [!Note]
+> Make sure you have [Docker](https://docs.docker.com/engine/install/) installed
 
-1. Dowload and open this repository with `git clone https://github.com/arc53/DocsGPT.git`
-2. Create an .env file in your root directory and set the env variable OPENAI_API_KEY with your openai api key and  VITE_API_STREAMING to true or false, depending on if you want streaming answers or not
+On Mac OS or Linux, write:
+
+`./setup.sh`
+
+It will install all the dependencies and allow you to download the local model, use OpenAI or use our LLM API.
+
+Otherwise, refer to this Guide:
+
+1. Download and open this repository with `git clone https://github.com/arc53/DocsGPT.git`
+2. Create a `.env` file in your root directory and set the env variables and `VITE_API_STREAMING` to true or false, depending on whether you want streaming answers or not.
    It should look like this inside:
-   
+
    ```
-   OPENAI_API_KEY=Yourkey
+   LLM_NAME=[docsgpt or openai or others] 
    VITE_API_STREAMING=true
+   API_KEY=[if LLM_NAME is openai]
    ```
-3. Run `./run-with-docker-compose.sh`
-4. Navigate to http://localhost:5173/
 
-To stop just run Ctrl + C
+   See optional environment variables in the [/.env-template](https://github.com/arc53/DocsGPT/blob/main/.env-template) and [/application/.env_sample](https://github.com/arc53/DocsGPT/blob/main/application/.env_sample) files.
 
-## Development environments
+3. Run [./run-with-docker-compose.sh](https://github.com/arc53/DocsGPT/blob/main/run-with-docker-compose.sh).
+4. Navigate to http://localhost:5173/.
 
-### Spin up mongo and redis
-For development only 2 containers are used from docker-compose.yaml (by deleting all services except for redis and mongo). 
+To stop, just run `Ctrl + C`.
+
+## Development Environments
+
+### Spin up Mongo and Redis
+
+For development, only two containers are used from [docker-compose.yaml](https://github.com/arc53/DocsGPT/blob/main/docker-compose.yaml) (by deleting all services except for Redis and Mongo).
 See file [docker-compose-dev.yaml](./docker-compose-dev.yaml).
 
 Run
+
 ```
 docker compose -f docker-compose-dev.yaml build
 docker compose -f docker-compose-dev.yaml up -d
 ```
 
-### Run the backend
+### Run the Backend
 
-Make sure you have Python 3.10 or 3.11 installed.
+> [!Note]
+> Make sure you have Python 3.10 or 3.11 installed.
+
+1. Export required environment variables or prepare a `.env` file in the `/application` folder:
+   - Copy [.env_sample](https://github.com/arc53/DocsGPT/blob/main/application/.env_sample) and create `.env`.
+
+(check out [`application/core/settings.py`](application/core/settings.py) if you want to see more config options.)
+
+2. (optional) Create a Python virtual environment:
+   You can follow the [Python official documentation](https://docs.python.org/3/tutorial/venv.html) for virtual environments.
+
+a) On Mac OS and Linux
 
-1. Export required environment variables
-```commandline
-export CELERY_BROKER_URL=redis://localhost:6379/0   
-export CELERY_RESULT_BACKEND=redis://localhost:6379/1
-export MONGO_URI=mongodb://localhost:27017/docsgpt
-```
-2. Prepare .env file
-Copy `.env_sample` and create `.env` with your OpenAI API token
-3. (optional) Create a python virtual environment
 ```commandline
 python -m venv venv
 . venv/bin/activate
 ```
-4. Change to `application/` subdir and install dependencies for the backend
+
+b) On Windows
+
+```commandline
+python -m venv venv
+ venv/Scripts/activate
+```
+
+3. Download embedding model and save it in the `model/` folder:
+You can use the script below, or download it manually from [here](https://d3dg1063dc54p9.cloudfront.net/models/embeddings/mpnet-base-v2.zip), unzip it and save it in the `model/` folder.
+
+```commandline
+wget https://d3dg1063dc54p9.cloudfront.net/models/embeddings/mpnet-base-v2.zip
+unzip mpnet-base-v2.zip -d model
+rm mpnet-base-v2.zip
+
+4. Change to the `application/` subdir by the command `cd application/` and install dependencies for the backend:
+
 ```commandline
-cd application/ 
 pip install -r requirements.txt
 ```
-5. Run the app `python wsgi.py`
-6. Start worker with `celery -A app.celery worker -l INFO`
 
-### Start frontend 
-Make sure you have Node version 16 or higher.
+5. Run the app using `flask --app application/app.py run --host=0.0.0.0 --port=7091`.
+6. Start worker with `celery -A application.app.celery worker -l INFO`.
 
-1. Navigate to `/frontend` folder
-2. Install dependencies
-`npm install`
-3. Run the app 
-`npm run dev`
+### Start Frontend
 
+> [!Note]
+> Make sure you have Node version 16 or higher.
 
+1. Navigate to the [/frontend](https://github.com/arc53/DocsGPT/tree/main/frontend) folder.
+2. Install the required packages `husky` and `vite` (ignore if already installed).
 
-Built with [🦜️🔗 LangChain](https://github.com/hwchase17/langchain)
+```commandline
+npm install husky -g
+npm install vite -g
+```
 
+3. Install dependencies by running `npm install --include=dev`.
+4. Run the app using `npm run dev`.
+
+## Contributing
+
+Please refer to the [CONTRIBUTING.md](CONTRIBUTING.md) file for information about how to get involved. We welcome issues, questions, and pull requests.
+
+## Code Of Conduct
+
+We as members, contributors, and leaders, pledge to make participation in our community a harassment-free experience for everyone, regardless of age, body size, visible or invisible disability, ethnicity, sex characteristics, gender identity and expression, level of experience, education, socio-economic status, nationality, personal appearance, race, religion, or sexual identity and orientation. Please refer to the [CODE_OF_CONDUCT.md](CODE_OF_CONDUCT.md) file for more information about contributing.
+
+## Many Thanks To Our Contributors⚡
+
+<a href="https://github.com/arc53/DocsGPT/graphs/contributors" alt="View Contributors">
+  <img src="https://contrib.rocks/image?repo=arc53/DocsGPT" alt="Contributors" />
+</a>
+
+## License
+
+The source code license is [MIT](https://opensource.org/license/mit/), as described in the [LICENSE](LICENSE) file.
+
+Built with [:bird: :link: LangChain](https://github.com/hwchase17/langchain)

application/.env_sample

@@ -1,9 +1,8 @@
 API_KEY=your_api_key
 EMBEDDINGS_KEY=your_api_key
-CELERY_BROKER_URL=redis://localhost:6379/0
-CELERY_RESULT_BACKEND=redis://localhost:6379/1
-MONGO_URI=mongodb://localhost:27017/docsgpt
 API_URL=http://localhost:7091
+FLASK_APP=application/app.py
+FLASK_DEBUG=true
 
 #For OPENAI on Azure
 OPENAI_API_BASE=

application/Dockerfile

@@ -1,19 +1,25 @@
-FROM python:3.10-slim-bullseye as builder
+FROM python:3.11-slim-bullseye as builder
 
 # Tiktoken requires Rust toolchain, so build it in a separate stage
 RUN apt-get update && apt-get install -y gcc curl
 RUN curl https://sh.rustup.rs -sSf | sh -s -- -y && apt-get install --reinstall libc6-dev -y
 ENV PATH="/root/.cargo/bin:${PATH}"
-RUN pip install --upgrade pip && pip install tiktoken==0.3.3
+RUN pip install --upgrade pip && pip install tiktoken==0.5.2
 COPY requirements.txt .
 RUN pip install -r requirements.txt
+RUN apt-get install -y wget unzip
+RUN wget https://d3dg1063dc54p9.cloudfront.net/models/embeddings/mpnet-base-v2.zip
+RUN unzip mpnet-base-v2.zip -d model
+RUN rm mpnet-base-v2.zip
 
-FROM python:3.10-slim-bullseye
+FROM python:3.11-slim-bullseye
 
 # Copy pre-built packages and binaries from builder stage
 COPY --from=builder /usr/local/ /usr/local/
 
 WORKDIR /app
+COPY --from=builder /model /app/model
+
 COPY . /app/application
 ENV FLASK_APP=app.py
 ENV FLASK_DEBUG=true

application/api/answer/routes.py

@@ -0,0 +1,374 @@
+import asyncio
+import os
+from flask import Blueprint, request, Response
+import json
+import datetime
+import logging
+import traceback
+
+from pymongo import MongoClient
+from bson.objectid import ObjectId
+from transformers import GPT2TokenizerFast
+
+
+
+from application.core.settings import settings
+from application.vectorstore.vector_creator import VectorCreator
+from application.llm.llm_creator import LLMCreator
+from application.error import bad_request
+
+
+
+logger = logging.getLogger(__name__)
+
+mongo = MongoClient(settings.MONGO_URI)
+db = mongo["docsgpt"]
+conversations_collection = db["conversations"]
+vectors_collection = db["vectors"]
+prompts_collection = db["prompts"]
+answer = Blueprint('answer', __name__)
+
+if settings.LLM_NAME == "gpt4":
+    gpt_model = 'gpt-4'
+elif settings.LLM_NAME == "anthropic":
+    gpt_model = 'claude-2'
+else:
+    gpt_model = 'gpt-3.5-turbo'
+
+# load the prompts
+current_dir = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+with open(os.path.join(current_dir, "prompts", "chat_combine_default.txt"), "r") as f:
+    chat_combine_template = f.read()
+
+with open(os.path.join(current_dir, "prompts", "chat_reduce_prompt.txt"), "r") as f:
+    chat_reduce_template = f.read()
+
+with open(os.path.join(current_dir, "prompts", "chat_combine_creative.txt"), "r") as f:
+    chat_combine_creative = f.read()
+
+with open(os.path.join(current_dir, "prompts", "chat_combine_strict.txt"), "r") as f:
+    chat_combine_strict = f.read()    
+
+api_key_set = settings.API_KEY is not None
+embeddings_key_set = settings.EMBEDDINGS_KEY is not None
+
+
+async def async_generate(chain, question, chat_history):
+    result = await chain.arun({"question": question, "chat_history": chat_history})
+    return result
+
+
+def count_tokens(string):
+    tokenizer = GPT2TokenizerFast.from_pretrained('gpt2')
+    return len(tokenizer(string)['input_ids'])
+
+
+def run_async_chain(chain, question, chat_history):
+    loop = asyncio.new_event_loop()
+    asyncio.set_event_loop(loop)
+    result = {}
+    try:
+        answer = loop.run_until_complete(async_generate(chain, question, chat_history))
+    finally:
+        loop.close()
+    result["answer"] = answer
+    return result
+
+
+def get_vectorstore(data):
+    if "active_docs" in data:
+        if data["active_docs"].split("/")[0] == "default":
+                vectorstore = ""
+        elif data["active_docs"].split("/")[0] == "local":
+            vectorstore = "indexes/" + data["active_docs"]
+        else:
+            vectorstore = "vectors/" + data["active_docs"]
+        if data["active_docs"] == "default":
+            vectorstore = ""
+    else:
+        vectorstore = ""
+    vectorstore = os.path.join("application", vectorstore)
+    return vectorstore
+
+
+def is_azure_configured():
+    return settings.OPENAI_API_BASE and settings.OPENAI_API_VERSION and settings.AZURE_DEPLOYMENT_NAME
+
+
+def complete_stream(question, docsearch, chat_history, api_key, prompt_id, conversation_id):
+    llm = LLMCreator.create_llm(settings.LLM_NAME, api_key=api_key)
+
+    if prompt_id == 'default':
+        prompt = chat_combine_template
+    elif prompt_id == 'creative':
+        prompt = chat_combine_creative
+    elif prompt_id == 'strict':
+        prompt = chat_combine_strict
+    else:
+        prompt = prompts_collection.find_one({"_id": ObjectId(prompt_id)})["content"]
+
+    docs = docsearch.search(question, k=2)
+    if settings.LLM_NAME == "llama.cpp":
+        docs = [docs[0]]
+    # join all page_content together with a newline
+    docs_together = "\n".join([doc.page_content for doc in docs])
+    p_chat_combine = prompt.replace("{summaries}", docs_together)
+    messages_combine = [{"role": "system", "content": p_chat_combine}]
+    source_log_docs = []
+    for doc in docs:
+        if doc.metadata:
+            source_log_docs.append({"title": doc.metadata['title'].split('/')[-1], "text": doc.page_content})
+        else:
+            source_log_docs.append({"title": doc.page_content, "text": doc.page_content})
+
+    if len(chat_history) > 1:
+        tokens_current_history = 0
+        # count tokens in history
+        chat_history.reverse()
+        for i in chat_history:
+            if "prompt" in i and "response" in i:
+                tokens_batch = count_tokens(i["prompt"]) + count_tokens(i["response"])
+                if tokens_current_history + tokens_batch < settings.TOKENS_MAX_HISTORY:
+                    tokens_current_history += tokens_batch
+                    messages_combine.append({"role": "user", "content": i["prompt"]})
+                    messages_combine.append({"role": "system", "content": i["response"]})
+    messages_combine.append({"role": "user", "content": question})
+
+    response_full = ""
+    completion = llm.gen_stream(model=gpt_model, engine=settings.AZURE_DEPLOYMENT_NAME,
+                                messages=messages_combine)
+    for line in completion:
+        data = json.dumps({"answer": str(line)})
+        response_full += str(line)
+        yield f"data: {data}\n\n"
+
+    # save conversation to database
+    if conversation_id is not None:
+        conversations_collection.update_one(
+            {"_id": ObjectId(conversation_id)},
+            {"$push": {"queries": {"prompt": question, "response": response_full, "sources": source_log_docs}}},
+        )
+
+    else:
+        # create new conversation
+        # generate summary
+        messages_summary = [{"role": "assistant", "content": "Summarise following conversation in no more than 3 "
+                                                             "words, respond ONLY with the summary, use the same "
+                                                             "language as the system \n\nUser: " + question + "\n\n" +
+                                                             "AI: " +
+                                                             response_full},
+                            {"role": "user", "content": "Summarise following conversation in no more than 3 words, "
+                                                        "respond ONLY with the summary, use the same language as the "
+                                                        "system"}]
+
+        completion = llm.gen(model=gpt_model, engine=settings.AZURE_DEPLOYMENT_NAME,
+                             messages=messages_summary, max_tokens=30)
+        conversation_id = conversations_collection.insert_one(
+            {"user": "local",
+             "date": datetime.datetime.utcnow(),
+             "name": completion,
+             "queries": [{"prompt": question, "response": response_full, "sources": source_log_docs}]}
+        ).inserted_id
+
+    # send data.type = "end" to indicate that the stream has ended as json
+    data = json.dumps({"type": "id", "id": str(conversation_id)})
+    yield f"data: {data}\n\n"
+    data = json.dumps({"type": "end"})
+    yield f"data: {data}\n\n"
+
+
+@answer.route("/stream", methods=["POST"])
+def stream():
+    data = request.get_json()
+    # get parameter from url question
+    question = data["question"]
+    history = data["history"]
+    # history to json object from string
+    history = json.loads(history)
+    conversation_id = data["conversation_id"]
+    if 'prompt_id' in data:
+        prompt_id = data["prompt_id"]
+    else:
+        prompt_id = 'default'
+
+    # check if active_docs is set
+
+    if not api_key_set:
+        api_key = data["api_key"]
+    else:
+        api_key = settings.API_KEY
+    if not embeddings_key_set:
+        embeddings_key = data["embeddings_key"]
+    else:
+        embeddings_key = settings.EMBEDDINGS_KEY
+    if "active_docs" in data:
+        vectorstore = get_vectorstore({"active_docs": data["active_docs"]})
+    else:
+        vectorstore = ""
+    docsearch = VectorCreator.create_vectorstore(settings.VECTOR_STORE, vectorstore, embeddings_key)
+
+    return Response(
+        complete_stream(question, docsearch,
+                        chat_history=history, api_key=api_key,
+                        prompt_id=prompt_id,
+                        conversation_id=conversation_id), mimetype="text/event-stream"
+    )
+
+
+@answer.route("/api/answer", methods=["POST"])
+def api_answer():
+    data = request.get_json()
+    question = data["question"]
+    history = data["history"]
+    if "conversation_id" not in data:
+        conversation_id = None
+    else:
+        conversation_id = data["conversation_id"]
+    print("-" * 5)
+    if not api_key_set:
+        api_key = data["api_key"]
+    else:
+        api_key = settings.API_KEY
+    if not embeddings_key_set:
+        embeddings_key = data["embeddings_key"]
+    else:
+        embeddings_key = settings.EMBEDDINGS_KEY
+    if 'prompt_id' in data:
+        prompt_id = data["prompt_id"]
+    else:
+        prompt_id = 'default'
+
+    if prompt_id == 'default':
+        prompt = chat_combine_template
+    elif prompt_id == 'creative':
+        prompt = chat_combine_creative
+    elif prompt_id == 'strict':
+        prompt = chat_combine_strict
+    else:
+        prompt = prompts_collection.find_one({"_id": ObjectId(prompt_id)})["content"]
+
+    # use try and except  to check for exception
+    try:
+        # check if the vectorstore is set
+        vectorstore = get_vectorstore(data)
+        # loading the index and the store and the prompt template
+        # Note if you have used other embeddings than OpenAI, you need to change the embeddings
+        docsearch = VectorCreator.create_vectorstore(settings.VECTOR_STORE, vectorstore, embeddings_key)
+
+
+        llm = LLMCreator.create_llm(settings.LLM_NAME, api_key=api_key)
+
+
+
+        docs = docsearch.search(question, k=2)
+        # join all page_content together with a newline
+        docs_together = "\n".join([doc.page_content for doc in docs])
+        p_chat_combine = prompt.replace("{summaries}", docs_together)
+        messages_combine = [{"role": "system", "content": p_chat_combine}]
+        source_log_docs = []
+        for doc in docs:
+            if doc.metadata:
+                source_log_docs.append({"title": doc.metadata['title'].split('/')[-1], "text": doc.page_content})
+            else:
+                source_log_docs.append({"title": doc.page_content, "text": doc.page_content})
+        # join all page_content together with a newline
+
+
+        if len(history) > 1:
+            tokens_current_history = 0
+            # count tokens in history
+            history.reverse()
+            for i in history:
+                if "prompt" in i and "response" in i:
+                    tokens_batch = count_tokens(i["prompt"]) + count_tokens(i["response"])
+                    if tokens_current_history + tokens_batch < settings.TOKENS_MAX_HISTORY:
+                        tokens_current_history += tokens_batch
+                        messages_combine.append({"role": "user", "content": i["prompt"]})
+                        messages_combine.append({"role": "system", "content": i["response"]})
+        messages_combine.append({"role": "user", "content": question})
+
+
+        completion = llm.gen(model=gpt_model, engine=settings.AZURE_DEPLOYMENT_NAME,
+                                    messages=messages_combine)
+
+
+        result = {"answer": completion, "sources": source_log_docs}
+        logger.debug(result)
+
+        # generate conversationId
+        if conversation_id is not None:
+            conversations_collection.update_one(
+                {"_id": ObjectId(conversation_id)},
+                {"$push": {"queries": {"prompt": question,
+                                       "response": result["answer"], "sources": result['sources']}}},
+            )
+
+        else:
+            # create new conversation
+            # generate summary
+            messages_summary = [
+                {"role": "assistant", "content": "Summarise following conversation in no more than 3 words, "
+                    "respond ONLY with the summary, use the same language as the system \n\n"
+                    "User: " + question + "\n\n" + "AI: " + result["answer"]},
+                {"role": "user", "content": "Summarise following conversation in no more than 3 words, "
+                    "respond ONLY with the summary, use the same language as the system"}
+            ]
+
+            completion = llm.gen(
+                model=gpt_model,
+                engine=settings.AZURE_DEPLOYMENT_NAME,
+                messages=messages_summary,
+                max_tokens=30
+            )
+            conversation_id = conversations_collection.insert_one(
+                {"user": "local",
+                "date": datetime.datetime.utcnow(),
+                "name": completion,
+                "queries": [{"prompt": question, "response": result["answer"], "sources": source_log_docs}]}
+            ).inserted_id
+
+        result["conversation_id"] = str(conversation_id)
+
+        # mock result
+        # result = {
+        #     "answer": "The answer is 42",
+        #     "sources": ["https://en.wikipedia.org/wiki/42_(number)", "https://en.wikipedia.org/wiki/42_(number)"]
+        # }
+        return result
+    except Exception as e:
+        # print whole traceback
+        traceback.print_exc()
+        print(str(e))
+        return bad_request(500, str(e))
+
+
+@answer.route("/api/search", methods=["POST"])
+def api_search():
+    data = request.get_json()
+    # get parameter from url question
+    question = data["question"]
+
+    if not embeddings_key_set:
+        if "embeddings_key" in data:
+            embeddings_key = data["embeddings_key"]
+        else:
+            embeddings_key = settings.EMBEDDINGS_KEY
+    else:
+        embeddings_key = settings.EMBEDDINGS_KEY
+    if "active_docs" in data:
+        vectorstore = get_vectorstore({"active_docs": data["active_docs"]})
+    else:
+        vectorstore = ""
+    docsearch = VectorCreator.create_vectorstore(settings.VECTOR_STORE, vectorstore, embeddings_key)
+
+    docs = docsearch.search(question, k=2)
+
+    source_log_docs = []
+    for doc in docs:
+        if doc.metadata:
+            source_log_docs.append({"title": doc.metadata['title'].split('/')[-1], "text": doc.page_content})
+        else:
+            source_log_docs.append({"title": doc.page_content, "text": doc.page_content})
+        #yield f"data:{data}\n\n"
+    return source_log_docs
+

application/api/internal/routes.py

@@ -0,0 +1,69 @@
+import os
+import datetime
+from flask import Blueprint, request, send_from_directory
+from pymongo import MongoClient
+from werkzeug.utils import secure_filename
+
+
+from application.core.settings import settings
+mongo = MongoClient(settings.MONGO_URI)
+db = mongo["docsgpt"]
+conversations_collection = db["conversations"]
+vectors_collection = db["vectors"]
+
+current_dir = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+
+internal = Blueprint('internal', __name__)
+@internal.route("/api/download", methods=["get"])
+def download_file():
+    user = secure_filename(request.args.get("user"))
+    job_name = secure_filename(request.args.get("name"))
+    filename = secure_filename(request.args.get("file"))
+    save_dir = os.path.join(current_dir, settings.UPLOAD_FOLDER, user, job_name)
+    return send_from_directory(save_dir, filename, as_attachment=True)
+
+
+
+@internal.route("/api/upload_index", methods=["POST"])
+def upload_index_files():
+    """Upload two files(index.faiss, index.pkl) to the user's folder."""
+    if "user" not in request.form:
+        return {"status": "no user"}
+    user = secure_filename(request.form["user"])
+    if "name" not in request.form:
+        return {"status": "no name"}
+    job_name = secure_filename(request.form["name"])
+    save_dir = os.path.join(current_dir, "indexes", user, job_name)
+    if settings.VECTOR_STORE == "faiss":
+        if "file_faiss" not in request.files:
+            print("No file part")
+            return {"status": "no file"}
+        file_faiss = request.files["file_faiss"]
+        if file_faiss.filename == "":
+            return {"status": "no file name"}
+        if "file_pkl" not in request.files:
+            print("No file part")
+            return {"status": "no file"}
+        file_pkl = request.files["file_pkl"]
+        if file_pkl.filename == "":
+            return {"status": "no file name"}
+        # saves index files
+        
+        if not os.path.exists(save_dir):
+            os.makedirs(save_dir)
+        file_faiss.save(os.path.join(save_dir, "index.faiss"))
+        file_pkl.save(os.path.join(save_dir, "index.pkl"))
+    # create entry in vectors_collection
+    vectors_collection.insert_one(
+        {
+            "user": user,
+            "name": job_name,
+            "language": job_name,
+            "location": save_dir,
+            "date": datetime.datetime.now().strftime("%d/%m/%Y %H:%M:%S"),
+            "model": settings.EMBEDDINGS_NAME,
+            "type": "local",
+        }
+    )
+    return {"status": "ok"}

application/api/user/routes.py

@@ -0,0 +1,321 @@
+import os
+from flask import Blueprint, request, jsonify
+import requests
+from pymongo import MongoClient
+from bson.objectid import ObjectId
+from werkzeug.utils import secure_filename
+
+from application.api.user.tasks import ingest
+
+from application.core.settings import settings
+from application.vectorstore.vector_creator import VectorCreator
+
+mongo = MongoClient(settings.MONGO_URI)
+db = mongo["docsgpt"]
+conversations_collection = db["conversations"]
+vectors_collection = db["vectors"]
+prompts_collection = db["prompts"]
+feedback_collection = db["feedback"]
+user = Blueprint('user', __name__)
+
+current_dir = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+@user.route("/api/delete_conversation", methods=["POST"])
+def delete_conversation():
+    # deletes a conversation from the database
+    conversation_id = request.args.get("id")
+    # write to mongodb
+    conversations_collection.delete_one(
+        {
+            "_id": ObjectId(conversation_id),
+        }
+    )
+
+    return {"status": "ok"}
+
+@user.route("/api/get_conversations", methods=["get"])
+def get_conversations():
+    # provides a list of conversations
+    conversations = conversations_collection.find().sort("date", -1)
+    list_conversations = []
+    for conversation in conversations:
+        list_conversations.append({"id": str(conversation["_id"]), "name": conversation["name"]})
+
+    #list_conversations = [{"id": "default", "name": "default"}, {"id": "jeff", "name": "jeff"}]
+
+    return jsonify(list_conversations)
+
+
+@user.route("/api/get_single_conversation", methods=["get"])
+def get_single_conversation():
+    # provides data for a conversation
+    conversation_id = request.args.get("id")
+    conversation = conversations_collection.find_one({"_id": ObjectId(conversation_id)})
+    return jsonify(conversation['queries'])
+
+@user.route("/api/update_conversation_name", methods=["POST"])
+def update_conversation_name():
+    # update data for a conversation
+    data = request.get_json()
+    id = data["id"]
+    name = data["name"]
+    conversations_collection.update_one({"_id": ObjectId(id)},{"$set":{"name":name}})
+    return {"status": "ok"}
+
+
+@user.route("/api/feedback", methods=["POST"])
+def api_feedback():
+    data = request.get_json()
+    question = data["question"]
+    answer = data["answer"]
+    feedback = data["feedback"]
+
+
+    feedback_collection.insert_one(
+        {
+            "question": question,
+            "answer": answer,
+            "feedback": feedback,
+        }
+    )
+    return {"status": "ok"}
+
+@user.route("/api/delete_by_ids", methods=["get"])
+def delete_by_ids():
+    """Delete by ID. These are the IDs in the vectorstore"""
+
+    ids = request.args.get("path")
+    if not ids:
+        return {"status": "error"}
+
+    if settings.VECTOR_STORE == "faiss":
+        result = vectors_collection.delete_index(ids=ids)
+        if result:
+            return {"status": "ok"}
+    return {"status": "error"}
+
+@user.route("/api/delete_old", methods=["get"])
+def delete_old():
+    """Delete old indexes."""
+    import shutil
+
+    path = request.args.get("path")
+    dirs = path.split("/")
+    dirs_clean = []
+    for i in range(0, len(dirs)):
+        dirs_clean.append(secure_filename(dirs[i]))
+    # check that path strats with indexes or vectors
+
+    if dirs_clean[0] not in ["indexes", "vectors"]:
+        return {"status": "error"}
+    path_clean = "/".join(dirs_clean)
+    vectors_collection.delete_one({"name": dirs_clean[-1], 'user': dirs_clean[-2]})
+    if settings.VECTOR_STORE == "faiss":
+        try:
+            shutil.rmtree(os.path.join(current_dir, path_clean))
+        except FileNotFoundError:
+            pass
+    else:
+        vetorstore = VectorCreator.create_vectorstore(
+            settings.VECTOR_STORE, path=os.path.join(current_dir, path_clean)
+        )
+        vetorstore.delete_index()
+        
+    return {"status": "ok"}
+
+@user.route("/api/upload", methods=["POST"])
+def upload_file():
+    """Upload a file to get vectorized and indexed."""
+    if "user" not in request.form:
+        return {"status": "no user"}
+    user = secure_filename(request.form["user"])
+    if "name" not in request.form:
+        return {"status": "no name"}
+    job_name = secure_filename(request.form["name"])
+    # check if the post request has the file part
+    if "file" not in request.files:
+        print("No file part")
+        return {"status": "no file"}
+    file = request.files["file"]
+    if file.filename == "":
+        return {"status": "no file name"}
+
+    if file:
+        filename = secure_filename(file.filename)
+        # save dir
+        save_dir = os.path.join(current_dir, settings.UPLOAD_FOLDER, user, job_name)
+        # create dir if not exists
+        if not os.path.exists(save_dir):
+            os.makedirs(save_dir)
+
+        file.save(os.path.join(save_dir, filename))
+        task = ingest.delay(settings.UPLOAD_FOLDER, [".rst", ".md", ".pdf", ".txt", ".docx", 
+        ".csv", ".epub", ".html", ".mdx"],
+         job_name, filename, user)
+        # task id
+        task_id = task.id
+        return {"status": "ok", "task_id": task_id}
+    else:
+        return {"status": "error"}
+
+@user.route("/api/task_status", methods=["GET"])
+def task_status():
+    """Get celery job status."""
+    task_id = request.args.get("task_id")
+    from application.celery import celery
+    task = celery.AsyncResult(task_id)
+    task_meta = task.info
+    return {"status": task.status, "result": task_meta}
+
+
+@user.route("/api/combine", methods=["GET"])
+def combined_json():
+    user = "local"
+    """Provide json file with combined available indexes."""
+    # get json from https://d3dg1063dc54p9.cloudfront.net/combined.json
+
+    data = [
+        {
+            "name": "default",
+            "language": "default",
+            "version": "",
+            "description": "default",
+            "fullName": "default",
+            "date": "default",
+            "docLink": "default",
+            "model": settings.EMBEDDINGS_NAME,
+            "location": "remote",
+        }
+    ]
+    # structure: name, language, version, description, fullName, date, docLink
+    # append data from vectors_collection
+    for index in vectors_collection.find({"user": user}):
+        data.append(
+            {
+                "name": index["name"],
+                "language": index["language"],
+                "version": "",
+                "description": index["name"],
+                "fullName": index["name"],
+                "date": index["date"],
+                "docLink": index["location"],
+                "model": settings.EMBEDDINGS_NAME,
+                "location": "local",
+            }
+        )
+    if settings.VECTOR_STORE == "faiss":
+        data_remote = requests.get("https://d3dg1063dc54p9.cloudfront.net/combined.json").json()
+        for index in data_remote:
+            index["location"] = "remote"
+            data.append(index)
+
+    return jsonify(data)
+
+
+@user.route("/api/docs_check", methods=["POST"])
+def check_docs():
+    # check if docs exist in a vectorstore folder
+    data = request.get_json()
+    # split docs on / and take first part
+    if data["docs"].split("/")[0] == "local":
+        return {"status": "exists"}
+    vectorstore = "vectors/" + data["docs"]
+    base_path = "https://raw.githubusercontent.com/arc53/DocsHUB/main/"
+    if os.path.exists(vectorstore) or data["docs"] == "default":
+        return {"status": "exists"}
+    else:
+        r = requests.get(base_path + vectorstore + "index.faiss")
+
+        if r.status_code != 200:
+            return {"status": "null"}
+        else:
+            if not os.path.exists(vectorstore):
+                os.makedirs(vectorstore)
+            with open(vectorstore + "index.faiss", "wb") as f:
+                f.write(r.content)
+
+            # download the store
+            r = requests.get(base_path + vectorstore + "index.pkl")
+            with open(vectorstore + "index.pkl", "wb") as f:
+                f.write(r.content)
+
+        return {"status": "loaded"}
+
+@user.route("/api/create_prompt", methods=["POST"])
+def create_prompt():
+    data = request.get_json()
+    content = data["content"]
+    name = data["name"]
+    if name == "":
+        return {"status": "error"}
+    user = "local"
+    resp = prompts_collection.insert_one(
+        {
+            "name": name,
+            "content": content,
+            "user": user,
+        }
+    )
+    new_id = str(resp.inserted_id)
+    return {"id": new_id}
+
+@user.route("/api/get_prompts", methods=["GET"])
+def get_prompts():
+    user = "local"
+    prompts = prompts_collection.find({"user": user})
+    list_prompts = []
+    list_prompts.append({"id": "default", "name": "default", "type": "public"})
+    list_prompts.append({"id": "creative", "name": "creative", "type": "public"})
+    list_prompts.append({"id": "strict", "name": "strict", "type": "public"})
+    for prompt in prompts:
+        list_prompts.append({"id": str(prompt["_id"]), "name": prompt["name"], "type": "private"})
+
+    return jsonify(list_prompts)
+
+@user.route("/api/get_single_prompt", methods=["GET"])
+def get_single_prompt():
+    prompt_id = request.args.get("id")
+    if prompt_id == 'default':
+        with open(os.path.join(current_dir, "prompts", "chat_combine_default.txt"), "r") as f:
+            chat_combine_template = f.read()
+        return jsonify({"content": chat_combine_template})
+    elif prompt_id == 'creative':
+        with open(os.path.join(current_dir, "prompts", "chat_combine_creative.txt"), "r") as f:
+            chat_reduce_creative = f.read()
+        return jsonify({"content": chat_reduce_creative})
+    elif prompt_id == 'strict':
+        with open(os.path.join(current_dir, "prompts", "chat_combine_strict.txt"), "r") as f:
+            chat_reduce_strict = f.read()   
+        return jsonify({"content": chat_reduce_strict})
+
+
+    prompt = prompts_collection.find_one({"_id": ObjectId(prompt_id)})
+    return jsonify({"content": prompt["content"]})
+
+@user.route("/api/delete_prompt", methods=["POST"])
+def delete_prompt():
+    data = request.get_json()
+    id = data["id"]
+    prompts_collection.delete_one(
+        {
+            "_id": ObjectId(id),
+        }
+    )
+    return {"status": "ok"}
+
+@user.route("/api/update_prompt", methods=["POST"])
+def update_prompt_name():
+    data = request.get_json()
+    id = data["id"]
+    name = data["name"]
+    content = data["content"]
+    # check if name is null
+    if name == "":
+        return {"status": "error"}
+    prompts_collection.update_one({"_id": ObjectId(id)},{"$set":{"name":name, "content": content}})
+    return {"status": "ok"}
+
+
+
+
+

application/api/user/tasks.py

@@ -0,0 +1,7 @@
+from application.worker import ingest_worker
+from application.celery import celery
+
+@celery.task(bind=True)
+def ingest(self, directory, formats, name_job, filename, user):
+    resp = ingest_worker(self, directory, formats, name_job, filename, user)
+    return resp

application/app.py

@@ -1,719 +1,44 @@
-import asyncio
-import datetime
-import http.client
-import json
-import logging
-import os
 import platform
-import traceback
-
 import dotenv
-import openai
-import requests
-from celery import Celery
-from celery.result import AsyncResult
-from flask import Flask, request, render_template, send_from_directory, jsonify, Response
-from langchain import FAISS
-from langchain import VectorDBQA, Cohere, OpenAI
-from langchain.chains import LLMChain, ConversationalRetrievalChain
-from langchain.chains.conversational_retrieval.prompts import CONDENSE_QUESTION_PROMPT
-from langchain.chains.question_answering import load_qa_chain
-from langchain.chat_models import ChatOpenAI, AzureChatOpenAI
-from langchain.embeddings import (
-    OpenAIEmbeddings,
-    HuggingFaceHubEmbeddings,
-    CohereEmbeddings,
-    HuggingFaceInstructEmbeddings,
-)
-from langchain.prompts import PromptTemplate
-from langchain.prompts.chat import (
-    ChatPromptTemplate,
-    SystemMessagePromptTemplate,
-    HumanMessagePromptTemplate,
-    AIMessagePromptTemplate,
-)
-from langchain.schema import HumanMessage, AIMessage
-from pymongo import MongoClient
-from werkzeug.utils import secure_filename
-
+from application.celery import celery
+from flask import Flask, request, redirect
 from application.core.settings import settings
-from application.error import bad_request
-from application.worker import ingest_worker
-from bson.objectid import ObjectId
-
-# os.environ["LANGCHAIN_HANDLER"] = "langchain"
-
-logger = logging.getLogger(__name__)
-if settings.LLM_NAME == "gpt4":
-    gpt_model = 'gpt-4'
-else:
-    gpt_model = 'gpt-3.5-turbo'
-
-
-if settings.SELF_HOSTED_MODEL:
-    from langchain.llms import HuggingFacePipeline
-    from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
-
-    model_id = settings.LLM_NAME # hf model id (Arc53/docsgpt-7b-falcon, Arc53/docsgpt-14b)
-    tokenizer = AutoTokenizer.from_pretrained(model_id)
-    model = AutoModelForCausalLM.from_pretrained(model_id)
-    pipe = pipeline(
-        "text-generation", model=model,
-        tokenizer=tokenizer, max_new_tokens=2000,
-        device_map="auto", eos_token_id=tokenizer.eos_token_id
-    )
-    hf = HuggingFacePipeline(pipeline=pipe)
-
-# Redirect PosixPath to WindowsPath on Windows
+from application.api.user.routes import user
+from application.api.answer.routes import answer
+from application.api.internal.routes import internal
 
 if platform.system() == "Windows":
     import pathlib
-
-    temp = pathlib.PosixPath
     pathlib.PosixPath = pathlib.WindowsPath
 
-# loading the .env file
 dotenv.load_dotenv()
 
-# load the prompts
-current_dir = os.path.dirname(os.path.abspath(__file__))
-with open(os.path.join(current_dir, "prompts", "combine_prompt.txt"), "r") as f:
-    template = f.read()
-
-with open(os.path.join(current_dir, "prompts", "combine_prompt_hist.txt"), "r") as f:
-    template_hist = f.read()
-
-with open(os.path.join(current_dir, "prompts", "question_prompt.txt"), "r") as f:
-    template_quest = f.read()
-
-with open(os.path.join(current_dir, "prompts", "chat_combine_prompt.txt"), "r") as f:
-    chat_combine_template = f.read()
-
-with open(os.path.join(current_dir, "prompts", "chat_reduce_prompt.txt"), "r") as f:
-    chat_reduce_template = f.read()
-
-api_key_set = settings.API_KEY is not None
-embeddings_key_set = settings.EMBEDDINGS_KEY is not None
-
 app = Flask(__name__)
-app.config["UPLOAD_FOLDER"] = UPLOAD_FOLDER = "inputs"
-app.config["CELERY_BROKER_URL"] = settings.CELERY_BROKER_URL
-app.config["CELERY_RESULT_BACKEND"] = settings.CELERY_RESULT_BACKEND
-app.config["MONGO_URI"] = settings.MONGO_URI
-celery = Celery()
+app.register_blueprint(user)
+app.register_blueprint(answer)
+app.register_blueprint(internal)
+app.config.update(
+    UPLOAD_FOLDER="inputs",
+    CELERY_BROKER_URL=settings.CELERY_BROKER_URL,
+    CELERY_RESULT_BACKEND=settings.CELERY_RESULT_BACKEND,
+    MONGO_URI=settings.MONGO_URI
+)
 celery.config_from_object("application.celeryconfig")
-mongo = MongoClient(app.config["MONGO_URI"])
-db = mongo["docsgpt"]
-vectors_collection = db["vectors"]
-conversations_collection = db["conversations"]
-
-
-async def async_generate(chain, question, chat_history):
-    result = await chain.arun({"question": question, "chat_history": chat_history})
-    return result
-
-
-def run_async_chain(chain, question, chat_history):
-    loop = asyncio.new_event_loop()
-    asyncio.set_event_loop(loop)
-    result = {}
-    try:
-        answer = loop.run_until_complete(async_generate(chain, question, chat_history))
-    finally:
-        loop.close()
-    result["answer"] = answer
-    return result
-
-
-def get_vectorstore(data):
-    if "active_docs" in data:
-        if data["active_docs"].split("/")[0] == "local":
-            if data["active_docs"].split("/")[1] == "default":
-                vectorstore = ""
-            else:
-                vectorstore = "indexes/" + data["active_docs"]
-        else:
-            vectorstore = "vectors/" + data["active_docs"]
-        if data["active_docs"] == "default":
-            vectorstore = ""
-    else:
-        vectorstore = ""
-    vectorstore = os.path.join("application", vectorstore)
-    return vectorstore
-
-
-def get_docsearch(vectorstore, embeddings_key):
-    if settings.EMBEDDINGS_NAME == "openai_text-embedding-ada-002":
-        if is_azure_configured():
-            os.environ["OPENAI_API_TYPE"] = "azure"
-            openai_embeddings = OpenAIEmbeddings(model=settings.AZURE_EMBEDDINGS_DEPLOYMENT_NAME)
-        else:
-            openai_embeddings = OpenAIEmbeddings(openai_api_key=embeddings_key)
-        docsearch = FAISS.load_local(vectorstore, openai_embeddings)
-    elif settings.EMBEDDINGS_NAME == "huggingface_sentence-transformers/all-mpnet-base-v2":
-        docsearch = FAISS.load_local(vectorstore, HuggingFaceHubEmbeddings())
-    elif settings.EMBEDDINGS_NAME == "huggingface_hkunlp/instructor-large":
-        docsearch = FAISS.load_local(vectorstore, HuggingFaceInstructEmbeddings())
-    elif settings.EMBEDDINGS_NAME == "cohere_medium":
-        docsearch = FAISS.load_local(vectorstore, CohereEmbeddings(cohere_api_key=embeddings_key))
-    return docsearch
-
-
-@celery.task(bind=True)
-def ingest(self, directory, formats, name_job, filename, user):
-    resp = ingest_worker(self, directory, formats, name_job, filename, user)
-    return resp
-
 
 @app.route("/")
 def home():
-    return render_template(
-        "index.html", api_key_set=api_key_set, llm_choice=settings.LLM_NAME, embeddings_choice=settings.EMBEDDINGS_NAME
-    )
-
-
-def complete_stream(question, docsearch, chat_history, api_key, conversation_id):
-    openai.api_key = api_key
-    if is_azure_configured():
-        logger.debug("in Azure")
-        openai.api_type = "azure"
-        openai.api_version = settings.OPENAI_API_VERSION
-        openai.api_base = settings.OPENAI_API_BASE
-        llm = AzureChatOpenAI(
-            openai_api_key=api_key,
-            openai_api_base=settings.OPENAI_API_BASE,
-            openai_api_version=settings.OPENAI_API_VERSION,
-            deployment_name=settings.AZURE_DEPLOYMENT_NAME,
-        )
+    if request.remote_addr in ('0.0.0.0', '127.0.0.1', 'localhost', '172.18.0.1'):
+        return redirect('http://localhost:5173')
     else:
-        logger.debug("plain OpenAI")
-        llm = ChatOpenAI(openai_api_key=api_key)
-    docs = docsearch.similarity_search(question, k=2)
-    # join all page_content together with a newline
-    docs_together = "\n".join([doc.page_content for doc in docs])
-    p_chat_combine = chat_combine_template.replace("{summaries}", docs_together)
-    messages_combine = [{"role": "system", "content": p_chat_combine}]
-    source_log_docs = []
-    for doc in docs:
-        if doc.metadata:
-            data = json.dumps({"type": "source", "doc": doc.page_content, "metadata": doc.metadata})
-            source_log_docs.append({"title": doc.metadata['title'].split('/')[-1], "text": doc.page_content})
-        else:
-            data = json.dumps({"type": "source", "doc": doc.page_content})
-            source_log_docs.append({"title": doc.page_content, "text": doc.page_content})
-        yield f"data:{data}\n\n"
+        return 'Welcome to DocsGPT Backend!'
 
-    if len(chat_history) > 1:
-        tokens_current_history = 0
-        # count tokens in history
-        chat_history.reverse()
-        for i in chat_history:
-            if "prompt" in i and "response" in i:
-                tokens_batch = llm.get_num_tokens(i["prompt"]) + llm.get_num_tokens(i["response"])
-                if tokens_current_history + tokens_batch < settings.TOKENS_MAX_HISTORY:
-                    tokens_current_history += tokens_batch
-                    messages_combine.append({"role": "user", "content": i["prompt"]})
-                    messages_combine.append({"role": "system", "content": i["response"]})
-    messages_combine.append({"role": "user", "content": question})
-    completion = openai.ChatCompletion.create(model=gpt_model, engine=settings.AZURE_DEPLOYMENT_NAME,
-                                              messages=messages_combine, stream=True, max_tokens=500, temperature=0)
-    reponse_full = ""
-    for line in completion:
-        if "content" in line["choices"][0]["delta"]:
-            # check if the delta contains content
-            data = json.dumps({"answer": str(line["choices"][0]["delta"]["content"])})
-            reponse_full += str(line["choices"][0]["delta"]["content"])
-            yield f"data: {data}\n\n"
-    # save conversation to database
-    if conversation_id is not None:
-        conversations_collection.update_one(
-            {"_id": ObjectId(conversation_id)},
-            {"$push": {"queries": {"prompt": question, "response": reponse_full, "sources": source_log_docs}}},
-        )
-
-    else:
-        # create new conversation
-        # generate summary
-        messages_summary = [{"role": "assistant", "content": "Summarise following conversation in no more than 3 "
-                                                             "words, respond ONLY with the summary, use the same "
-                                                             "language as the system \n\nUser: " + question + "\n\n" +
-                                                             "AI: " +
-                                                             reponse_full},
-                            {"role": "user", "content": "Summarise following conversation in no more than 3 words, "
-                                                        "respond ONLY with the summary, use the same language as the "
-                                                        "system"}]
-        completion = openai.ChatCompletion.create(model='gpt-3.5-turbo', engine=settings.AZURE_DEPLOYMENT_NAME,
-                                                  messages=messages_summary, max_tokens=30, temperature=0)
-        conversation_id = conversations_collection.insert_one(
-            {"user": "local",
-             "date": datetime.datetime.utcnow(),
-             "name": completion["choices"][0]["message"]["content"],
-             "queries": [{"prompt": question, "response": reponse_full, "sources": source_log_docs}]}
-        ).inserted_id
-
-    # send data.type = "end" to indicate that the stream has ended as json
-    data = json.dumps({"type": "id", "id": str(conversation_id)})
-    yield f"data: {data}\n\n"
-    data = json.dumps({"type": "end"})
-    yield f"data: {data}\n\n"
-
-
-@app.route("/stream", methods=["POST"])
-def stream():
-    data = request.get_json()
-    # get parameter from url question
-    question = data["question"]
-    history = data["history"]
-    # history to json object from string
-    history = json.loads(history)
-    conversation_id = data["conversation_id"]
-
-    # check if active_docs is set
-
-    if not api_key_set:
-        api_key = data["api_key"]
-    else:
-        api_key = settings.API_KEY
-    if not embeddings_key_set:
-        embeddings_key = data["embeddings_key"]
-    else:
-        embeddings_key = settings.EMBEDDINGS_KEY
-    if "active_docs" in data:
-        vectorstore = get_vectorstore({"active_docs": data["active_docs"]})
-    else:
-        vectorstore = ""
-    docsearch = get_docsearch(vectorstore, embeddings_key)
-
-    # question = "Hi"
-    return Response(
-        complete_stream(question, docsearch,
-                        chat_history=history, api_key=api_key,
-                        conversation_id=conversation_id), mimetype="text/event-stream"
-    )
-
-
-def is_azure_configured():
-    return settings.OPENAI_API_BASE and settings.OPENAI_API_VERSION and settings.AZURE_DEPLOYMENT_NAME
-
-
-@app.route("/api/answer", methods=["POST"])
-def api_answer():
-    data = request.get_json()
-    question = data["question"]
-    history = data["history"]
-    if "conversation_id" not in data:
-        conversation_id = None
-    else:
-        conversation_id = data["conversation_id"]
-    print("-" * 5)
-    if not api_key_set:
-        api_key = data["api_key"]
-    else:
-        api_key = settings.API_KEY
-    if not embeddings_key_set:
-        embeddings_key = data["embeddings_key"]
-    else:
-        embeddings_key = settings.EMBEDDINGS_KEY
-
-    # use try and except  to check for exception
-    try:
-        # check if the vectorstore is set
-        vectorstore = get_vectorstore(data)
-        # loading the index and the store and the prompt template
-        # Note if you have used other embeddings than OpenAI, you need to change the embeddings
-        docsearch = get_docsearch(vectorstore, embeddings_key)
-
-        q_prompt = PromptTemplate(
-            input_variables=["context", "question"], template=template_quest, template_format="jinja2"
-        )
-        if settings.LLM_NAME == "openai_chat":
-            if is_azure_configured():
-                logger.debug("in Azure")
-                llm = AzureChatOpenAI(
-                    openai_api_key=api_key,
-                    openai_api_base=settings.OPENAI_API_BASE,
-                    openai_api_version=settings.OPENAI_API_VERSION,
-                    deployment_name=settings.AZURE_DEPLOYMENT_NAME,
-                )
-            else:
-                logger.debug("plain OpenAI")
-                llm = ChatOpenAI(openai_api_key=api_key, model_name=gpt_model)  # optional parameter: model_name="gpt-4"
-            messages_combine = [SystemMessagePromptTemplate.from_template(chat_combine_template)]
-            if history:
-                tokens_current_history = 0
-                # count tokens in history
-                history.reverse()
-                for i in history:
-                    if "prompt" in i and "response" in i:
-                        tokens_batch = llm.get_num_tokens(i["prompt"]) + llm.get_num_tokens(i["response"])
-                        if tokens_current_history + tokens_batch < settings.TOKENS_MAX_HISTORY:
-                            tokens_current_history += tokens_batch
-                            messages_combine.append(HumanMessagePromptTemplate.from_template(i["prompt"]))
-                            messages_combine.append(AIMessagePromptTemplate.from_template(i["response"]))
-            messages_combine.append(HumanMessagePromptTemplate.from_template("{question}"))
-            p_chat_combine = ChatPromptTemplate.from_messages(messages_combine)
-        elif settings.LLM_NAME == "openai":
-            llm = OpenAI(openai_api_key=api_key, temperature=0)
-        elif settings.SELF_HOSTED_MODEL:
-            llm = hf
-        elif settings.LLM_NAME == "cohere":
-            llm = Cohere(model="command-xlarge-nightly", cohere_api_key=api_key)
-        else:
-            raise ValueError("unknown LLM model")
-
-        if settings.LLM_NAME == "openai_chat":
-            question_generator = LLMChain(llm=llm, prompt=CONDENSE_QUESTION_PROMPT)
-            doc_chain = load_qa_chain(llm, chain_type="map_reduce", combine_prompt=p_chat_combine)
-            chain = ConversationalRetrievalChain(
-                retriever=docsearch.as_retriever(k=2),
-                question_generator=question_generator,
-                combine_docs_chain=doc_chain,
-            )
-            chat_history = []
-            # result = chain({"question": question, "chat_history": chat_history})
-            # generate async with async generate method
-            result = run_async_chain(chain, question, chat_history)
-        elif settings.SELF_HOSTED_MODEL:
-            question_generator = LLMChain(llm=llm, prompt=CONDENSE_QUESTION_PROMPT)
-            doc_chain = load_qa_chain(llm, chain_type="map_reduce", combine_prompt=p_chat_combine)
-            chain = ConversationalRetrievalChain(
-                retriever=docsearch.as_retriever(k=2),
-                question_generator=question_generator,
-                combine_docs_chain=doc_chain,
-            )
-            chat_history = []
-            # result = chain({"question": question, "chat_history": chat_history})
-            # generate async with async generate method
-            result = run_async_chain(chain, question, chat_history)
-
-        else:
-            qa_chain = load_qa_chain(
-                llm=llm, chain_type="map_reduce", combine_prompt=chat_combine_template, question_prompt=q_prompt
-            )
-            chain = VectorDBQA(combine_documents_chain=qa_chain, vectorstore=docsearch, k=3)
-            result = chain({"query": question})
-
-        print(result)
-
-        # some formatting for the frontend
-        if "result" in result:
-            result["answer"] = result["result"]
-        result["answer"] = result["answer"].replace("\\n", "\n")
-        try:
-            result["answer"] = result["answer"].split("SOURCES:")[0]
-        except Exception:
-            pass
-
-        sources = docsearch.similarity_search(question, k=2)
-        sources_doc = []
-        for doc in sources:
-            if doc.metadata:
-                sources_doc.append({'title': doc.metadata['title'], 'text': doc.page_content})
-            else:
-                sources_doc.append({'title': doc.page_content, 'text': doc.page_content})
-        result['sources'] = sources_doc
-
-        # generate conversationId
-        if conversation_id is not None:
-            conversations_collection.update_one(
-                {"_id": ObjectId(conversation_id)},
-                {"$push": {"queries": {"prompt": question,
-                                       "response": result["answer"], "sources": result['sources']}}},
-            )
-
-        else:
-            # create new conversation
-            # generate summary
-            messages_summary = [AIMessage(content="Summarise following conversation in no more than 3 " +
-                                                  "words, respond ONLY with the summary, use the same " +
-                                                  "language as the system \n\nUser: " + question + "\n\nAI: " +
-                                                  result["answer"]),
-                                HumanMessage(content="Summarise following conversation in no more than 3 words, " +
-                                                     "respond ONLY with the summary, use the same language as the " +
-                                                     "system")]
-
-
-            # completion = openai.ChatCompletion.create(model='gpt-3.5-turbo', engine=settings.AZURE_DEPLOYMENT_NAME,
-            #                                           messages=messages_summary, max_tokens=30, temperature=0)
-            completion = llm.predict_messages(messages_summary)
-            conversation_id = conversations_collection.insert_one(
-                {"user": "local",
-                 "date": datetime.datetime.utcnow(),
-                 "name": completion.content,
-                 "queries": [{"prompt": question, "response": result["answer"], "sources": result['sources']}]}
-            ).inserted_id
-
-        result["conversation_id"] = str(conversation_id)
-
-        # mock result
-        # result = {
-        #     "answer": "The answer is 42",
-        #     "sources": ["https://en.wikipedia.org/wiki/42_(number)", "https://en.wikipedia.org/wiki/42_(number)"]
-        # }
-        return result
-    except Exception as e:
-        # print whole traceback
-        traceback.print_exc()
-        print(str(e))
-        return bad_request(500, str(e))
-
-
-@app.route("/api/docs_check", methods=["POST"])
-def check_docs():
-    # check if docs exist in a vectorstore folder
-    data = request.get_json()
-    # split docs on / and take first part
-    if data["docs"].split("/")[0] == "local":
-        return {"status": "exists"}
-    vectorstore = "vectors/" + data["docs"]
-    base_path = "https://raw.githubusercontent.com/arc53/DocsHUB/main/"
-    if os.path.exists(vectorstore) or data["docs"] == "default":
-        return {"status": "exists"}
-    else:
-        r = requests.get(base_path + vectorstore + "index.faiss")
-
-        if r.status_code != 200:
-            return {"status": "null"}
-        else:
-            if not os.path.exists(vectorstore):
-                os.makedirs(vectorstore)
-            with open(vectorstore + "index.faiss", "wb") as f:
-                f.write(r.content)
-
-            # download the store
-            r = requests.get(base_path + vectorstore + "index.pkl")
-            with open(vectorstore + "index.pkl", "wb") as f:
-                f.write(r.content)
-
-        return {"status": "loaded"}
-
-
-@app.route("/api/feedback", methods=["POST"])
-def api_feedback():
-    data = request.get_json()
-    question = data["question"]
-    answer = data["answer"]
-    feedback = data["feedback"]
-
-    print("-" * 5)
-    print("Question: " + question)
-    print("Answer: " + answer)
-    print("Feedback: " + feedback)
-    print("-" * 5)
-    response = requests.post(
-        url="https://86x89umx77.execute-api.eu-west-2.amazonaws.com/docsgpt-feedback",
-        headers={
-            "Content-Type": "application/json; charset=utf-8",
-        },
-        data=json.dumps({"answer": answer, "question": question, "feedback": feedback}),
-    )
-    return {"status": http.client.responses.get(response.status_code, "ok")}
-
-
-@app.route("/api/combine", methods=["GET"])
-def combined_json():
-    user = "local"
-    """Provide json file with combined available indexes."""
-    # get json from https://d3dg1063dc54p9.cloudfront.net/combined.json
-
-    data = [
-        {
-            "name": "default",
-            "language": "default",
-            "version": "",
-            "description": "default",
-            "fullName": "default",
-            "date": "default",
-            "docLink": "default",
-            "model": settings.EMBEDDINGS_NAME,
-            "location": "local",
-        }
-    ]
-    # structure: name, language, version, description, fullName, date, docLink
-    # append data from vectors_collection
-    for index in vectors_collection.find({"user": user}):
-        data.append(
-            {
-                "name": index["name"],
-                "language": index["language"],
-                "version": "",
-                "description": index["name"],
-                "fullName": index["name"],
-                "date": index["date"],
-                "docLink": index["location"],
-                "model": settings.EMBEDDINGS_NAME,
-                "location": "local",
-            }
-        )
-
-    data_remote = requests.get("https://d3dg1063dc54p9.cloudfront.net/combined.json").json()
-    for index in data_remote:
-        index["location"] = "remote"
-        data.append(index)
-
-    return jsonify(data)
-
-
-@app.route("/api/upload", methods=["POST"])
-def upload_file():
-    """Upload a file to get vectorized and indexed."""
-    if "user" not in request.form:
-        return {"status": "no user"}
-    user = secure_filename(request.form["user"])
-    if "name" not in request.form:
-        return {"status": "no name"}
-    job_name = secure_filename(request.form["name"])
-    # check if the post request has the file part
-    if "file" not in request.files:
-        print("No file part")
-        return {"status": "no file"}
-    file = request.files["file"]
-    if file.filename == "":
-        return {"status": "no file name"}
-
-    if file:
-        filename = secure_filename(file.filename)
-        # save dir
-        save_dir = os.path.join(app.config["UPLOAD_FOLDER"], user, job_name)
-        # create dir if not exists
-        if not os.path.exists(save_dir):
-            os.makedirs(save_dir)
-
-        file.save(os.path.join(save_dir, filename))
-        task = ingest.delay("temp", [".rst", ".md", ".pdf", ".txt"], job_name, filename, user)
-        # task id
-        task_id = task.id
-        return {"status": "ok", "task_id": task_id}
-    else:
-        return {"status": "error"}
-
-
-@app.route("/api/task_status", methods=["GET"])
-def task_status():
-    """Get celery job status."""
-    task_id = request.args.get("task_id")
-    task = AsyncResult(task_id)
-    task_meta = task.info
-    return {"status": task.status, "result": task_meta}
-
-
-### Backgound task api
-@app.route("/api/upload_index", methods=["POST"])
-def upload_index_files():
-    """Upload two files(index.faiss, index.pkl) to the user's folder."""
-    if "user" not in request.form:
-        return {"status": "no user"}
-    user = secure_filename(request.form["user"])
-    if "name" not in request.form:
-        return {"status": "no name"}
-    job_name = secure_filename(request.form["name"])
-    if "file_faiss" not in request.files:
-        print("No file part")
-        return {"status": "no file"}
-    file_faiss = request.files["file_faiss"]
-    if file_faiss.filename == "":
-        return {"status": "no file name"}
-    if "file_pkl" not in request.files:
-        print("No file part")
-        return {"status": "no file"}
-    file_pkl = request.files["file_pkl"]
-    if file_pkl.filename == "":
-        return {"status": "no file name"}
-
-    # saves index files
-    save_dir = os.path.join("indexes", user, job_name)
-    if not os.path.exists(save_dir):
-        os.makedirs(save_dir)
-    file_faiss.save(os.path.join(save_dir, "index.faiss"))
-    file_pkl.save(os.path.join(save_dir, "index.pkl"))
-    # create entry in vectors_collection
-    vectors_collection.insert_one(
-        {
-            "user": user,
-            "name": job_name,
-            "language": job_name,
-            "location": save_dir,
-            "date": datetime.datetime.now().strftime("%d/%m/%Y %H:%M:%S"),
-            "model": settings.EMBEDDINGS_NAME,
-            "type": "local",
-        }
-    )
-    return {"status": "ok"}
-
-
-@app.route("/api/download", methods=["get"])
-def download_file():
-    user = secure_filename(request.args.get("user"))
-    job_name = secure_filename(request.args.get("name"))
-    filename = secure_filename(request.args.get("file"))
-    save_dir = os.path.join(app.config["UPLOAD_FOLDER"], user, job_name)
-    return send_from_directory(save_dir, filename, as_attachment=True)
-
-
-@app.route("/api/delete_old", methods=["get"])
-def delete_old():
-    """Delete old indexes."""
-    import shutil
-
-    path = request.args.get("path")
-    dirs = path.split("/")
-    dirs_clean = []
-    for i in range(1, len(dirs)):
-        dirs_clean.append(secure_filename(dirs[i]))
-    # check that path strats with indexes or vectors
-    if dirs[0] not in ["indexes", "vectors"]:
-        return {"status": "error"}
-    path_clean = "/".join(dirs)
-    vectors_collection.delete_one({"location": path})
-    try:
-        shutil.rmtree(path_clean)
-    except FileNotFoundError:
-        pass
-    return {"status": "ok"}
-
-
-@app.route("/api/get_conversations", methods=["get"])
-def get_conversations():
-    # provides a list of conversations
-    conversations = conversations_collection.find().sort("date", -1)
-    list_conversations = []
-    for conversation in conversations:
-        list_conversations.append({"id": str(conversation["_id"]), "name": conversation["name"]})
-
-    #list_conversations = [{"id": "default", "name": "default"}, {"id": "jeff", "name": "jeff"}]
-
-    return jsonify(list_conversations)
-
-@app.route("/api/get_single_conversation", methods=["get"])
-def get_single_conversation():
-    # provides data for a conversation
-    conversation_id = request.args.get("id")
-    conversation = conversations_collection.find_one({"_id": ObjectId(conversation_id)})
-    return jsonify(conversation['queries'])
-
-@app.route("/api/delete_conversation", methods=["POST"])
-def delete_conversation():
-    # deletes a conversation from the database
-    conversation_id = request.args.get("id")
-    # write to mongodb
-    conversations_collection.delete_one(
-        {
-            "_id": ObjectId(conversation_id),
-        }
-    )
-
-    return {"status": "ok"}
-
-
-# handling CORS
 @app.after_request
 def after_request(response):
     response.headers.add("Access-Control-Allow-Origin", "*")
     response.headers.add("Access-Control-Allow-Headers", "Content-Type,Authorization")
     response.headers.add("Access-Control-Allow-Methods", "GET,PUT,POST,DELETE,OPTIONS")
-    response.headers.add("Access-Control-Allow-Credentials", "true")
     return response
 
-
 if __name__ == "__main__":
     app.run(debug=True, port=7091)
+

application/celery.py

@@ -0,0 +1,9 @@
+from celery import Celery
+from application.core.settings import settings
+
+def make_celery(app_name=__name__):
+    celery = Celery(app_name, broker=settings.CELERY_BROKER_URL, backend=settings.CELERY_RESULT_BACKEND)
+    celery.conf.update(settings)
+    return celery
+
+celery = make_celery()

application/core/settings.py

@@ -1,26 +1,43 @@
 from pathlib import Path
+from typing import Optional
+import os
 
-from pydantic import BaseSettings
+from pydantic_settings import BaseSettings
+current_dir = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
 
 
 class Settings(BaseSettings):
-    LLM_NAME: str = "openai_chat"
-    EMBEDDINGS_NAME: str = "openai_text-embedding-ada-002"
+    LLM_NAME: str = "docsgpt"
+    EMBEDDINGS_NAME: str = "huggingface_sentence-transformers/all-mpnet-base-v2"
     CELERY_BROKER_URL: str = "redis://localhost:6379/0"
     CELERY_RESULT_BACKEND: str = "redis://localhost:6379/1"
     MONGO_URI: str = "mongodb://localhost:27017/docsgpt"
-    MODEL_PATH: str = "./models/gpt4all-model.bin"
+    MODEL_PATH: str = os.path.join(current_dir, "models/docsgpt-7b-f16.gguf")
     TOKENS_MAX_HISTORY: int = 150
-    SELF_HOSTED_MODEL: bool = False
+    UPLOAD_FOLDER: str = "inputs"
+    VECTOR_STORE: str = "faiss"  # "faiss" or "elasticsearch"
 
     API_URL: str = "http://localhost:7091"  # backend url for celery worker
 
-    API_KEY: str = None  # LLM api key
-    EMBEDDINGS_KEY: str = None  # api key for embeddings (if using openai, just copy API_KEY
-    OPENAI_API_BASE: str = None  # azure openai api base url
-    OPENAI_API_VERSION: str = None  # azure openai api version
-    AZURE_DEPLOYMENT_NAME: str = None  # azure deployment name for answering
-    AZURE_EMBEDDINGS_DEPLOYMENT_NAME: str = None  # azure deployment name for embeddings
+    API_KEY: Optional[str] = None  # LLM api key
+    EMBEDDINGS_KEY: Optional[str] = None  # api key for embeddings (if using openai, just copy API_KEY)
+    OPENAI_API_BASE: Optional[str] = None  # azure openai api base url
+    OPENAI_API_VERSION: Optional[str] = None  # azure openai api version
+    AZURE_DEPLOYMENT_NAME: Optional[str] = None  # azure deployment name for answering
+    AZURE_EMBEDDINGS_DEPLOYMENT_NAME: Optional[str] = None  # azure deployment name for embeddings
+
+    # elasticsearch
+    ELASTIC_CLOUD_ID: Optional[str] = None # cloud id for elasticsearch
+    ELASTIC_USERNAME: Optional[str] = None # username for elasticsearch
+    ELASTIC_PASSWORD: Optional[str] = None # password for elasticsearch
+    ELASTIC_URL: Optional[str] = None # url for elasticsearch
+    ELASTIC_INDEX: Optional[str] = "docsgpt" # index name for elasticsearch
+
+    # SageMaker config
+    SAGEMAKER_ENDPOINT: Optional[str] = None # SageMaker endpoint name
+    SAGEMAKER_REGION: Optional[str] = None # SageMaker region name
+    SAGEMAKER_ACCESS_KEY: Optional[str] = None # SageMaker access key
+    SAGEMAKER_SECRET_KEY: Optional[str] = None # SageMaker secret key
 
 
 path = Path(__file__).parent.parent.absolute()

application/llm/anthropic.py

@@ -0,0 +1,40 @@
+from application.llm.base import BaseLLM
+from application.core.settings import settings
+
+class AnthropicLLM(BaseLLM):
+
+    def __init__(self, api_key=None):
+        from anthropic import Anthropic, HUMAN_PROMPT, AI_PROMPT
+        self.api_key = api_key or settings.ANTHROPIC_API_KEY  # If not provided, use a default from settings
+        self.anthropic = Anthropic(api_key=self.api_key)
+        self.HUMAN_PROMPT = HUMAN_PROMPT
+        self.AI_PROMPT = AI_PROMPT
+
+    def gen(self, model, messages, engine=None, max_tokens=300, stream=False, **kwargs):
+        context = messages[0]['content']
+        user_question = messages[-1]['content']
+        prompt = f"### Context \n {context} \n ### Question \n {user_question}"
+        if stream:
+            return self.gen_stream(model, prompt, max_tokens, **kwargs)
+
+        completion = self.anthropic.completions.create(
+            model=model,
+            max_tokens_to_sample=max_tokens,
+            stream=stream,
+            prompt=f"{self.HUMAN_PROMPT} {prompt}{self.AI_PROMPT}",
+        )
+        return completion.completion
+
+    def gen_stream(self, model, messages, engine=None, max_tokens=300, **kwargs):
+        context = messages[0]['content']
+        user_question = messages[-1]['content']
+        prompt = f"### Context \n {context} \n ### Question \n {user_question}"
+        stream_response = self.anthropic.completions.create(
+            model=model,
+            prompt=f"{self.HUMAN_PROMPT} {prompt}{self.AI_PROMPT}",
+            max_tokens_to_sample=max_tokens,
+            stream=True,
+        )
+
+        for completion in stream_response:
+            yield completion.completion

application/llm/base.py

@@ -0,0 +1,14 @@
+from abc import ABC, abstractmethod
+
+
+class BaseLLM(ABC):
+    def __init__(self):
+        pass
+
+    @abstractmethod
+    def gen(self, *args, **kwargs):
+        pass
+
+    @abstractmethod
+    def gen_stream(self, *args, **kwargs):
+        pass

application/llm/docsgpt_provider.py

@@ -0,0 +1,49 @@
+from application.llm.base import BaseLLM
+import json
+import requests
+
+class DocsGPTAPILLM(BaseLLM):
+
+    def __init__(self, *args, **kwargs):
+        self.endpoint =  "https://llm.docsgpt.co.uk"
+
+
+    def gen(self, model, engine, messages, stream=False, **kwargs):
+        context = messages[0]['content']
+        user_question = messages[-1]['content']
+        prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
+
+        response = requests.post(
+            f"{self.endpoint}/answer",
+            json={
+                "prompt": prompt,
+                "max_new_tokens": 30
+            }
+        )
+        response_clean = response.json()['a'].split("###")[0]
+
+        return response_clean
+
+    def gen_stream(self, model, engine, messages, stream=True, **kwargs):
+        context = messages[0]['content']
+        user_question = messages[-1]['content']
+        prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
+
+        # send prompt to endpoint /stream
+        response = requests.post(
+            f"{self.endpoint}/stream",
+            json={
+                "prompt": prompt,
+                "max_new_tokens": 256
+            },
+            stream=True
+        )
+    
+        for line in response.iter_lines():
+            if line:
+                #data = json.loads(line)
+                data_str = line.decode('utf-8')
+                if data_str.startswith("data: "):
+                    data = json.loads(data_str[6:])
+                    yield data['a']
+                    

application/llm/huggingface.py

@@ -0,0 +1,44 @@
+from application.llm.base import BaseLLM
+
+class HuggingFaceLLM(BaseLLM):
+
+    def __init__(self, api_key, llm_name='Arc53/DocsGPT-7B',q=False):
+        global hf
+        
+        from langchain.llms import HuggingFacePipeline
+        if q:
+            import torch
+            from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline, BitsAndBytesConfig
+            tokenizer = AutoTokenizer.from_pretrained(llm_name)
+            bnb_config = BitsAndBytesConfig(
+                            load_in_4bit=True,
+                            bnb_4bit_use_double_quant=True,
+                            bnb_4bit_quant_type="nf4",
+                            bnb_4bit_compute_dtype=torch.bfloat16
+                        )
+            model = AutoModelForCausalLM.from_pretrained(llm_name,quantization_config=bnb_config)
+        else:
+            from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
+            tokenizer = AutoTokenizer.from_pretrained(llm_name)
+            model = AutoModelForCausalLM.from_pretrained(llm_name)
+        
+        pipe = pipeline(
+            "text-generation", model=model,
+            tokenizer=tokenizer, max_new_tokens=2000,
+            device_map="auto", eos_token_id=tokenizer.eos_token_id
+        )
+        hf = HuggingFacePipeline(pipeline=pipe)
+
+    def gen(self, model, engine, messages, stream=False, **kwargs):
+        context = messages[0]['content']
+        user_question = messages[-1]['content']
+        prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
+
+        result = hf(prompt)
+
+        return result.content
+
+    def gen_stream(self, model, engine, messages, stream=True, **kwargs):
+
+        raise NotImplementedError("HuggingFaceLLM Streaming is not implemented yet.")
+

application/llm/llama_cpp.py

@@ -0,0 +1,39 @@
+from application.llm.base import BaseLLM
+from application.core.settings import settings
+
+class LlamaCpp(BaseLLM):
+
+    def __init__(self, api_key, llm_name=settings.MODEL_PATH, **kwargs):
+        global llama
+        try:
+            from llama_cpp import Llama
+        except ImportError:
+            raise ImportError("Please install llama_cpp using pip install llama-cpp-python")
+
+        llama = Llama(model_path=llm_name, n_ctx=2048)
+
+    def gen(self, model, engine, messages, stream=False, **kwargs):
+        context = messages[0]['content']
+        user_question = messages[-1]['content']
+        prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
+
+        result = llama(prompt, max_tokens=150, echo=False)
+
+        # import sys
+        # print(result['choices'][0]['text'].split('### Answer \n')[-1], file=sys.stderr)
+        
+        return result['choices'][0]['text'].split('### Answer \n')[-1]
+
+    def gen_stream(self, model, engine, messages, stream=True, **kwargs):
+        context = messages[0]['content']
+        user_question = messages[-1]['content']
+        prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
+
+        result = llama(prompt, max_tokens=150, echo=False, stream=stream)
+
+        # import sys
+        # print(list(result), file=sys.stderr)
+
+        for item in result:
+            for choice in item['choices']:
+                yield choice['text']

application/llm/llm_creator.py

@@ -0,0 +1,26 @@
+from application.llm.openai import OpenAILLM, AzureOpenAILLM
+from application.llm.sagemaker import SagemakerAPILLM
+from application.llm.huggingface import HuggingFaceLLM
+from application.llm.llama_cpp import LlamaCpp
+from application.llm.anthropic import AnthropicLLM
+from application.llm.docsgpt_provider import DocsGPTAPILLM
+
+
+
+class LLMCreator:
+    llms = {
+        'openai': OpenAILLM,
+        'azure_openai': AzureOpenAILLM,
+        'sagemaker': SagemakerAPILLM,
+        'huggingface': HuggingFaceLLM,
+        'llama.cpp': LlamaCpp,
+        'anthropic': AnthropicLLM,
+        'docsgpt': DocsGPTAPILLM
+    }
+
+    @classmethod
+    def create_llm(cls, type, *args, **kwargs):
+        llm_class = cls.llms.get(type.lower())
+        if not llm_class:
+            raise ValueError(f"No LLM class found for type {type}")
+        return llm_class(*args, **kwargs)

application/llm/openai.py

@@ -0,0 +1,60 @@
+from application.llm.base import BaseLLM
+from application.core.settings import settings
+
+class OpenAILLM(BaseLLM):
+
+    def __init__(self, api_key):
+        global openai
+        from openai import OpenAI
+        
+        self.client = OpenAI(
+                api_key=api_key, 
+            )
+        self.api_key = api_key
+
+    def _get_openai(self):
+        # Import openai when needed
+        import openai
+        
+        return openai
+
+    def gen(self, model, engine, messages, stream=False, **kwargs):
+        response = self.client.chat.completions.create(model=model,
+            messages=messages,
+            stream=stream,
+            **kwargs)
+
+        return response.choices[0].message.content
+
+    def gen_stream(self, model, engine, messages, stream=True, **kwargs):
+        response = self.client.chat.completions.create(model=model,
+            messages=messages,
+            stream=stream,
+            **kwargs)
+
+        for line in response:
+            # import sys
+            # print(line.choices[0].delta.content, file=sys.stderr)
+            if line.choices[0].delta.content is not None:
+                yield line.choices[0].delta.content
+
+
+class AzureOpenAILLM(OpenAILLM):
+
+    def __init__(self, openai_api_key, openai_api_base, openai_api_version, deployment_name):
+        super().__init__(openai_api_key)
+        self.api_base = settings.OPENAI_API_BASE,
+        self.api_version = settings.OPENAI_API_VERSION,
+        self.deployment_name = settings.AZURE_DEPLOYMENT_NAME,
+        from openai import AzureOpenAI
+        self.client = AzureOpenAI(
+            api_key=openai_api_key,  
+            api_version=settings.OPENAI_API_VERSION,
+            api_base=settings.OPENAI_API_BASE,
+            deployment_name=settings.AZURE_DEPLOYMENT_NAME,
+        )
+
+    def _get_openai(self):
+        openai = super()._get_openai()
+
+        return openai

application/llm/sagemaker.py

@@ -0,0 +1,139 @@
+from application.llm.base import BaseLLM
+from application.core.settings import settings
+import json
+import io
+
+
+
+class LineIterator:
+    """
+    A helper class for parsing the byte stream input. 
+    
+    The output of the model will be in the following format:
+    ```
+    b'{"outputs": [" a"]}\n'
+    b'{"outputs": [" challenging"]}\n'
+    b'{"outputs": [" problem"]}\n'
+    ...
+    ```
+    
+    While usually each PayloadPart event from the event stream will contain a byte array 
+    with a full json, this is not guaranteed and some of the json objects may be split across
+    PayloadPart events. For example:
+    ```
+    {'PayloadPart': {'Bytes': b'{"outputs": '}}
+    {'PayloadPart': {'Bytes': b'[" problem"]}\n'}}
+    ```
+    
+    This class accounts for this by concatenating bytes written via the 'write' function
+    and then exposing a method which will return lines (ending with a '\n' character) within
+    the buffer via the 'scan_lines' function. It maintains the position of the last read 
+    position to ensure that previous bytes are not exposed again. 
+    """
+    
+    def __init__(self, stream):
+        self.byte_iterator = iter(stream)
+        self.buffer = io.BytesIO()
+        self.read_pos = 0
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        while True:
+            self.buffer.seek(self.read_pos)
+            line = self.buffer.readline()
+            if line and line[-1] == ord('\n'):
+                self.read_pos += len(line)
+                return line[:-1]
+            try:
+                chunk = next(self.byte_iterator)
+            except StopIteration:
+                if self.read_pos < self.buffer.getbuffer().nbytes:
+                    continue
+                raise
+            if 'PayloadPart' not in chunk:
+                print('Unknown event type:' + chunk)
+                continue
+            self.buffer.seek(0, io.SEEK_END)
+            self.buffer.write(chunk['PayloadPart']['Bytes'])
+
+class SagemakerAPILLM(BaseLLM):
+
+    def __init__(self, *args, **kwargs):
+        import boto3
+        runtime = boto3.client(
+            'runtime.sagemaker',
+            aws_access_key_id='xxx',
+            aws_secret_access_key='xxx',
+            region_name='us-west-2'
+        )
+
+        
+        self.endpoint =  settings.SAGEMAKER_ENDPOINT
+        self.runtime = runtime
+
+
+    def gen(self, model, engine, messages, stream=False, **kwargs):
+        context = messages[0]['content']
+        user_question = messages[-1]['content']
+        prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
+    
+
+        # Construct payload for endpoint
+        payload = {
+            "inputs": prompt,
+            "stream": False,
+            "parameters": {
+                "do_sample": True,
+                "temperature": 0.1,
+                "max_new_tokens": 30,
+                "repetition_penalty": 1.03,
+                "stop": ["</s>", "###"]
+            }
+        }
+        body_bytes = json.dumps(payload).encode('utf-8')
+
+        # Invoke the endpoint
+        response = self.runtime.invoke_endpoint(EndpointName=self.endpoint,
+                                        ContentType='application/json',
+                                        Body=body_bytes)
+        result = json.loads(response['Body'].read().decode())
+        import sys
+        print(result[0]['generated_text'], file=sys.stderr)
+        return result[0]['generated_text'][len(prompt):]
+
+    def gen_stream(self, model, engine, messages, stream=True, **kwargs):
+        context = messages[0]['content']
+        user_question = messages[-1]['content']
+        prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
+    
+
+        # Construct payload for endpoint
+        payload = {
+            "inputs": prompt,
+            "stream": True,
+            "parameters": {
+                "do_sample": True,
+                "temperature": 0.1,
+                "max_new_tokens": 512,
+                "repetition_penalty": 1.03,
+                "stop": ["</s>", "###"]
+            }
+        }
+        body_bytes = json.dumps(payload).encode('utf-8')
+
+        # Invoke the endpoint
+        response = self.runtime.invoke_endpoint_with_response_stream(EndpointName=self.endpoint,
+                                        ContentType='application/json',
+                                        Body=body_bytes)
+        #result = json.loads(response['Body'].read().decode())
+        event_stream = response['Body']
+        start_json = b'{'
+        for line in LineIterator(event_stream):
+            if line != b'' and start_json in line:
+                #print(line)
+                data = json.loads(line[line.find(start_json):].decode('utf-8'))
+                if data['token']['text'] not in ["</s>", "###"]:
+                    print(data['token']['text'],end='')
+                    yield data['token']['text']

application/package.json

@@ -1,5 +0,0 @@
-{
-  "devDependencies": {
-    "tailwindcss": "^3.2.4"
-  }
-}

application/parser/file/bulk.py

@@ -62,7 +62,6 @@ class SimpleDirectoryReader(BaseReader):
             file_extractor: Optional[Dict[str, BaseParser]] = None,
             num_files_limit: Optional[int] = None,
             file_metadata: Optional[Callable[[str], Dict]] = None,
-            chunk_size_max: int = 2048,
     ) -> None:
         """Initialize with parameters."""
         super().__init__()

application/parser/file/html_parser.py

@@ -57,7 +57,7 @@ class HTMLParser(BaseParser):
         title_indexes = [i for i, isd_el in enumerate(isd) if isd_el['type'] == 'Title']
 
         # Creating 'Chunks' - List of lists of strings 
-        # each list starting with with isd_el['type'] = 'Title' and all the data till the next 'Title'
+        # each list starting with isd_el['type'] = 'Title' and all the data till the next 'Title'
         # Each Chunk can be thought of as an individual set of data, which can be sent to the model
         # Where Each Title is grouped together with the data under it
 
@@ -69,10 +69,10 @@ class HTMLParser(BaseParser):
                 Chunks.append([])
             Chunks[-1].append(isd_el['text'])
 
-        # Removing all the chunks with sum of lenth of all the strings in the chunk < 25
+        # Removing all the chunks with sum of length of all the strings in the chunk < 25
         # TODO: This value can be an user defined variable
         for chunk in Chunks:
-            # sum of lenth of all the strings in the chunk
+            # sum of length of all the strings in the chunk
             sum = 0
             sum += len(str(chunk))
             if sum < 25:

application/parser/file/openapi3_parser.py

@@ -0,0 +1,51 @@
+from urllib.parse import urlparse
+
+from openapi_parser import parse
+
+try:
+    from application.parser.file.base_parser import BaseParser
+except ModuleNotFoundError:
+    from base_parser import BaseParser
+
+
+class OpenAPI3Parser(BaseParser):
+    def init_parser(self) -> None:
+        return super().init_parser()
+
+    def get_base_urls(self, urls):
+        base_urls = []
+        for i in urls:
+            parsed_url = urlparse(i)
+            base_url = parsed_url.scheme + "://" + parsed_url.netloc
+            if base_url not in base_urls:
+                base_urls.append(base_url)
+        return base_urls
+
+    def get_info_from_paths(self, path):
+        info = ""
+        if path.operations:
+            for operation in path.operations:
+                info += (
+                    f"\n{operation.method.value}="
+                    f"{operation.responses[0].description}"
+                )
+        return info
+
+    def parse_file(self, file_path):
+        data = parse(file_path)
+        results = ""
+        base_urls = self.get_base_urls(link.url for link in data.servers)
+        base_urls = ",".join([base_url for base_url in base_urls])
+        results += f"Base URL:{base_urls}\n"
+        i = 1
+        for path in data.paths:
+            info = self.get_info_from_paths(path)
+            results += (
+                f"Path{i}: {path.url}\n"
+                f"description: {path.description}\n"
+                f"parameters: {path.parameters}\nmethods: {info}\n"
+            )
+            i += 1
+        with open("results.txt", "w") as f:
+            f.write(results)
+        return results

application/parser/file/rst_parser.py

@@ -27,7 +27,7 @@ class RstParser(BaseParser):
             remove_interpreters: bool = True,
             remove_directives: bool = True,
             remove_whitespaces_excess: bool = True,
-            # Be carefull with remove_characters_excess, might cause data loss
+            # Be careful with remove_characters_excess, might cause data loss
             remove_characters_excess: bool = True,
             **kwargs: Any,
     ) -> None:

application/parser/open_ai_func.py

@@ -1,14 +1,14 @@
 import os
 
 import tiktoken
-from langchain.embeddings import OpenAIEmbeddings
-from langchain.vectorstores import FAISS
+from application.vectorstore.vector_creator import VectorCreator
+from application.core.settings import settings
 from retry import retry
 
 
-# from langchain.embeddings import HuggingFaceEmbeddings
-# from langchain.embeddings import HuggingFaceInstructEmbeddings
-# from langchain.embeddings import CohereEmbeddings
+# from langchain_community.embeddings import HuggingFaceEmbeddings
+# from langchain_community.embeddings import HuggingFaceInstructEmbeddings
+# from langchain_community.embeddings import CohereEmbeddings
 
 
 def num_tokens_from_string(string: str, encoding_name: str) -> int:
@@ -33,12 +33,23 @@ def call_openai_api(docs, folder_name, task_status):
         os.makedirs(f"{folder_name}")
 
     from tqdm import tqdm
-    docs_test = [docs[0]]
-    docs.pop(0)
     c1 = 0
+    if settings.VECTOR_STORE == "faiss":
+        docs_init = [docs[0]]
+        docs.pop(0)
 
-    store = FAISS.from_documents(docs_test, OpenAIEmbeddings(openai_api_key=os.getenv("EMBEDDINGS_KEY")))
-
+        store = VectorCreator.create_vectorstore(
+            settings.VECTOR_STORE,
+            docs_init = docs_init,
+            path=f"{folder_name}",
+            embeddings_key=os.getenv("EMBEDDINGS_KEY")
+        )
+    else:
+        store = VectorCreator.create_vectorstore(
+            settings.VECTOR_STORE,
+            path=f"{folder_name}",
+            embeddings_key=os.getenv("EMBEDDINGS_KEY")
+        )
     # Uncomment for MPNet embeddings
     # model_name = "sentence-transformers/all-mpnet-base-v2"
     # hf = HuggingFaceEmbeddings(model_name=model_name)
@@ -57,7 +68,8 @@ def call_openai_api(docs, folder_name, task_status):
             store.save_local(f"{folder_name}")
             break
         c1 += 1
-    store.save_local(f"{folder_name}")
+    if settings.VECTOR_STORE == "faiss":
+        store.save_local(f"{folder_name}")
 
 
 def get_user_permission(docs, folder_name):

application/prompts/chat_combine_default.txt

@@ -0,0 +1,9 @@
+You are a helpful AI assistant, DocsGPT, specializing in document assistance, designed to offer detailed and informative responses. 
+If appropriate, your answers can include code examples, formatted as follows:
+```(language)
+(code)
+```
+You effectively utilize chat history, ensuring relevant and tailored responses. 
+If a question doesn't align with your context, you provide friendly and helpful replies.
+----------------
+{summaries}

application/prompts/chat_combine_strict.txt

@@ -0,0 +1,13 @@
+You are an AI Assistant, DocsGPT, adept at offering document assistance. 
+Your expertise lies in providing answer on top of provided context. 
+You can leverage the chat history if needed.
+Answer the question based on the context below. 
+Keep the answer concise. Respond "Irrelevant context" if not sure about the answer.
+If question is not related to the context, respond "Irrelevant context".
+When using code examples, use the following format:
+```(language)
+(code)
+```
+ ----------------
+ Context:
+ {summaries}

application/prompts/combine_prompt.txt

@@ -1,25 +0,0 @@
-You are a DocsGPT, friendly and helpful AI assistant by Arc53 that provides help with documents. You give thorough answers with code examples if possible.
-
-QUESTION: How to merge tables in pandas?
-=========
-Content: pandas provides various facilities for easily combining together Series or DataFrame with various kinds of set logic for the indexes and relational algebra functionality in the case of join / merge-type operations.
-Source: 28-pl
-Content: pandas provides a single function, merge(), as the entry point for all standard database join operations between DataFrame or named Series objects: \n\npandas.merge(left, right, how='inner', on=None, left_on=None, right_on=None, left_index=False, right_index=False, sort=False, suffixes=('_x', '_y'), copy=True, indicator=False, validate=None)
-Source: 30-pl
-=========
-FINAL ANSWER: To merge two tables in pandas, you can use the pd.merge() function. The basic syntax is: \n\npd.merge(left, right, on, how) \n\nwhere left and right are the two tables to merge, on is the column to merge on, and how is the type of merge to perform. \n\nFor example, to merge the two tables df1 and df2 on the column 'id', you can use: \n\npd.merge(df1, df2, on='id', how='inner')
-SOURCES: 28-pl 30-pl
-
-QUESTION: How are you?
-=========
-CONTENT:
-SOURCE:
-=========
-FINAL ANSWER: I am fine, thank you. How are you?
-SOURCES:
-
-QUESTION: {{ question }}
-=========
-{{ summaries }}
-=========
-FINAL ANSWER:

application/prompts/combine_prompt_hist.txt

@@ -1,33 +0,0 @@
-You are a DocsGPT, friendly and helpful AI assistant by Arc53 that provides help with documents. You give thorough answers with code examples if possible.
-
-QUESTION: How to merge tables in pandas?
-=========
-Content: pandas provides various facilities for easily combining together Series or DataFrame with various kinds of set logic for the indexes and relational algebra functionality in the case of join / merge-type operations.
-Source: 28-pl
-Content: pandas provides a single function, merge(), as the entry point for all standard database join operations between DataFrame or named Series objects: \n\npandas.merge(left, right, how='inner', on=None, left_on=None, right_on=None, left_index=False, right_index=False, sort=False, suffixes=('_x', '_y'), copy=True, indicator=False, validate=None)
-Source: 30-pl
-=========
-FINAL ANSWER: To merge two tables in pandas, you can use the pd.merge() function. The basic syntax is: \n\npd.merge(left, right, on, how) \n\nwhere left and right are the two tables to merge, on is the column to merge on, and how is the type of merge to perform. \n\nFor example, to merge the two tables df1 and df2 on the column 'id', you can use: \n\npd.merge(df1, df2, on='id', how='inner')
-SOURCES: 28-pl 30-pl
-
-QUESTION: How are you?
-=========
-CONTENT:
-SOURCE:
-=========
-FINAL ANSWER: I am fine, thank you. How are you?
-SOURCES:
-
-QUESTION: {{ historyquestion }}
-=========
-CONTENT:
-SOURCE:
-=========
-FINAL ANSWER: {{ historyanswer }}
-SOURCES:
-
-QUESTION: {{ question }}
-=========
-{{ summaries }}
-=========
-FINAL ANSWER:

application/prompts/question_prompt.txt

@@ -1,4 +0,0 @@
-Use the following portion of a long document to see if any of the text is relevant to answer the question.
-{{ context }}
-Question: {{ question }}
-Provide all relevant text to the question verbatim. Summarize if needed. If nothing relevant return "-".

application/requirements.txt

@@ -1,104 +1,33 @@
-aiodns==3.0.0
-aiohttp==3.8.5
-aiohttp-retry==2.8.3
-aiosignal==1.3.1
-aleph-alpha-client==2.16.1
-amqp==5.1.1
-async-timeout==4.0.2
-attrs==22.2.0
-billiard==3.6.4.0
-blobfile==2.0.1
-boto3==1.28.20
-celery==5.2.7
-cffi==1.15.1
-charset-normalizer==3.1.0
-click==8.1.3
-click-didyoumean==0.3.0
-click-plugins==1.1.1
-click-repl==0.2.0
-cryptography==41.0.3
-dataclasses-json==0.5.7
-decorator==5.1.1
-dill==0.3.6
-dnspython==2.3.0
-ecdsa==0.18.0
-entrypoints==0.4
-faiss-cpu==1.7.3
-filelock==3.9.0
-Flask==2.2.5
-Flask-Cors==3.0.10
-frozenlist==1.3.3
-geojson==2.5.0
-gunicorn==20.1.0
-greenlet==2.0.2
-gpt4all==0.1.7
-huggingface-hub==0.15.1
-humbug==0.3.2
-idna==3.4
-itsdangerous==2.1.2
-Jinja2==3.1.2
-jmespath==1.0.1
-joblib==1.2.0
-kombu==5.2.4
-langchain==0.0.263
-loguru==0.6.0
-lxml==4.9.2
-MarkupSafe==2.1.2
-marshmallow==3.19.0
-marshmallow-enum==1.5.1
-mpmath==1.3.0
-multidict==6.0.4
-multiprocess==0.70.14
-mypy-extensions==1.0.0
-networkx==3.0
-npx
+anthropic==0.12.0
+boto3==1.34.6
+celery==5.3.6
+dataclasses_json==0.6.3
+docx2txt==0.8
+EbookLib==0.18
+elasticsearch==8.12.0
+escodegen==1.0.11
+esprima==4.0.1
+faiss-cpu==1.7.4
+Flask==3.0.1
+gunicorn==21.2.0
+html2text==2020.1.16
+javalang==0.13.0
+langchain==0.1.4
+langchain-openai==0.0.5
 nltk==3.8.1
-numcodecs==0.11.0
-numpy==1.24.2
-openai==0.27.8
-packaging==23.0
-pathos==0.3.0
-Pillow==9.4.0
-pox==0.3.2
-ppft==1.7.6.6
-prompt-toolkit==3.0.38
-py==1.11.0
-pyasn1==0.4.8
-pycares==4.3.0
-pycparser==2.21
-pycryptodomex==3.17
-pydantic==1.10.5
-PyJWT==2.6.0
-pymongo==4.3.3
-pyowm==3.3.0
+openapi3_parser==1.1.16
+pandas==2.2.0
+pydantic_settings==2.1.0
+pymongo==4.6.1
 PyPDF2==3.0.1
-PySocks==1.7.1
-pytest
-python-dateutil==2.8.2
-python-dotenv==1.0.0
-python-jose==3.3.0
-pytz==2022.7.1
-PyYAML==6.0
-redis==4.5.4
-regex==2022.10.31
-requests==2.31.0
+python-dotenv==1.0.1
+redis==5.0.1
+Requests==2.31.0
 retry==0.9.2
-rsa==4.9
-scikit-learn==1.2.2
-scipy==1.10.1
-sentencepiece
-six==1.16.0
-SQLAlchemy==1.4.46
-sympy==1.11.1
-tenacity==8.2.2
-threadpoolctl==3.1.0
-tiktoken
-tqdm==4.65.0
-transformers==4.30.0
-typer==0.7.0
-typing-inspect==0.8.0
-typing_extensions==4.5.0
-urllib3==1.26.14
-vine==5.0.0
-wcwidth==0.2.6
-yarl==1.8.2
+sentence-transformers
+tiktoken==0.5.2
+torch==2.1.2
+tqdm==4.66.1
+transformers==4.36.2
+unstructured==0.12.2
+Werkzeug==3.0.1

application/static/favicon/site.webmanifest

@@ -1 +0,0 @@
-{"name":"","short_name":"","icons":[{"src":"/android-chrome-192x192.png","sizes":"192x192","type":"image/png"},{"src":"/android-chrome-512x512.png","sizes":"512x512","type":"image/png"}],"theme_color":"#ffffff","background_color":"#ffffff","display":"standalone"}

application/static/src/authapi.js

@@ -1,19 +0,0 @@
-function resetApiKey() {
-  const modal = document.getElementById("modal");
-  modal.classList.toggle("hidden");
-}
-
-const apiKeyForm = document.getElementById("api-key-form");
-if (apiKeyForm) {
-  apiKeyForm.addEventListener("submit", function(event) {
-    event.preventDefault();
-
-    const apiKeyInput = document.getElementById("api-key-input");
-    const apiKey = apiKeyInput.value;
-
-    localStorage.setItem("apiKey", apiKey);
-
-    apiKeyInput.value = "";
-    modal.classList.toggle("hidden");
-  });
-}

application/static/src/chat.js

@@ -1,76 +0,0 @@
-var form = document.getElementById('message-form');
-var errorModal = document.getElementById('error-alert')
-document.getElementById('close').addEventListener('click',()=>{
-    errorModal.classList.toggle('hidden')
-})
-
-
-function submitForm(event){
-    event.preventDefault()
-    var message = document.getElementById("message-input").value;
-    console.log(message.length)
-    if(message.length === 0){
-        return
-    }
-    msg_html = '<div class="bg-blue-500 text-white p-2 rounded-lg mb-2 self-end"><p class="text-sm">'
-    msg_html += message
-    msg_html += '</p></div>'
-    document.getElementById("messages").innerHTML += msg_html;
-    let chatWindow = document.getElementById("messages-container");
-    chatWindow.scrollTop = chatWindow.scrollHeight;
-    document.getElementById("message-input").value = "";
-    document.getElementById("button-submit").innerHTML = '<i class="fa fa-circle-o-notch fa-spin"></i> Thinking...';
-    document.getElementById("button-submit").disabled = true;
-    if (localStorage.getItem('activeDocs') == null) {
-        localStorage.setItem('activeDocs', 'default')
-    }
-
-    
-    fetch('/api/answer', {
-        method: 'POST',
-        headers: {
-            'Content-Type': 'application/json',
-        },
-
-        body: JSON.stringify({question: message,
-            api_key: localStorage.getItem('apiKey'),
-            embeddings_key: localStorage.getItem('apiKey'),
-            history: localStorage.getItem('chatHistory'),
-            active_docs: localStorage.getItem('activeDocs')}),
-    }).then((response)=> response.json())
-    .then(data => {
-            console.log('Success:', data);
-            if(data.error){
-            document.getElementById('text-error').textContent = `Error : ${JSON.stringify(data.message)}`
-            errorModal.classList.toggle('hidden')
-            }
-            if(data.answer){
-            msg_html = '<div class="bg-indigo-500 text-white p-2 rounded-lg mb-2 self-start"><code class="text-sm">'
-            data.answer = data.answer.replace(/\n/g, "<br>");
-            msg_html += data.answer
-            msg_html += '</code></div>'
-            document.getElementById("messages").innerHTML += msg_html;
-            let chatWindow = document.getElementById("messages-container");
-            chatWindow.scrollTop = chatWindow.scrollHeight;
-            }
-            document.getElementById("button-submit").innerHTML = 'Send';
-            document.getElementById("button-submit").disabled = false;
-            let chatHistory = [message, data.answer || ''];
-            localStorage.setItem('chatHistory', JSON.stringify(chatHistory));
-
-            
-
-
-        })
-        .catch((error) => {
-            console.error('Error:', error);
-            // console.log(error);
-            // document.getElementById("button-submit").innerHTML = 'Send';
-            // document.getElementById("button-submit").disabled = false;
-
-        });
-}
-
-//window.addEventListener('submit',submitForm)
-// rewrite using id = button-submit
-document.getElementById("button-submit").addEventListener('click',submitForm)

application/static/src/choiceChange.js

@@ -1,15 +0,0 @@
-document.getElementById("select-docs").addEventListener("change", function() {
-localStorage.setItem('activeDocs', this.value)
-     fetch('/api/docs_check', {
-         method: 'POST',
-         headers: {
-             'Content-Type': 'application/json',
-         },
-         body: JSON.stringify({docs: this.value}),
-     }).then(response => response.json()).then(
-            data => {
-                console.log('Success:', data);
-            }
-     )
-});
-

application/static/src/input.css

@@ -1,37 +0,0 @@
-@tailwind base;
-@tailwind components;
-@tailwind utilities;
-
-
-
-
-@media screen and (max-width: 1024px) {
-  .text-lg {
-    font-size: 3.125rem;
-    margin: 2rem;
-    line-height: inherit;
-  }
-  .text-sm {
-    font-size: 2.5rem;
-    margin: 1.5rem;
-    line-height: inherit;
-  }
-
-}
-
-
-.loader {
-  border: 16px solid #f3f3f3; /* Light grey */
-  border-top: 16px solid #3498db; /* Blue */
-  border-radius: 50%;
-  width: 120px;
-  height: 120px;
-  animation: spin 2s linear infinite;
-}
-
-@keyframes spin {
-  0% { transform: rotate(0deg); }
-  100% { transform: rotate(360deg); }
-}
-
-

application/templates/index.html

@@ -1,215 +0,0 @@
-<!DOCTYPE html>
-<html>
-  <head>
-    <title>DocsGPT 🦖 Preview</title>
-    <link href="{{url_for('static',filename='dist/css/output.css')}}" rel="stylesheet">
-      <link rel="favicon" href="{{ url_for('static', filename='favicon/favicon.ico') }}">
-      <link rel="apple-touch-icon" sizes="180x180" href="{{ url_for('static', filename='favicon/apple-touch-icon.png') }}">
-    <link rel="icon" type="image/png" sizes="32x32" href="{{ url_for('static', filename='favicon/favicon-32x32.png') }}">
-    <link rel="icon" type="image/png" sizes="16x16" href="{{ url_for('static', filename='favicon/favicon-16x16.png') }}">
-    <link rel="manifest" href="{{ url_for('static', filename='favicon//site.webmanifest') }}">
-    <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/4.7.0/css/font-awesome.min.css">
-
-
-
-  </head>
-
-
-  <body>
-    
-
-
-    <header class="bg-white p-2 flex justify-between items-center">
-      <h1 class="text-lg font-medium">DocsGPT 🦖 Preview</h1>
-        <div>
-      <a href="https://github.com/arc53/docsgpt" class="text-blue-500 hover:text-blue-800 text-sm">About</a>
-            {% if not api_key_set %}
-      <button class="text-sm text-yellow-500 hover:text-yellow-800" onclick="resetApiKey()">Reset Key</button>
-        {% endif %}
-            </div>
-    </header>
-
-  
- <!-- Alert Info  -->
- <div class="border flex justify-between 
-  w-auto px-4 py-3 rounded relative 
-  hidden" style="background-color: rgb(197, 51, 51);color: white;" id="error-alert" role="alert">
-  <span class="block sm:inline" id="text-error"></span>
-  <strong class="text-xl align-center alert-del" style="cursor: pointer;" id="close">&times;</strong>
-</div>
-
-
-    <div class="lg:flex ml-2 mr-2">
-      <div class="lg:w-3/4 min-h-screen max-h-screen">
-        <div class="w-full flex flex-col h-5/6">
-          <div id="messages-container" style="overflow: auto;" class="sm:max-lg:mb-[12rem]">
-
-            <div id="messages" class="w-full flex flex-col mt-2" >
-              <div class="bg-indigo-500 text-white p-2 rounded-lg mb-2 self-start">
-                <p class="text-sm">Hello, ask me anything about this library. Im here to help</p>
-              </div>
-              <div class="bg-blue-500 text-white p-2 rounded-lg mb-2 self-end">
-                <p class="text-sm">How to merge tables?</p>
-              </div>
-              <div class="bg-indigo-500 text-white p-2 rounded-lg mb-2 self-start">
-                <p class="text-sm">To merge two tables in pandas, you can use the pd.merge() function. The basic syntax is:<br>
-pd.merge(left, right, on, how)<br>
-where left and right are the two tables to merge, on is the column to merge on, and how is the type of merge to perform.<br>
-For example, to merge the two tables df1 and df2 on the column 'key', you can use:<br>
-pd.merge(df1, df2, on='key', how='left')<br>
-This will return a new DataFrame with all the columns from both tables, and only the rows that match the 'key' column. </p>
-              </div>
-
-          </div>
-        </div>
-
-        <div class="fixed bottom-0 w-full mt-4 mb-2 lg:w-3/4">
-        <form id="message-form" autocomplete="off" class="flex items-stretch">
-          <input autocomplete="off" id="message-input" class="bg-white p-2 rounded-lg ml-2 text-sm w-full" type="text" placeholder="Type your message here...">
-          <button id="button-submit" class="bg-blue-500 text-white p-2 rounded-lg ml-2 mr-2 text-sm sm:max-lg:p-5" type="submit">Send</button>
-        </form>
-        </div>
-
-        
-
-
-    </div>
-        </div>
-        <div class="lg:w-1/4 p-2 sm:max-lg:hidden">
-          <p class="text-sm">This is a chatbot that uses the GPT-3, Faiss and <a href="https://github.com/hwchase17/langchain" class="text-blue-500 hover:text-blue-800">LangChain</a> to answer questions</p>
-          <br>
-          <p class="text-sm">The source code is available on <a href="https://github.com/arc53/docsgpt" class="text-blue-500 hover:text-blue-800">Github</a></p><br>
-          <p class="text-sm">Currently It uses python pandas documentation, so it will respond to information relevant to pandas. If you want to train it on different documentation -  <a href="https://github.com/arc53/docsgpt/wiki/How-to-train-on-other-documentation" class="text-blue-500 hover:text-blue-800"> please follow this guide </a></p><br>
-          <p class="text-sm">If you want to launch it on your own server - <a href="https://github.com/arc53/docsgpt/wiki/How-to-train-on-other-documentation" class="text-blue-500 hover:text-blue-800"> follow this guide </a></p><br>
-            <label  class="block mb-2 text-sm font-medium text-gray-900">Select documentation from DocsHUB</label>
-            <select id="select-docs" class="bg-gray-50 border border-gray-300 text-gray-900 text-sm rounded-lg focus:ring-blue-500 focus:border-blue-500 block w-full p-2.5">
-              <option selected value="default">Choose documentation</option>
-              <option value="default">Default</option>
-            </select>
-            <form action="/api/upload" method="post" enctype="multipart/form-data" class="mt-2">
-                <input type="file" name="file" class="py-4" id="file-upload">
-                <input type="text" name="user" value="local" hidden>
-                <input type="text" name="name" placeholder="Name:">
-
-
-              <button type="submit" class="py-2 px-4 text-white bg-blue-500 rounded-md hover:bg-blue-600 focus:outline-none focus:ring-2 focus:ring-offset-2 focus:ring-blue-500">
-                Upload
-              </button>
-            </form>
-
-
-
-        </div>
-    </div>
-
-  <div class="flex items-center justify-center h-full">
-    
- 
-</div>
-
-
-
-
-{% if not api_key_set %}
-
-<div class="fixed z-10 overflow-y-auto top-0 w-full left-0 show" id="modal">
-  <div class="flex items-center justify-center min-height-100vh pt-4 px-4 pb-20 text-center sm:block sm:p-0">
-    <div class="fixed inset-0 transition-opacity">
-      <div class="absolute inset-0 bg-gray-900 opacity-75" />
-    </div>
-    <span class="hidden sm:inline-block sm:align-middle sm:h-screen">&#8203;</span>
-    <div class=" text-sm inline-block align-center bg-white rounded-lg text-left overflow-hidden shadow-xl transform transition-all sm:my-8 sm:align-middle sm:max-w-lg sm:w-full" role="dialog" aria-modal="true" aria-labelledby="modal-headline">
-       <form id="api-key-form">
-        <div class="bg-white px-4 pt-5 pb-4 sm:p-6 sm:pb-4">
-        <h2>Before you can start using DocsGPT we need you to provide an API key for llm. Currently, we support only OpenAI but soon many more. You can find it <a class="text-blue-500 hover:text-blue-800" href="https://platform.openai.com/account/api-keys">here</a></h2><br>
-        <label>OpenAI API key:</label>
-
-        <input id="api-key-input" type="password" class="w-full bg-gray-100 p-2 mt-2 mb-3" placeholder="Paste you Api Key here">
-
-      </div>
-      <div class="bg-gray-200 px-4 py-3 text-right">
-        <button type="submit" class="py-2 px-4 bg-blue-500 text-white rounded hover:bg-blue-700 mr-2">Save</button>
-
-      </div>
-            </form>
-    </div>
-  </div>
-</div>
-{% endif %}
-
-
-
-      <script>
-          function docsIndex() {
-                // loads latest index from https://raw.githubusercontent.com/arc53/DocsHUB/main/combined.json
-                // and stores it in localStorage
-                fetch('/api/combine')
-                    .then(response => response.json())
-                    .then(data => {
-                        localStorage.setItem("docsIndex", JSON.stringify(data));
-                        localStorage.setItem("docsIndexDate", Date.now());
-                        generateOptions()
-                    }
-
-                )
-
-            }
-          function generateOptions(){
-                docsIndex = localStorage.getItem('docsIndex')
-                // create option on select with id select-docs
-                var select = document.getElementById("select-docs");
-                // convert docsIndex to json
-                docsIndex = JSON.parse(docsIndex)
-                // create option for each key in docsIndex
-                for (var key in docsIndex) {
-                    var option = document.createElement("option");
-                    if (docsIndex[key].location == 'docshub'){
-                        if (docsIndex[key].name == docsIndex[key].language) {
-                            option.text = docsIndex[key].name + " " + docsIndex[key].version;
-                            option.value = docsIndex[key].name + "/" + ".project" + "/" + docsIndex[key].version + "/{{ embeddings_choice }}/";
-                            if (docsIndex[key].model == "{{ embeddings_choice }}") {
-                                select.add(option);
-                            }
-                        }
-                        else {
-                            option.text = docsIndex[key].name + " " + docsIndex[key].version;
-                            option.value = docsIndex[key].language + "/" + docsIndex[key].name + "/" + docsIndex[key].version + "/{{ embeddings_choice }}/";
-                            if (docsIndex[key].model == "{{ embeddings_choice }}") {
-                                select.add(option);
-                            }
-                        }
-                    }
-                    else {
-                        option.text = docsIndex[key].name;
-                        option.value = docsIndex[key].location + "/" + docsIndex[key].name;
-                        select.add(option);
-                    }
-                }
-
-          }
-        {% if not api_key_set %}
-        if (localStorage.getItem('apiKey') === null) {
-            console.log("apiKey is not set")
-            document.getElementById('modal').classList.toggle('hidden')
-        }
-        {% endif %}
-        if (localStorage.getItem('docsIndex') === null) {
-            console.log("docsIndex is not set")
-            docsIndex()
-        }
-        else if (localStorage.getItem("docsIndexDate") < Date.now() - 900000) {
-            console.log("docsIndex is older than 15 minutes")
-            docsIndex()
-        }
-
-        generateOptions()
-
-  </script>
-    {% if not api_key_set %}
-    <script src="{{url_for('static',filename='src/authapi.js')}}"></script>
-    {% endif %}
-  <script src="{{url_for('static',filename='src/chat.js')}}"></script>
-  <script src="{{url_for('static',filename='src/choiceChange.js')}}"></script>
-
-  </body>
-</html>

application/vectorstore/base.py

@@ -0,0 +1,56 @@
+from abc import ABC, abstractmethod
+import os
+from langchain_community.embeddings import (
+    HuggingFaceEmbeddings,
+    CohereEmbeddings,
+    HuggingFaceInstructEmbeddings,
+)
+from langchain_openai import OpenAIEmbeddings
+from application.core.settings import settings
+
+class BaseVectorStore(ABC):
+    def __init__(self):
+        pass
+
+    @abstractmethod
+    def search(self, *args, **kwargs):
+        pass
+
+    def is_azure_configured(self):
+        return settings.OPENAI_API_BASE and settings.OPENAI_API_VERSION and settings.AZURE_DEPLOYMENT_NAME
+
+    def _get_embeddings(self, embeddings_name, embeddings_key=None):
+        embeddings_factory = {
+            "openai_text-embedding-ada-002": OpenAIEmbeddings,
+            "huggingface_sentence-transformers/all-mpnet-base-v2": HuggingFaceEmbeddings,
+            "huggingface_hkunlp/instructor-large": HuggingFaceInstructEmbeddings,
+            "cohere_medium": CohereEmbeddings
+        }
+        
+        if embeddings_name not in embeddings_factory:
+            raise ValueError(f"Invalid embeddings_name: {embeddings_name}")
+
+        if embeddings_name == "openai_text-embedding-ada-002":
+            if self.is_azure_configured():
+                os.environ["OPENAI_API_TYPE"] = "azure"
+                embedding_instance = embeddings_factory[embeddings_name](
+                    model=settings.AZURE_EMBEDDINGS_DEPLOYMENT_NAME
+                )
+            else:
+                embedding_instance = embeddings_factory[embeddings_name](
+                    openai_api_key=embeddings_key
+                )
+        elif embeddings_name == "cohere_medium":
+            embedding_instance = embeddings_factory[embeddings_name](
+                cohere_api_key=embeddings_key
+            )
+        elif embeddings_name == "huggingface_sentence-transformers/all-mpnet-base-v2":
+            embedding_instance = embeddings_factory[embeddings_name](
+                #model_name="./model/all-mpnet-base-v2",
+                model_kwargs={"device": "cpu"},
+            )
+        else:
+            embedding_instance = embeddings_factory[embeddings_name]()
+            
+        return embedding_instance
+

application/vectorstore/document_class.py

@@ -0,0 +1,8 @@
+class Document(str):
+    """Class for storing a piece of text and associated metadata."""
+
+    def __new__(cls, page_content: str, metadata: dict):
+        instance = super().__new__(cls, page_content)
+        instance.page_content = page_content
+        instance.metadata = metadata
+        return instance

application/vectorstore/elasticsearch.py

@@ -0,0 +1,213 @@
+from application.vectorstore.base import BaseVectorStore
+from application.core.settings import settings
+from application.vectorstore.document_class import Document
+import elasticsearch
+
+
+
+
+class ElasticsearchStore(BaseVectorStore):
+    _es_connection = None  # Class attribute to hold the Elasticsearch connection
+
+    def __init__(self, path, embeddings_key, index_name=settings.ELASTIC_INDEX):
+        super().__init__()
+        self.path = path.replace("application/indexes/", "").rstrip("/")
+        self.embeddings_key = embeddings_key
+        self.index_name = index_name
+        
+        if ElasticsearchStore._es_connection is None:
+            connection_params = {}
+            if settings.ELASTIC_URL:
+                connection_params["hosts"] = [settings.ELASTIC_URL]
+                connection_params["http_auth"] = (settings.ELASTIC_USERNAME, settings.ELASTIC_PASSWORD)
+            elif settings.ELASTIC_CLOUD_ID:
+                connection_params["cloud_id"] = settings.ELASTIC_CLOUD_ID
+                connection_params["basic_auth"] = (settings.ELASTIC_USERNAME, settings.ELASTIC_PASSWORD)
+            else:
+                raise ValueError("Please provide either elasticsearch_url or cloud_id.")
+
+            
+
+            ElasticsearchStore._es_connection = elasticsearch.Elasticsearch(**connection_params)
+            
+        self.docsearch = ElasticsearchStore._es_connection
+
+    def connect_to_elasticsearch(
+        *,
+        es_url = None,
+        cloud_id = None,
+        api_key = None,
+        username = None,
+        password = None,
+    ):
+        try:
+            import elasticsearch
+        except ImportError:
+            raise ImportError(
+                "Could not import elasticsearch python package. "
+                "Please install it with `pip install elasticsearch`."
+            )
+
+        if es_url and cloud_id:
+            raise ValueError(
+                "Both es_url and cloud_id are defined. Please provide only one."
+            )
+
+        connection_params = {}
+
+        if es_url:
+            connection_params["hosts"] = [es_url]
+        elif cloud_id:
+            connection_params["cloud_id"] = cloud_id
+        else:
+            raise ValueError("Please provide either elasticsearch_url or cloud_id.")
+
+        if api_key:
+            connection_params["api_key"] = api_key
+        elif username and password:
+            connection_params["basic_auth"] = (username, password)
+
+        es_client = elasticsearch.Elasticsearch(
+            **connection_params,
+        )
+        try:
+            es_client.info()
+        except Exception as e:
+            raise e
+
+        return es_client
+
+    def search(self, question, k=2, index_name=settings.ELASTIC_INDEX, *args, **kwargs):
+        embeddings = self._get_embeddings(settings.EMBEDDINGS_NAME, self.embeddings_key)
+        vector = embeddings.embed_query(question)
+        knn = {
+            "filter": [{"match": {"metadata.store.keyword": self.path}}],
+            "field": "vector",
+            "k": k,
+            "num_candidates": 100,
+            "query_vector": vector,
+        }
+        full_query = {
+            "knn": knn,
+            "query": {
+                "bool": {
+                    "must": [
+                        {
+                            "match": {
+                                "text": {
+                                    "query": question,
+                                }
+                            }
+                        }
+                    ],
+                    "filter": [{"match": {"metadata.store.keyword": self.path}}],
+                }
+            },
+            "rank": {"rrf": {}},
+        }
+        resp = self.docsearch.search(index=self.index_name, query=full_query['query'], size=k, knn=full_query['knn'])
+        # create Documents objects from the results page_content ['_source']['text'], metadata ['_source']['metadata']
+        doc_list = []
+        for hit in resp['hits']['hits']:
+            
+            doc_list.append(Document(page_content = hit['_source']['text'], metadata = hit['_source']['metadata']))
+        return doc_list
+
+    def _create_index_if_not_exists(
+            self, index_name, dims_length
+        ):
+
+        if self._es_connection.indices.exists(index=index_name):
+            print(f"Index {index_name} already exists.")
+
+        else:
+
+            indexSettings = self.index(
+                dims_length=dims_length,
+            )
+            self._es_connection.indices.create(index=index_name, **indexSettings)
+
+    def index(
+            self,
+            dims_length,
+        ):
+        return {
+            "mappings": {
+                "properties": {
+                    "vector": {
+                        "type": "dense_vector",
+                        "dims": dims_length,
+                        "index": True,
+                        "similarity": "cosine",
+                    },
+                }
+            }
+        }
+
+    def add_texts(
+        self,
+        texts,
+        metadatas = None,
+        ids = None,
+        refresh_indices = True,
+        create_index_if_not_exists = True,
+        bulk_kwargs = None,
+        **kwargs,
+        ):
+        
+        from elasticsearch.helpers import BulkIndexError, bulk
+
+        bulk_kwargs = bulk_kwargs or {}
+        import uuid
+        embeddings = []
+        ids = ids or [str(uuid.uuid4()) for _ in texts]
+        requests = []
+        embeddings = self._get_embeddings(settings.EMBEDDINGS_NAME, self.embeddings_key)
+
+        vectors = embeddings.embed_documents(list(texts))
+
+        dims_length = len(vectors[0])
+
+        if create_index_if_not_exists:
+            self._create_index_if_not_exists(
+                index_name=self.index_name, dims_length=dims_length
+            )
+
+        for i, (text, vector) in enumerate(zip(texts, vectors)):
+            metadata = metadatas[i] if metadatas else {}
+
+            requests.append(
+                {
+                    "_op_type": "index",
+                    "_index": self.index_name,
+                    "text": text,
+                    "vector": vector,
+                    "metadata": metadata,
+                    "_id": ids[i],
+                }
+            )
+
+
+        if len(requests) > 0:
+            try:
+                success, failed = bulk(
+                    self._es_connection,
+                    requests,
+                    stats_only=True,
+                    refresh=refresh_indices,
+                    **bulk_kwargs,
+                )
+                return ids
+            except BulkIndexError as e:
+                print(f"Error adding texts: {e}")
+                firstError = e.errors[0].get("index", {}).get("error", {})
+                print(f"First error reason: {firstError.get('reason')}")
+                raise e
+
+        else:
+            return []
+
+    def delete_index(self):
+        self._es_connection.delete_by_query(index=self.index_name, query={"match": {
+                                      "metadata.store.keyword": self.path}},)
+

application/vectorstore/faiss.py

@@ -0,0 +1,46 @@
+from langchain_community.vectorstores import FAISS
+from application.vectorstore.base import BaseVectorStore
+from application.core.settings import settings
+
+class FaissStore(BaseVectorStore):
+
+    def __init__(self, path, embeddings_key, docs_init=None):
+        super().__init__()
+        self.path = path
+        embeddings = self._get_embeddings(settings.EMBEDDINGS_NAME, embeddings_key)
+        if docs_init:
+            self.docsearch = FAISS.from_documents(
+                docs_init, embeddings
+            )
+        else:
+            self.docsearch = FAISS.load_local(
+                self.path, embeddings
+            )
+        self.assert_embedding_dimensions(embeddings)
+
+    def search(self, *args, **kwargs):
+        return self.docsearch.similarity_search(*args, **kwargs)
+
+    def add_texts(self, *args, **kwargs):
+        return self.docsearch.add_texts(*args, **kwargs)
+
+    def save_local(self, *args, **kwargs):
+        return self.docsearch.save_local(*args, **kwargs)
+
+    def delete_index(self, *args, **kwargs):
+        return self.docsearch.delete(*args, **kwargs)
+
+    def assert_embedding_dimensions(self, embeddings):
+        """
+        Check that the word embedding dimension of the docsearch index matches
+        the dimension of the word embeddings used 
+        """
+        if settings.EMBEDDINGS_NAME == "huggingface_sentence-transformers/all-mpnet-base-v2":
+            try:
+                word_embedding_dimension = embeddings.client[1].word_embedding_dimension
+            except AttributeError as e:
+                raise AttributeError("word_embedding_dimension not found in embeddings.client[1]") from e
+            docsearch_index_dimension = self.docsearch.index.d
+            if word_embedding_dimension != docsearch_index_dimension:
+                raise ValueError(f"word_embedding_dimension ({word_embedding_dimension}) " +
+                                 f"!= docsearch_index_word_embedding_dimension ({docsearch_index_dimension})")

application/vectorstore/mongodb.py

@@ -0,0 +1,126 @@
+from application.vectorstore.base import BaseVectorStore
+from application.core.settings import settings
+from application.vectorstore.document_class import Document
+
+class MongoDBVectorStore(BaseVectorStore):
+    def __init__(
+        self,
+        path: str = "",
+        embeddings_key: str = "embeddings",
+        collection: str = "documents",
+        index_name: str = "vector_search_index",
+        text_key: str = "text",
+        embedding_key: str = "embedding",
+        database: str = "docsgpt",
+    ):
+        self._index_name = index_name
+        self._text_key = text_key
+        self._embedding_key = embedding_key
+        self._embeddings_key = embeddings_key
+        self._mongo_uri = settings.MONGO_URI
+        self._path = path.replace("application/indexes/", "").rstrip("/")
+        self._embedding = self._get_embeddings(settings.EMBEDDINGS_NAME, embeddings_key)
+
+        try:
+            import pymongo
+        except ImportError:
+            raise ImportError(
+                "Could not import pymongo python package. "
+                "Please install it with `pip install pymongo`."
+            )
+
+        self._client = pymongo.MongoClient(self._mongo_uri)
+        self._database = self._client[database]
+        self._collection = self._database[collection]
+
+        
+    def search(self, question, k=2, *args, **kwargs):
+        query_vector = self._embedding.embed_query(question)
+
+        pipeline = [
+            {
+                "$vectorSearch": {
+                    "queryVector": query_vector, 
+                    "path": self._embedding_key,
+                    "limit": k, 
+                    "numCandidates": k * 10, 
+                    "index": self._index_name,
+                    "filter": {
+                        "store": {"$eq": self._path}
+                    }
+                }
+            }
+        ]
+
+        cursor = self._collection.aggregate(pipeline)
+        
+        results = []
+        for doc in cursor:
+            text = doc[self._text_key]
+            doc.pop("_id")
+            doc.pop(self._text_key)
+            doc.pop(self._embedding_key)
+            metadata = doc
+            results.append(Document(text, metadata))
+        return results
+    
+    def _insert_texts(self, texts, metadatas):
+        if not texts:
+            return []
+        embeddings = self._embedding.embed_documents(texts)
+        to_insert = [
+            {self._text_key: t, self._embedding_key: embedding, **m}
+            for t, m, embedding in zip(texts, metadatas, embeddings)
+        ]
+        # insert the documents in MongoDB Atlas
+        insert_result = self._collection.insert_many(to_insert)
+        return insert_result.inserted_ids
+    
+    def add_texts(self,
+        texts,
+        metadatas = None,
+        ids = None,
+        refresh_indices = True,
+        create_index_if_not_exists = True,
+        bulk_kwargs = None,
+        **kwargs,):
+
+
+        #dims = self._embedding.client[1].word_embedding_dimension
+        # # check if index exists
+        # if create_index_if_not_exists:
+        #     # check if index exists
+        #     info = self._collection.index_information()
+        #     if self._index_name not in info:
+        #         index_mongo = {
+        #         "fields": [{
+        #             "type": "vector",
+        #             "path": self._embedding_key,
+        #             "numDimensions": dims,
+        #             "similarity": "cosine",
+        #         },
+        #         {
+        #             "type": "filter",
+        #             "path": "store"
+        #         }]
+        #         }
+        #         self._collection.create_index(self._index_name, index_mongo)
+
+        batch_size = 100
+        _metadatas = metadatas or ({} for _ in texts)
+        texts_batch = []
+        metadatas_batch = []
+        result_ids = []
+        for i, (text, metadata) in enumerate(zip(texts, _metadatas)):
+            texts_batch.append(text)
+            metadatas_batch.append(metadata)
+            if (i + 1) % batch_size == 0:
+                result_ids.extend(self._insert_texts(texts_batch, metadatas_batch))
+                texts_batch = []
+                metadatas_batch = []
+        if texts_batch:
+            result_ids.extend(self._insert_texts(texts_batch, metadatas_batch))
+        return result_ids
+    
+    def delete_index(self, *args, **kwargs):
+        self._collection.delete_many({"store": self._path})

application/vectorstore/vector_creator.py

@@ -0,0 +1,18 @@
+from application.vectorstore.faiss import FaissStore
+from application.vectorstore.elasticsearch import ElasticsearchStore
+from application.vectorstore.mongodb import MongoDBVectorStore
+
+
+class VectorCreator:
+    vectorstores = {
+        'faiss': FaissStore,
+        'elasticsearch':ElasticsearchStore,
+        'mongodb': MongoDBVectorStore,
+    }
+
+    @classmethod
+    def create_vectorstore(cls, type, *args, **kwargs):
+        vectorstore_class = cls.vectorstores.get(type.lower())
+        if not vectorstore_class:
+            raise ValueError(f"No vectorstore class found for type {type}")
+        return vectorstore_class(*args, **kwargs)

application/worker.py

@@ -20,15 +20,34 @@ except FileExistsError:
     pass
 
 
+# Define a function to extract metadata from a given filename.
 def metadata_from_filename(title):
-    return {'title': title}
+    store = '/'.join(title.split('/')[1:3])
+    return {'title': title, 'store': store}
 
 
+# Define a function to generate a random string of a given length.
 def generate_random_string(length):
     return ''.join([string.ascii_letters[i % 52] for i in range(length)])
 
+current_dir = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
 
+# Define the main function for ingesting and processing documents.
 def ingest_worker(self, directory, formats, name_job, filename, user):
+    """
+    Ingest and process documents.
+
+    Args:
+        self: Reference to the instance of the task.
+        directory (str): Specifies the directory for ingesting ('inputs' or 'temp').
+        formats (list of str): List of file extensions to consider for ingestion (e.g., [".rst", ".md"]).
+        name_job (str): Name of the job for this ingestion task.
+        filename (str): Name of the file to be ingested.
+        user (str): Identifier for the user initiating the ingestion.
+
+    Returns:
+        dict: Information about the completed ingestion task, including input parameters and a "limited" flag.
+    """
     # directory = 'inputs' or 'temp'
     # formats = [".rst", ".md"]
     input_files = None
@@ -43,9 +62,13 @@ def ingest_worker(self, directory, formats, name_job, filename, user):
     min_tokens = 150
     max_tokens = 1250
     full_path = directory + '/' + user + '/' + name_job
+    import sys
+    print(full_path, file=sys.stderr)
     # check if API_URL env variable is set
     file_data = {'name': name_job, 'file': filename, 'user': user}
     response = requests.get(urljoin(settings.API_URL, "/api/download"), params=file_data)
+    # check if file is in the response
+    print(response, file=sys.stderr)
     file = response.content
 
     if not os.path.exists(full_path):
@@ -78,11 +101,15 @@ def ingest_worker(self, directory, formats, name_job, filename, user):
     # get files from outputs/inputs/index.faiss and outputs/inputs/index.pkl
     # and send them to the server (provide user and name in form)
     file_data = {'name': name_job, 'user': user}
-    files = {'file_faiss': open(full_path + '/index.faiss', 'rb'),
-             'file_pkl': open(full_path + '/index.pkl', 'rb')}
-    response = requests.post(urljoin(settings.API_URL, "/api/upload_index"), files=files, data=file_data)
+    if settings.VECTOR_STORE == "faiss":
+        files = {'file_faiss': open(full_path + '/index.faiss', 'rb'),
+                'file_pkl': open(full_path + '/index.pkl', 'rb')}
+        response = requests.post(urljoin(settings.API_URL, "/api/upload_index"), files=files, data=file_data)
+        response = requests.get(urljoin(settings.API_URL, "/api/delete_old?path=" + full_path))
+    else:
+        response = requests.post(urljoin(settings.API_URL, "/api/upload_index"), data=file_data)
 
-    response = requests.get(urljoin(settings.API_URL, "/api/delete_old?path="))
+    
     # delete local
     shutil.rmtree(full_path)
 

codecov.yml

@@ -0,0 +1,2 @@
+ignore:
+  - "*/tests/*"

docker-compose-local.yaml

@@ -0,0 +1,26 @@
+version: "3.9"
+
+services:
+  frontend:
+    build: ./frontend
+    environment:
+      - VITE_API_HOST=http://localhost:7091
+      - VITE_API_STREAMING=$VITE_API_STREAMING
+      - VITE_EMBEDDINGS_NAME=$EMBEDDINGS_NAME
+    ports:
+      - "5173:5173"
+
+  redis:
+    image: redis:6-alpine
+    ports:
+      - 6379:6379
+
+  mongo:
+    image: mongo:6
+    ports:
+      - 27017:27017
+    volumes:
+      - mongodb_data_container:/data/db
+
+volumes:
+  mongodb_data_container:

docker-compose-mock.yaml

@@ -0,0 +1,22 @@
+version: "3.9"
+
+services:
+  frontend:
+    build: ./frontend
+    environment:
+      - VITE_API_HOST=http://localhost:7091
+      - VITE_API_STREAMING=$VITE_API_STREAMING
+    ports:
+      - "5173:5173"
+    depends_on:
+      - mock-backend
+
+  mock-backend:
+    build: ./mock-backend
+    ports:
+      - "7091:7091"
+
+  redis:
+    image: redis:6-alpine
+    ports:
+      - 6379:6379

docker-compose.yaml

@@ -14,12 +14,12 @@ services:
   backend:
     build: ./application
     environment:
-      - API_KEY=$OPENAI_API_KEY
-      - EMBEDDINGS_KEY=$OPENAI_API_KEY
+      - API_KEY=$API_KEY
+      - EMBEDDINGS_KEY=$API_KEY
+      - LLM_NAME=$LLM_NAME
       - CELERY_BROKER_URL=redis://redis:6379/0
       - CELERY_RESULT_BACKEND=redis://redis:6379/1
       - MONGO_URI=mongodb://mongo:27017/docsgpt
-      - SELF_HOSTED_MODEL=$SELF_HOSTED_MODEL
     ports:
       - "7091:7091"
     volumes:
@@ -34,8 +34,9 @@ services:
     build: ./application
     command: celery -A application.app.celery worker -l INFO
     environment:
-      - API_KEY=$OPENAI_API_KEY
-      - EMBEDDINGS_KEY=$OPENAI_API_KEY
+      - API_KEY=$API_KEY
+      - EMBEDDINGS_KEY=$API_KEY
+      - LLM_NAME=$LLM_NAME
       - CELERY_BROKER_URL=redis://redis:6379/0
       - CELERY_RESULT_BACKEND=redis://redis:6379/1
       - MONGO_URI=mongodb://mongo:27017/docsgpt

docs/README.md

@@ -0,0 +1,51 @@
+# nextra-docsgpt
+
+## Setting Up Docs Folder of DocsGPT Locally
+
+### 1. Clone the DocsGPT repository:
+
+```bash
+git clone https://github.com/arc53/DocsGPT.git
+```
+### 2. Navigate to the docs folder:
+
+```bash
+cd DocsGPT/docs
+```
+
+The docs folder contains the markdown files that make up the documentation. The majority of the files are in the pages directory. Some notable files in this folder include:
+
+`index.mdx`: The main documentation file.
+`_app.js`: This file is used to customize the default Next.js application shell.
+`theme.config.jsx`: This file is for configuring the Nextra theme for the documentation.
+
+### 3. Verify that you have Node.js and npm installed in your system. You can check by running:
+
+```bash
+node --version
+npm --version
+```
+
+### 4. If not installed, download Node.js and npm from the respective official websites.
+
+### 5. Once you have Node.js and npm running, proceed to install yarn - another package manager that helps to manage project dependencies:
+
+```bash
+npm install --global yarn
+```
+
+### 6. Install the project dependencies using yarn:
+
+```bash
+yarn install
+```
+
+### 7. After the successful installation of the project dependencies, start the local server:
+
+```bash
+yarn dev
+```
+
+- Now, you should be able to view the docs on your local environment by visiting `http://localhost:5000`. You can explore the different markdown files and make changes as you see fit.
+
+- **Footnotes:** This guide assumes you have Node.js and npm installed. The guide involves running a local server using yarn, and viewing the documentation offline. If you encounter any issues, it may be worth verifying your Node.js and npm installations and whether you have installed yarn correctly.

docs/next.config.js

@@ -0,0 +1,9 @@
+const withNextra = require('nextra')({
+    theme: 'nextra-theme-docs',
+    themeConfig: './theme.config.jsx'
+  })
+   
+  module.exports = withNextra()
+   
+  // If you have other Next.js configurations, you can pass them as the parameter:
+  // module.exports = withNextra({ /* other next.js config */ })

docs/package.json

@@ -0,0 +1,17 @@
+{
+"scripts":{
+      "dev": "next dev",
+    "build": "next build",
+    "start": "next start"
+  },
+    "license": "MIT",
+  "dependencies": {
+    "@vercel/analytics": "^1.0.2",
+    "docsgpt": "^0.2.4",
+    "next": "^13.5.1",
+    "nextra": "^2.12.3",
+    "nextra-theme-docs": "^2.12.3",
+    "react": "^18.2.0",
+    "react-dom": "^18.2.0"
+  }
+}

docs/pages/Deploying/Hosting-the-app.md

@@ -0,0 +1,110 @@
+# Self-hosting DocsGPT on Amazon Lightsail
+
+Here's a step-by-step guide on how to set up an Amazon Lightsail instance to host DocsGPT.
+
+## Configuring your instance
+
+(If you know how to create a Lightsail instance, you can skip to the recommended configuration part by clicking [here](#connecting-to-your-newly-created-instance)).
+
+### 1. Create an AWS Account: 
+If you haven't already, create or log in to your AWS account at https://lightsail.aws.amazon.com.
+
+### 2. Create an Instance: 
+
+a. Click "Create Instance."
+
+b. Select the "Instance location." In most cases, the default location works fine.
+
+c. Choose "Linux/Unix" as the image and "Ubuntu 20.04 LTS" as the Operating System.
+
+d. Configure the instance plan based on your requirements. A "1 GB, 1vCPU, 40GB SSD, and 2TB transfer" setup is recommended for most scenarios.
+
+e. Give your instance a unique name and click "Create Instance."
+
+PS: It may take a few minutes for the instance setup to complete.
+
+### Connecting to Your newly created Instance
+
+Your instance will be ready a few minutes after creation. To access it, open the instance and click "Connect using SSH."
+
+#### Clone the DocsGPT Repository
+
+A terminal window will pop up, and the first step will be to clone the DocsGPT Git repository:
+
+`git clone https://github.com/arc53/DocsGPT.git`
+
+#### Download the package information
+
+Once it has finished cloning the repository, it is time to download the package information from all sources. To do so, simply enter the following command:
+
+`sudo apt update`
+
+#### Install Docker and Docker Compose
+
+DocsGPT backend and worker use Python, Frontend is written on React and the whole application is containerized using Docker. To install Docker and Docker Compose, enter the following commands:
+
+`sudo apt install docker.io`
+
+And now install docker-compose:
+
+`sudo apt install docker-compose`
+
+#### Access the DocsGPT Folder
+
+Enter the following command to access the folder in which the DocsGPT docker-compose file is present.
+
+`cd DocsGPT/`
+
+#### Prepare the Environment
+
+Inside the DocsGPT folder create a `.env` file and copy the contents of `.env_sample` into it.
+
+`nano .env`
+
+Make sure your `.env` file looks like this:
+
+```
+OPENAI_API_KEY=(Your OpenAI API key)
+VITE_API_STREAMING=true
+SELF_HOSTED_MODEL=false
+```
+
+To save the file, press CTRL+X, then Y, and then ENTER.
+
+Next, set the correct IP for the Backend by opening the docker-compose.yml file:
+
+`nano docker-compose.yml`
+
+And Change line 7 to: `VITE_API_HOST=http://localhost:7091`
+to this `VITE_API_HOST=http://<your instance public IP>:7091`
+
+This will allow the frontend to connect to the backend.
+
+#### Running the Application
+
+You're almost there! Now that all the necessary bits and pieces have been installed, it is time to run the application. To do so, use the following command:
+
+`sudo docker-compose up -d`
+
+Launching it for the first time will take a few minutes to download all the necessary dependencies and build.
+
+Once this is done you can go ahead and close the terminal window.
+
+#### Enabling Ports 
+
+a. Before you are able to access your live instance, you must first enable the port that it is using.
+
+b. Open your Lightsail instance and head to "Networking".
+
+c. Then click on "Add rule" under "IPv4 Firewall", enter `5173` as your port, and hit "Create". 
+Repeat the process for port `7091`.
+
+#### Access your instance
+
+Your instance is now available at your Public IP Address on port 5173. Enjoy using DocsGPT!
+
+## Other Deployment Options
+
+- [Deploy DocsGPT on Civo Compute Cloud](https://dev.to/rutamhere/deploying-docsgpt-on-civo-compute-c)
+- [Deploy DocsGPT on DigitalOcean Droplet](https://dev.to/rutamhere/deploying-docsgpt-on-digitalocean-droplet-50ea)
+- [Deploy DocsGPT on Kamatera Performance Cloud](https://dev.to/rutamhere/deploying-docsgpt-on-kamatera-performance-cloud-1bj)

docs/pages/Deploying/Quickstart.md

@@ -0,0 +1,128 @@
+## Launching Web App
+**Note**: Make sure you have Docker installed
+
+**On macOS or Linux:**
+Just run the following command:
+
+```bash
+./setup.sh
+```
+
+This command will install all the necessary dependencies and provide you with an option to use our LLM API, download the local model or use OpenAI.
+
+If you prefer to follow manual steps, refer to this guide:
+
+1. Open and download this repository with 
+   ```bash
+   git clone https://github.com/arc53/DocsGPT.git
+   ```
+2. Create a `.env` file in your root directory and set your `API_KEY` with your [OpenAI API key](https://platform.openai.com/account/api-keys). (optional in case you want to use OpenAI)
+3. Run the following commands:
+   ```bash
+   docker-compose build && docker-compose up
+   ```
+4. Navigate to http://localhost:5173/.
+
+To stop, simply press **Ctrl + C**.
+
+**For WINDOWS:**
+
+To run the setup on Windows, you have two options: using the Windows Subsystem for Linux (WSL) or using Git Bash or Command Prompt.
+
+**Option 1: Using Windows Subsystem for Linux (WSL):**
+
+1. Install WSL if you haven't already. You can follow the official Microsoft documentation for installation: (https://learn.microsoft.com/en-us/windows/wsl/install).
+2. After setting up WSL, open the WSL terminal.
+3. Clone the repository and create the `.env` file:
+   ```bash
+   git clone https://github.com/arc53/DocsGPT.git
+   cd DocsGPT
+   echo "API_KEY=Yourkey" > .env
+   echo "VITE_API_STREAMING=true" >> .env
+   ```
+4. Run the following command to start the setup with Docker Compose:
+   ```bash
+   ./run-with-docker-compose.sh
+   ```
+6. Open your web browser and navigate to http://localhost:5173/.
+7. To stop the setup, just press **Ctrl + C** in the WSL terminal
+
+**Option 2: Using Git Bash or Command Prompt (CMD):**
+
+1. Install Git for Windows if you haven't already. Download it from the official website: (https://gitforwindows.org/).
+2. Open Git Bash or Command Prompt.
+3. Clone the repository and create the `.env` file:
+   ```bash
+   git clone https://github.com/arc53/DocsGPT.git
+   cd DocsGPT
+   echo "API_KEY=Yourkey" > .env
+   echo "VITE_API_STREAMING=true" >> .env
+   ```
+4. Run the following command to start the setup with Docker Compose:
+   ```bash
+   ./run-with-docker-compose.sh
+   ```
+5. Open your web browser and navigate to http://localhost:5173/.
+6. To stop the setup, just press **Ctrl + C** in the Git Bash or Command Prompt terminal.
+
+These steps should help you set up and run the project on Windows using either WSL or Git Bash/Command Prompt. 
+**Important:** Ensure that Docker is installed and properly configured on your Windows system for these steps to work.
+
+
+For WINDOWS:
+
+To run the given setup on Windows, you can use the Windows Subsystem for Linux (WSL) or a Git Bash terminal to execute similar commands. Here are the steps adapted for Windows:
+
+Option 1: Using Windows Subsystem for Linux (WSL):
+
+1. Install WSL if you haven't already. You can follow the official Microsoft documentation for installation: (https://learn.microsoft.com/en-us/windows/wsl/install).
+2. After setting up WSL, open the WSL terminal.
+3. Clone the repository and create the `.env` file:
+   ```bash
+   git clone https://github.com/arc53/DocsGPT.git
+   cd DocsGPT
+   echo "API_KEY=Yourkey" > .env
+   echo "VITE_API_STREAMING=true" >> .env
+   ```
+4. Run the following command to start the setup with Docker Compose:
+    ```bash
+    ./run-with-docker-compose.sh
+    ```
+5. Open your web browser and navigate to http://localhost:5173/.
+6. To stop the setup, just press **Ctrl + C** in the WSL terminal.
+
+Option 2: Using Git Bash or Command Prompt (CMD):
+
+1. Install Git for Windows if you haven't already. You can download it from the official website: (https://gitforwindows.org/).
+2. Open Git Bash or Command Prompt.
+3. Clone the repository and create the `.env` file:
+   ```bash
+   git clone https://github.com/arc53/DocsGPT.git
+   cd DocsGPT
+   echo "API_KEY=Yourkey" > .env
+   echo "VITE_API_STREAMING=true" >> .env
+   ```
+4. Run the following command to start the setup with Docker Compose:
+   ```bash
+   ./run-with-docker-compose.sh
+   ```
+5. Open your web browser and navigate to http://localhost:5173/.
+6. To stop the setup, just press **Ctrl + C** in the Git Bash or Command Prompt terminal.
+
+These steps should help you set up and run the project on Windows using either WSL or Git Bash/Command Prompt. Make sure you have Docker installed and properly configured on your Windows system for this to work.
+
+
+### Chrome Extension
+
+#### Installing the Chrome extension:
+To enhance your DocsGPT experience, you can install the DocsGPT Chrome extension. Here's how:
+
+1. In the DocsGPT GitHub repository, click on the **Code** button and select **Download ZIP**.
+2. Unzip the downloaded file to a location you can easily access.
+3. Open the Google Chrome browser and click on the three dots menu (upper right corner).
+4. Select **More Tools** and then **Extensions**.
+5. Turn on the **Developer mode** switch in the top right corner of the **Extensions page**.
+6. Click on the **Load unpacked** button.
+7. Select the **Chrome** folder where the DocsGPT files have been unzipped (docsgpt-main > extensions > chrome).
+8. The extension should now be added to Google Chrome and can be managed on the Extensions page.
+9. To disable or remove the extension, simply turn off the toggle switch on the extension card or click the **Remove** button.

docs/pages/Deploying/Railway-Deploying.md

@@ -0,0 +1,254 @@
+
+# Self-hosting DocsGPT on Railway
+
+  
+
+Here's a step-by-step guide on how to host DocsGPT on Railway App.
+
+  
+
+At first Clone and set up the project locally to run , test and Modify.
+
+  
+
+### 1. Clone and GitHub SetUp
+
+a. Open Terminal (Windows Shell or Git bash(recommended)).
+
+  
+
+b. Type `git clone https://github.com/arc53/DocsGPT.git`
+
+  
+
+#### Download the package information
+
+  
+
+Once it has finished cloning the repository, it is time to download the package information from all sources. To do so, simply enter the following command:
+
+  
+
+`sudo apt update`
+
+  
+
+#### Install Docker and Docker Compose
+
+  
+
+DocsGPT backend and worker use Python, Frontend is written on React and the whole application is containerized using Docker. To install Docker and Docker Compose, enter the following commands:
+
+  
+
+`sudo apt install docker.io`
+
+  
+
+And now install docker-compose:
+
+  
+
+`sudo apt install docker-compose`
+
+  
+
+#### Access the DocsGPT Folder
+
+  
+
+Enter the following command to access the folder in which the DocsGPT docker-compose file is present.
+
+  
+
+`cd DocsGPT/`
+
+  
+
+#### Prepare the Environment
+
+  
+
+Inside the DocsGPT folder create a `.env` file and copy the contents of `.env_sample` into it.
+
+  
+
+`nano .env`
+
+  
+
+Make sure your `.env` file looks like this:
+
+  
+
+```
+
+OPENAI_API_KEY=(Your OpenAI API key)
+
+VITE_API_STREAMING=true
+
+SELF_HOSTED_MODEL=false
+
+```
+
+  
+
+To save the file, press CTRL+X, then Y, and then ENTER.
+
+  
+
+Next, set the correct IP for the Backend by opening the docker-compose.yml file:
+
+  
+
+`nano docker-compose.yml`
+
+  
+
+And Change line 7 to: `VITE_API_HOST=http://localhost:7091`
+
+to this `VITE_API_HOST=http://<your instance public IP>:7091`
+
+  
+
+This will allow the frontend to connect to the backend.
+
+  
+
+#### Running the Application
+
+  
+
+You're almost there! Now that all the necessary bits and pieces have been installed, it is time to run the application. To do so, use the following command:
+
+  
+
+`sudo docker-compose up -d`
+
+  
+
+Launching it for the first time will take a few minutes to download all the necessary dependencies and build.
+
+  
+
+Once this is done you can go ahead and close the terminal window.
+
+  
+
+### 2. Pushing it to your own Repository
+
+  
+
+a. Create a Repository on your GitHub.
+
+  
+
+b. Open Terminal in the same directory of the Cloned project.
+
+  
+
+c. Type `git init`
+
+  
+
+d. `git add .`
+
+  
+
+e. `git commit -m "first-commit"`
+
+  
+
+f. `git remote add origin <your  repository  link>`
+
+  
+
+g. `git push git push --set-upstream origin master`
+
+Your local files will now be pushed to your GitHub Account. :)
+  
+
+### 3. Create a Railway Account:
+
+  
+
+If you haven't already, create or log in to your railway account do it by visiting [Railway](https://railway.app/)
+
+  
+
+Signup via **GitHub** [Recommended].
+
+  
+
+### 4. Start New Project:
+
+  
+
+a. Open Railway app and Click on "Start New Project."
+
+  
+
+b. Choose any from the list of options available (Recommended "**Deploy from GitHub Repo**")
+
+  
+
+c. Choose the required Repository from your GitHub.
+
+  
+
+d. Configure and allow access to modify your GitHub content from the pop-up window.
+
+  
+
+e. Agree to all the terms and conditions.
+
+  
+
+PS: It may take a few minutes for the account setup to complete.
+
+  
+
+#### You will get A free trial of $5 (use it for trial and then purchase if satisfied and needed)
+
+  
+
+### 5. Connecting to Your newly Railway app with GitHub
+
+  
+
+a. Choose DocsGPT repo from the list of your GitHub repository that you want to deploy now.
+
+  
+
+b. Click on Deploy now.
+
+  
+
+![Three Tabs will be there](/Railway-selection.png)
+
+  
+
+c. Select Variables Tab.
+
+  
+
+d. Upload the env file here that you used for local setup.
+
+  
+
+e. Go to Settings Tab now.
+
+  
+
+f. Go to "Networking" and click on Generate Domain Name, to get the URL of your hosted project.
+
+  
+
+g. You can update the Root directory, build command, installation command as per need.
+
+*[However recommended not the disturb these options and leave them as default if not that needed.]*
+
+  
+  
+
+Your own DocsGPT is now available at the Generated domain URl. :)

docs/pages/Deploying/_meta.json

@@ -0,0 +1,14 @@
+{
+  "Hosting-the-app": {
+    "title": "☁️ Hosting DocsGPT",
+    "href": "/Deploying/Hosting-the-app"
+  },
+  "Quickstart": {
+    "title": "⚡️Quickstart",
+    "href": "/Deploying/Quickstart"
+  },
+  "Railway-Deploying": {
+    "title": "🚂Deploying on Railway",
+    "href": "/Deploying/Railway-Deploying"
+  }
+}

docs/pages/Developing/API-docs.md

@@ -0,0 +1,229 @@
+# API Endpoints Documentation
+
+*Currently, the application provides the following main API endpoints:*
+
+
+### 1. /api/answer 
+**Description:**
+
+This endpoint is used to request answers to user-provided questions.
+
+**Request:**
+
+**Method**: `POST`
+
+**Headers**: Content-Type should be set to `application/json; charset=utf-8`
+
+**Request Body**: JSON object with the following fields:
+* `question` — The user's question.
+* `history`  —  (Optional) Previous conversation history.
+* `api_key`— Your API key.
+* `embeddings_key`  —  Your embeddings key.
+* `active_docs` — The location of active documentation.
+
+Here is a JavaScript Fetch Request example:
+```js
+// answer (POST http://127.0.0.1:5000/api/answer)
+fetch("http://127.0.0.1:5000/api/answer", {
+      "method": "POST",
+      "headers": {
+            "Content-Type": "application/json; charset=utf-8"
+      },
+      "body": JSON.stringify({"question":"Hi","history":null,"api_key":"OPENAI_API_KEY","embeddings_key":"OPENAI_API_KEY",
+      "active_docs": "javascript/.project/ES2015/openai_text-embedding-ada-002/"})
+})
+.then((res) => res.text())
+.then(console.log.bind(console))
+```
+
+**Response**
+
+In response, you will get a JSON document containing the `answer`, `query` and `result`:
+```json
+{
+  "answer": "Hi there! How can I help you?\n",
+  "query": "Hi",
+  "result": "Hi there! How can I help you?\nSOURCES:"
+}
+```
+
+### 2. /api/docs_check
+
+**Description:**
+
+This endpoint will make sure documentation is loaded on the server (just run it every time user is switching between libraries (documentations)).
+
+**Request:**
+
+**Method**: `POST`
+
+**Headers**: Content-Type should be set to `application/json; charset=utf-8`
+
+**Request Body**: JSON object with the field:
+* `docs` — The location of the documentation:
+```js
+// docs_check (POST http://127.0.0.1:5000/api/docs_check)
+fetch("http://127.0.0.1:5000/api/docs_check", {
+      "method": "POST",
+      "headers": {
+            "Content-Type": "application/json; charset=utf-8"
+      },
+      "body": JSON.stringify({"docs":"javascript/.project/ES2015/openai_text-embedding-ada-002/"})
+})
+.then((res) => res.text())
+.then(console.log.bind(console))
+```
+
+**Response:**
+
+In response, you will get a JSON document like this one indicating whether the documentation exists or not:
+```json
+{
+  "status": "exists"
+}
+```
+
+
+### 3. /api/combine
+**Description:**
+
+This endpoint provides information about available vectors and their locations with a simple GET request.
+
+**Request:**
+
+**Method**: `GET`
+
+**Response:**
+
+Response will include:
+* `date`
+* `description`
+* `docLink`
+* `fullName`
+* `language`
+* `location` (local or docshub)
+* `model`
+* `name`
+* `version`
+
+Example of JSON in Docshub and local:
+
+<img width="295" alt="image" src="https://user-images.githubusercontent.com/15183589/224714085-f09f51a4-7a9a-4efb-bd39-798029bb4273.png">
+
+### 4. /api/upload
+**Description:**
+
+This endpoint is used to upload a file that needs to be trained, response is JSON with task ID, which can be used to check on task's progress.
+
+**Request:**
+
+**Method**: `POST`
+
+**Request Body**: A multipart/form-data form with file upload and additional fields, including `user` and `name`.
+
+HTML example:
+
+```html
+<form action="/api/upload" method="post" enctype="multipart/form-data" class="mt-2">
+    <input type="file" name="file" class="py-4" id="file-upload">
+    <input type="text" name="user" value="local" hidden>
+    <input type="text" name="name" placeholder="Name:">
+    
+    <button type="submit" class="py-2 px-4 text-white bg-purple-30 rounded-md hover:bg-purple-30 focus:outline-none focus:ring-2 focus:ring-offset-2 focus:ring-purple-30">
+        Upload
+    </button>
+</form>
+```
+
+**Response:**
+
+JSON response with a status and a task ID that can be used to check the task's progress.
+
+
+### 5. /api/task_status
+**Description:**
+
+This endpoint is used to get the status of a task (`task_id`) from `/api/upload`
+
+**Request:**
+
+**Method**: `GET`
+
+**Query Parameter**: `task_id` (task ID to check)
+
+**Sample JavaScript Fetch Request:**
+```js
+// Task status (Get http://127.0.0.1:5000/api/task_status)
+fetch("http://localhost:5001/api/task_status?task_id=YOUR_TASK_ID", {
+      "method": "GET",
+      "headers": {
+            "Content-Type": "application/json; charset=utf-8"
+      },
+})
+.then((res) => res.text())
+.then(console.log.bind(console))
+```
+
+**Response:**
+
+There are two types of responses:
+
+1. While the task is still running, the 'current' value will show progress from 0 to 100.
+   ```json
+   {
+     "result": {
+       "current": 1
+     },
+     "status": "PROGRESS"
+   }
+   ```
+
+2. When task is completed:
+   ```json
+   {
+     "result": {
+       "directory": "temp",
+       "filename": "install.rst",
+       "formats": [
+         ".rst",
+         ".md",
+         ".pdf"
+       ],
+       "name_job": "somename",
+       "user": "local"
+     },
+     "status": "SUCCESS"
+   }
+   ```
+
+### 6. /api/delete_old
+**Description:**
+
+This endpoint is used to delete old Vector Stores.
+
+**Request:**
+
+**Method**: `GET`
+
+**Query Parameter**: `task_id`
+
+**Sample JavaScript Fetch Request:**
+```js
+// delete_old (GET http://127.0.0.1:5000/api/delete_old)
+fetch("http://localhost:5001/api/delete_old?task_id=YOUR_TASK_ID", {
+      "method": "GET",
+      "headers": {
+            "Content-Type": "application/json; charset=utf-8"
+      },
+})
+.then((res) => res.text())
+.then(console.log.bind(console))
+
+```
+**Response:**
+
+JSON response indicating the status of the operation:
+
+```json
+{ "status": "ok" }
+```

docs/pages/Developing/_meta.json

@@ -0,0 +1,6 @@
+{
+  "API-docs": {
+    "title": "🗂️️ API-docs",
+    "href": "/Developing/API-docs"
+  }
+}

docs/pages/Extensions/Chatwoot-extension.md

@@ -0,0 +1,44 @@
+## Chatwoot Extension Setup Guide
+
+### Step 1: Prepare and Start DocsGPT
+
+- **Launch DocsGPT**: Follow the instructions in our [DocsGPT Wiki](https://github.com/arc53/DocsGPT/wiki) to start DocsGPT. Make sure to load your documentation.
+
+### Step 2: Get Access Token from Chatwoot
+
+- Go to Chatwoot.
+- In your profile settings (located at the bottom left), scroll down and copy the **Access Token**.
+
+### Step 3: Set Up Chatwoot Extension
+
+- Navigate to `/extensions/chatwoot`.
+- Copy the `.env_sample` file and create a new file named `.env`.
+- Fill in the values in the `.env` file as follows:
+
+```env
+docsgpt_url=<Docsgpt_API_URL>
+chatwoot_url=<Chatwoot_URL>
+docsgpt_key=<OpenAI_API_Key or Other_LLM_Key>
+chatwoot_token=<Token from Step 2>
+```
+
+### Step 4: Start the Extension
+
+- Use the command `flask run` to start the extension.
+
+### Step 5: Optional - Extra Validation
+
+- In app.py, uncomment lines 12-13 and 71-75.
+- Add the following lines to your .env file:
+```account_id=(optional) 1
+assignee_id=(optional) 1
+```
+These Chatwoot values help ensure you respond to the correct widget and handle questions assigned to a specific user.
+
+### Stopping Bot Responses for Specific User or Session
+
+- If you want the bot to stop responding to questions for a specific user or session, add a label `human-requested` in your conversation.
+
+### Additional Notes
+
+- For further details on training on other documentation, refer to our [wiki](https://github.com/arc53/DocsGPT/wiki/How-to-train-on-other-documentation).

docs/pages/Extensions/_meta.json

@@ -0,0 +1,10 @@
+{
+  "Chatwoot-extension": {
+    "title": "💬️ Chatwoot Extension",
+    "href": "/Extensions/Chatwoot-extension"
+  },
+  "react-widget": {
+      "title": "🏗️ Widget setup",
+      "href": "/Extensions/react-widget"
+    }
+}

docs/pages/Extensions/react-widget.md

@@ -0,0 +1,47 @@
+### Setting up the DocsGPT Widget in Your React Project
+
+### Introduction:
+The DocsGPT Widget is a powerful tool that allows you to integrate AI-powered documentation assistance into your web applications. This guide will walk you through the installation and usage of the DocsGPT Widget in your React project. Whether you're building a web app or a knowledge base, this widget can enhance your user experience.
+
+### Installation
+First, make sure you have Node.js and npm installed in your project. Then go to your project and install a new dependency: `npm install docsgpt`.
+
+### Usage
+In the file where you want to use the widget, import it and include the CSS file:
+```js
+import { DocsGPTWidget } from "docsgpt";
+import "docsgpt/dist/style.css";
+```
+
+
+Now, you can use the widget in your component like this :
+```jsx
+<DocsGPTWidget
+  apiHost="https://your-docsgpt-api.com"
+  selectDocs="local/docs.zip"
+  apiKey=""
+/>
+```
+DocsGPTWidget takes 3 **props**:
+1. `apiHost` — The URL of your DocsGPT API.
+2. `selectDocs` — The documentation source that you want to use for your widget (e.g. `default` or `local/docs1.zip`).
+3. `apiKey` — Usually, it's empty.
+
+### How to use DocsGPTWidget with [Nextra](https://nextra.site/) (Next.js + MDX)
+Install your widget as described above and then go to your `pages/` folder and create a new file `_app.js` with the following content:
+```js
+import { DocsGPTWidget } from "docsgpt";
+import "docsgpt/dist/style.css";
+
+export default function MyApp({ Component, pageProps }) {
+    return (
+        <>
+            <Component {...pageProps} />
+            <DocsGPTWidget selectDocs="local/docsgpt-sep.zip/"/>
+        </>
+    )
+}
+```  
+
+For more information about React, refer to this [link here](https://react.dev/learn)
+

docs/pages/Guides/Customising-prompts.md

@@ -0,0 +1,27 @@
+# Customizing the Main Prompt
+
+Customizing the main prompt for DocsGPT gives you the ability to tailor the AI's responses to your specific requirements. By modifying the prompt text, you can achieve more accurate and relevant answers. Here's how you can do it:
+
+1. Navigate to `/application/prompts/combine_prompt.txt`.
+
+2. Open the `combine_prompt.txt` file and modify the prompt text to suit your needs. You can experiment with different phrasings and structures to observe how the model responds. The main prompt serves as guidance to the AI model on how to generate responses.
+
+## Example Prompt Modification
+
+**Original Prompt:**
+```markdown
+You are a DocsGPT, friendly and helpful AI assistant by Arc53 that provides help with documents. You give thorough answers with code examples if possible.
+Use the following pieces of context to help answer the users question. If it's not relevant to the question, provide friendly responses.
+You have access to chat history, and can use it to help answer the question.
+When using code examples, use the following format:
+
+(code)
+{summaries}
+```
+
+Feel free to customize the prompt to align it with your specific use case or the kind of responses you want from the AI. For example, you can focus on specific document types, industries, or topics to get more targeted results.
+
+## Conclusion
+
+Customizing the main prompt for DocsGPT allows you to tailor the AI's responses to your unique requirements. Whether you need in-depth explanations, code examples, or specific insights, you can achieve it by modifying the main prompt. Remember to experiment and fine-tune your prompts to get the best results.
+

docs/pages/Guides/How-to-train-on-other-documentation.md

@@ -0,0 +1,63 @@
+## How to train on other documentation
+
+This AI can utilize any documentation, but it requires preparation for similarity search. Follow these steps to get your documentation ready:
+
+**Step 1: Prepare Your Documentation**
+![video-example-of-how-to-do-it](https://d3dg1063dc54p9.cloudfront.net/videos/how-to-vectorise.gif)
+
+Start by going to `/scripts/` folder.
+
+If you open this file, you will see that it uses RST files from the folder to create a `index.faiss` and `index.pkl`. 
+
+It currently uses OPENAI to create the vector store, so make sure your documentation is not too large. Using Pandas cost me around $3-$4.
+
+You can typically find documentation on GitHub in the `docs/` folder for most open-source projects.
+
+### 1. Find documentation in .rst/.md format and create a folder with it in your scripts directory.
+- Name it `inputs/`.
+- Put all your .rst/.md files in there.  
+- The search is recursive, so you don't need to flatten them.
+
+If there are no .rst/.md files, convert whatever you find to a .txt file and feed it. (Don't forget to change the extension in the script).
+
+### Step 2: Configure Your OpenAI API Key
+1. Create a .env file in the scripts/ folder.
+  - Add your OpenAI API key inside: OPENAI_API_KEY=<your-api-key>.
+
+### Step 3: Run the Ingestion Script
+
+`python ingest.py ingest`
+
+It will provide you with the estimated cost.
+
+### Step 4: Move `index.faiss` and `index.pkl` generated in `scripts/output` to `application/` folder. 
+
+
+### Step 5: Run the Web App
+Once you run it, it will use new context relevant to your documentation.Make sure you select default in the dropdown in the UI.
+
+## Customization 
+You can learn more about options while running ingest.py by running:
+  - Make sure you select 'default' from the dropdown in the UI.
+
+## Customization
+You can learn more about options while running ingest.py by executing:
+`python ingest.py --help`
+|              Options             |                                                                                                                                |
+|:--------------------------------:|:------------------------------------------------------------------------------------------------------------------------------:|
+|            **ingest**            | Runs 'ingest' function, converting documentation to Faiss plus Index format                                                  |
+| --dir TEXT                       | List of paths to directory for index creation. E.g. --dir inputs --dir inputs2 [default: inputs]                               |
+| --file TEXT                      | File paths to use (Optional; overrides directory) E.g. --files inputs/1.md --files inputs/2.md                                 |
+| --recursive / --no-recursive     | Whether to recursively search in subdirectories [default: recursive]                                                           |
+| --limit INTEGER                  | Maximum number of files to read                                                                                                |
+| --formats TEXT                   | List of required extensions (list with .) Currently supported: .rst, .md, .pdf, .docx, .csv, .epub, .html [default: .rst, .md] |
+| --exclude / --no-exclude         | Whether to exclude hidden files (dotfiles) [default: exclude]                                                                  |
+| -y, --yes                        | Whether to skip price confirmation                                                                                             |
+| --sample / --no-sample           | Whether to output sample of the first 5 split documents. [default: no-sample]                                                  |
+| --token-check / --no-token-check | Whether to group small documents and split large. Improves semantics. [default: token-check]                                   |
+| --min_tokens INTEGER             | Minimum number of tokens to not group. [default: 150]                                                                          |
+| --max_tokens INTEGER             | Maximum number of tokens to not split. [default: 2000]                                                                         |
+|                                  |                                                                                                                                |
+|            **convert**           | Creates documentation in .md format from source code                                                                           |
+| --dir TEXT                       | Path to a directory with source code. E.g. --dir inputs [default: inputs]                                                      |
+| --formats TEXT                   | Source code language from which to create documentation. Supports py, js and java.  E.g. --formats py [default: py]            |

docs/pages/Guides/How-to-use-different-LLM.md

@@ -0,0 +1,48 @@
+# Setting Up Local Language Models for Your App
+
+Your app relies on two essential models: Embeddings and Text Generation. While OpenAI's default models work seamlessly, you have the flexibility to switch providers or even run the models locally.
+
+## Step 1: Configure Environment Variables
+
+Navigate to the `.env` file or set the following environment variables:
+
+```env
+LLM_NAME=<your Text Generation model>
+API_KEY=<API key for Text Generation>
+EMBEDDINGS_NAME=<LLM for Embeddings>
+EMBEDDINGS_KEY=<API key for Embeddings>
+VITE_API_STREAMING=<true or false>
+```
+
+You can omit the keys if users provide their own. Ensure you set `LLM_NAME` and `EMBEDDINGS_NAME`.
+
+## Step 2: Choose Your Models
+
+**Options for `LLM_NAME`:**
+- openai ([More details](https://platform.openai.com/docs/models))
+- anthropic ([More details](https://docs.anthropic.com/claude/reference/selecting-a-model))
+- manifest ([More details](https://python.langchain.com/docs/integrations/llms/manifest))
+- cohere ([More details](https://docs.cohere.com/docs/llmu))
+- llama.cpp ([More details](https://python.langchain.com/docs/integrations/llms/llamacpp))
+- huggingface (Arc53/DocsGPT-7B by default)
+- sagemaker ([Mode details](https://aws.amazon.com/sagemaker/))
+
+
+Note: for huggingface you can choose any model inside application/llm/huggingface.py or pass llm_name on init, loads   
+
+**Options for `EMBEDDINGS_NAME`:**
+- openai_text-embedding-ada-002
+- huggingface_sentence-transformers/all-mpnet-base-v2
+- huggingface_hkunlp/instructor-large
+- cohere_medium
+
+If you want to be completely local, set `EMBEDDINGS_NAME` to `huggingface_sentence-transformers/all-mpnet-base-v2`. 
+
+For llama.cpp Download the required model and place it in the `models/` folder.
+
+Alternatively, for local Llama setup, run `setup.sh` and choose option 1. The script handles the DocsGPT model addition.
+
+## Step 3: Local Hosting for Privacy
+
+If working with sensitive data, host everything locally by setting `LLM_NAME`, llama.cpp or huggingface, use any model available on Hugging Face, for llama.cpp you need to convert it into gguf format.
+That's it! Your app is now configured for local and private hosting, ensuring optimal security for critical data.

docs/pages/Guides/My-AI-answers-questions-using-external-knowledge.md

@@ -0,0 +1,21 @@
+# Avoiding hallucinations
+
+If your AI uses external knowledge and is not explicit enough, it is ok, because we try to make DocsGPT friendly.
+
+But if you want to adjust it, here is a simple way:-
+
+- Got to `application/prompts/chat_combine_prompt.txt`
+
+- And change it to
+
+
+```
+
+You are a DocsGPT, friendly and helpful AI assistant by Arc53 that provides help with documents. You give thorough answers with code examples, if possible.
+Write an answer for the question below based on the provided context.
+If the context provides insufficient information, reply "I cannot answer".
+You have access to chat history and can use it to help answer the question.
+----------------
+{summaries}
+
+```

docs/pages/Guides/_meta.json

@@ -0,0 +1,18 @@
+{
+  "Customising-prompts": {
+    "title": "🏗️️ Customising Prompts",
+    "href": "/Guides/Customising-prompts"
+  },
+  "How-to-train-on-other-documentation": {
+    "title": "📥 Training on docs",
+    "href": "/Guides/How-to-train-on-other-documentation"
+  },
+  "How-to-use-different-LLM": {
+    "title": "⚙️️ How to use different LLM's",
+    "href": "/Guides/How-to-use-different-LLM"
+  },
+  "My-AI-answers-questions-using-external-knowledge": {
+    "title": "💭️ Avoiding hallucinations",
+    "href": "/Guides/My-AI-answers-questions-using-external-knowledge"
+  }
+}

docs/pages/_app.js

@@ -0,0 +1,11 @@
+import { DocsGPTWidget } from "docsgpt";
+import "docsgpt/dist/style.css";
+
+export default function MyApp({ Component, pageProps }) {
+  return (
+    <>
+      <Component {...pageProps} />
+        <DocsGPTWidget selectDocs="local/docsgpt-sep.zip/"/>
+    </>
+  )
+}

docs/pages/index.mdx

@@ -0,0 +1,37 @@
+---
+title: 'Home'
+---
+import { Cards, Card } from 'nextra/components'
+import deployingGuides from './Deploying/_meta.json';
+import developingGuides from './Developing/_meta.json';
+import extensionGuides from './Extensions/_meta.json';
+import mainGuides from './Guides/_meta.json';
+
+
+
+
+export const allGuides = {
+  ...deployingGuides,
+  ...developingGuides,
+  ...extensionGuides,
+  ...mainGuides,
+};
+
+###  **DocsGPT 🦖**
+
+DocsGPT 🦖 is an innovative open-source tool designed to simplify the retrieval of information from project documentation using advanced GPT models 🤖. Eliminate lengthy manual searches 🔍 and enhance your documentation experience with DocsGPT, and consider contributing to its AI-powered future 🚀.
+
+![video-example-of-docs-gpt](https://d3dg1063dc54p9.cloudfront.net/videos/demov3.gif)
+
+Try it yourself: [https://docsgpt.arc53.com/](https://docsgpt.arc53.com/)
+
+<Cards
+      num={3}
+      children={Object.keys(allGuides).map((key, i) => (
+        <Card
+          key={i}
+          title={allGuides[key].title}
+          href={allGuides[key].href}
+        />
+      ))}
+    />