diff --git a/README.md b/README.md index 03393629..598cfa1d 100644 --- a/README.md +++ b/README.md @@ -49,11 +49,11 @@ - [x] Manually updating chunks in the app UI (Feb 2025) - [x] Devcontainer for easy development (Feb 2025) - [x] ReACT agent (March 2025) -- [ ] Anthropic Tool compatibility -- [ ] New input box in the conversation menu -- [ ] Add triggerable actions / tools (webhook) +- [ ] Chatbots menu re-design to handle tools, agent types, and more (April 2025) +- [ ] New input box in the conversation menu (April 2025) +- [ ] Anthropic Tool compatibility (April 2025) +- [ ] Add triggerable actions / tools (webhook) (April 2025) - [ ] Add OAuth 2.0 authentication for tools and sources -- [ ] Chatbots menu re-design to handle tools, agent types, and more - [ ] Agent scheduling You can find our full roadmap [here](https://github.com/orgs/arc53/projects/2). Please don't hesitate to contribute or create issues, it helps us improve DocsGPT! @@ -95,13 +95,15 @@ A more detailed [Quickstart](https://docs.docsgpt.cloud/quickstart) is available ./setup.sh ``` -This interactive script will guide you through setting up DocsGPT. It offers four options: using the public API, running locally, connecting to a local inference engine, or using a cloud API provider. The script will automatically configure your `.env` file and handle necessary downloads and installations based on your chosen option. - **For Windows:** -2. **Follow the Docker Deployment Guide:** +2. **Run the PowerShell setup script:** - Please refer to the [Docker Deployment documentation](https://docs.docsgpt.cloud/Deploying/Docker-Deploying) for detailed step-by-step instructions on setting up DocsGPT using Docker. + ```powershell + PowerShell -ExecutionPolicy Bypass -File .\setup.ps1 + ``` + +Either script will guide you through setting up DocsGPT. Four options available: using the public API, running locally, connecting to a local inference engine, or using a cloud API provider. Scripts will automatically configure your `.env` file and handle necessary downloads and installations based on your chosen option. **Navigate to http://localhost:5173/** @@ -110,7 +112,7 @@ To stop DocsGPT, open a terminal in the `DocsGPT` directory and run: ```bash docker compose -f deployment/docker-compose.yaml down ``` -(or use the specific `docker compose down` command shown after running `setup.sh`). +(or use the specific `docker compose down` command shown after running the setup script). > [!Note] > For development environment setup instructions, please refer to the [Development Environment Guide](https://docs.docsgpt.cloud/Deploying/Development-Environment). diff --git a/application/agents/classic_agent.py b/application/agents/classic_agent.py index 4f4df5c5..bf472cd0 100644 --- a/application/agents/classic_agent.py +++ b/application/agents/classic_agent.py @@ -48,10 +48,13 @@ class ClassicAgent(BaseAgent): ): yield {"answer": resp.message.content} else: - completion = self.llm.gen_stream( - model=self.gpt_model, messages=messages, tools=self.tools - ) - for line in completion: + # completion = self.llm.gen_stream( + # model=self.gpt_model, messages=messages, tools=self.tools + # ) + # log type of resp + logger.info(f"Response type: {type(resp)}") + logger.info(f"Response: {resp}") + for line in resp: if isinstance(line, str): yield {"answer": line} diff --git a/application/agents/llm_handler.py b/application/agents/llm_handler.py index 21d972d9..7fe794f8 100644 --- a/application/agents/llm_handler.py +++ b/application/agents/llm_handler.py @@ -15,7 +15,7 @@ class LLMHandler(ABC): @abstractmethod def handle_response(self, agent, resp, tools_dict, messages, attachments=None, **kwargs): pass - + def prepare_messages_with_attachments(self, agent, messages, attachments=None): """ Prepare messages with attachment content if available. @@ -33,15 +33,53 @@ class LLMHandler(ABC): logger.info(f"Preparing messages with {len(attachments)} attachments") - # Check if the LLM has its own custom attachment handling implementation - if hasattr(agent.llm, "prepare_messages_with_attachments") and agent.llm.__class__.__name__ != "BaseLLM": - logger.info(f"Using {agent.llm.__class__.__name__}'s own prepare_messages_with_attachments method") - return agent.llm.prepare_messages_with_attachments(messages, attachments) + supported_types = agent.llm.get_supported_attachment_types() - # Otherwise, append attachment content to the system prompt + supported_attachments = [] + unsupported_attachments = [] + + for attachment in attachments: + mime_type = attachment.get('mime_type') + if not mime_type: + import mimetypes + file_path = attachment.get('path') + if file_path: + mime_type = mimetypes.guess_type(file_path)[0] or 'application/octet-stream' + else: + unsupported_attachments.append(attachment) + continue + + if mime_type in supported_types: + supported_attachments.append(attachment) + else: + unsupported_attachments.append(attachment) + + # Process supported attachments with the LLM's custom method + prepared_messages = messages + if supported_attachments: + logger.info(f"Processing {len(supported_attachments)} supported attachments with {agent.llm.__class__.__name__}'s method") + prepared_messages = agent.llm.prepare_messages_with_attachments(messages, supported_attachments) + + # Process unsupported attachments with the default method + if unsupported_attachments: + logger.info(f"Processing {len(unsupported_attachments)} unsupported attachments with default method") + prepared_messages = self._append_attachment_content_to_system(prepared_messages, unsupported_attachments) + + return prepared_messages + + def _append_attachment_content_to_system(self, messages, attachments): + """ + Default method to append attachment content to the system prompt. + + Args: + messages (list): List of message dictionaries. + attachments (list): List of attachment dictionaries with content. + + Returns: + list: Messages with attachment context added to the system prompt. + """ prepared_messages = messages.copy() - # Build attachment content string attachment_texts = [] for attachment in attachments: logger.info(f"Adding attachment {attachment.get('id')} to context") @@ -122,12 +160,13 @@ class OpenAILLMHandler(LLMHandler): return resp else: - + text_buffer = "" while True: tool_calls = {} for chunk in resp: if isinstance(chunk, str) and len(chunk) > 0: - return + yield chunk + continue elif hasattr(chunk, "delta"): chunk_delta = chunk.delta @@ -206,12 +245,17 @@ class OpenAILLMHandler(LLMHandler): } ) tool_calls = {} + if hasattr(chunk_delta, "content") and chunk_delta.content: + # Add to buffer or yield immediately based on your preference + text_buffer += chunk_delta.content + yield text_buffer + text_buffer = "" if ( hasattr(chunk, "finish_reason") and chunk.finish_reason == "stop" ): - return + return resp elif isinstance(chunk, str) and len(chunk) == 0: continue @@ -227,7 +271,7 @@ class GoogleLLMHandler(LLMHandler): from google.genai import types messages = self.prepare_messages_with_attachments(agent, messages, attachments) - + while True: if not stream: response = agent.llm.gen( @@ -298,6 +342,9 @@ class GoogleLLMHandler(LLMHandler): "content": [function_response_part.to_json_dict()], } ) + else: + tool_call_found = False + yield result if not tool_call_found: return response diff --git a/application/api/answer/routes.py b/application/api/answer/routes.py index 9af1e27e..8f44385b 100644 --- a/application/api/answer/routes.py +++ b/application/api/answer/routes.py @@ -657,6 +657,7 @@ class Answer(Resource): source_log_docs = [] tool_calls = [] stream_ended = False + thought = "" for line in complete_stream( question=question, @@ -679,6 +680,8 @@ class Answer(Resource): source_log_docs = event["source"] elif event["type"] == "tool_calls": tool_calls = event["tool_calls"] + elif event["type"] == "thought": + thought = event["thought"] elif event["type"] == "error": logger.error(f"Error from stream: {event['error']}") return bad_request(500, event["error"]) @@ -710,6 +713,7 @@ class Answer(Resource): conversation_id, question, response_full, + thought, source_log_docs, tool_calls, llm, @@ -876,14 +880,7 @@ def get_attachments_content(attachment_ids, user): ) if attachment_doc: - attachments.append( - { - "id": str(attachment_doc["_id"]), - "content": attachment_doc["content"], - "token_count": attachment_doc.get("token_count", 0), - "path": attachment_doc.get("path", ""), - } - ) + attachments.append(attachment_doc) except Exception as e: logger.error(f"Error retrieving attachment {attachment_id}: {e}") diff --git a/application/api/user/routes.py b/application/api/user/routes.py index a4bc97e7..48e8c08a 100644 --- a/application/api/user/routes.py +++ b/application/api/user/routes.py @@ -2792,25 +2792,25 @@ class StoreAttachment(Resource): user = secure_filename(decoded_token.get("sub")) try: + attachment_id = ObjectId() original_filename = secure_filename(file.filename) - folder_name = original_filename + save_dir = os.path.join( - current_dir, settings.UPLOAD_FOLDER, user, "attachments", folder_name + current_dir, + settings.UPLOAD_FOLDER, + user, + "attachments", + str(attachment_id), ) os.makedirs(save_dir, exist_ok=True) - # Create directory structure: user/attachments/filename/ + file_path = os.path.join(save_dir, original_filename) - # Handle filename conflicts - if os.path.exists(file_path): - name_parts = os.path.splitext(original_filename) - timestamp = datetime.datetime.now().strftime("%Y%m%d%H%M%S") - new_filename = f"{name_parts[0]}_{timestamp}{name_parts[1]}" - file_path = os.path.join(save_dir, new_filename) - original_filename = new_filename - file.save(file_path) - file_info = {"folder": folder_name, "filename": original_filename} + file_info = { + "filename": original_filename, + "attachment_id": str(attachment_id), + } current_app.logger.info(f"Saved file: {file_path}") # Start async task to process single file diff --git a/application/llm/base.py b/application/llm/base.py index 0fce208c..0607159d 100644 --- a/application/llm/base.py +++ b/application/llm/base.py @@ -55,3 +55,12 @@ class BaseLLM(ABC): def _supports_tools(self): raise NotImplementedError("Subclass must implement _supports_tools method") + + def get_supported_attachment_types(self): + """ + Return a list of MIME types supported by this LLM for file uploads. + + Returns: + list: List of supported MIME types + """ + return [] # Default: no attachments supported diff --git a/application/llm/google_ai.py b/application/llm/google_ai.py index 5e33550c..c049eaa2 100644 --- a/application/llm/google_ai.py +++ b/application/llm/google_ai.py @@ -1,5 +1,9 @@ from google import genai from google.genai import types +import os +import logging +import mimetypes +import json from application.llm.base import BaseLLM @@ -9,6 +13,138 @@ class GoogleLLM(BaseLLM): super().__init__(*args, **kwargs) self.api_key = api_key self.user_api_key = user_api_key + self.client = genai.Client(api_key=self.api_key) + + def get_supported_attachment_types(self): + """ + Return a list of MIME types supported by Google Gemini for file uploads. + + Returns: + list: List of supported MIME types + """ + return [ + 'application/pdf', + 'image/png', + 'image/jpeg', + 'image/jpg', + 'image/webp', + 'image/gif' + ] + + def prepare_messages_with_attachments(self, messages, attachments=None): + """ + Process attachments using Google AI's file API for more efficient handling. + + Args: + messages (list): List of message dictionaries. + attachments (list): List of attachment dictionaries with content and metadata. + + Returns: + list: Messages formatted with file references for Google AI API. + """ + if not attachments: + return messages + + prepared_messages = messages.copy() + + # Find the user message to attach files to the last one + user_message_index = None + for i in range(len(prepared_messages) - 1, -1, -1): + if prepared_messages[i].get("role") == "user": + user_message_index = i + break + + if user_message_index is None: + user_message = {"role": "user", "content": []} + prepared_messages.append(user_message) + user_message_index = len(prepared_messages) - 1 + + if isinstance(prepared_messages[user_message_index].get("content"), str): + text_content = prepared_messages[user_message_index]["content"] + prepared_messages[user_message_index]["content"] = [ + {"type": "text", "text": text_content} + ] + elif not isinstance(prepared_messages[user_message_index].get("content"), list): + prepared_messages[user_message_index]["content"] = [] + + files = [] + for attachment in attachments: + mime_type = attachment.get('mime_type') + if not mime_type: + file_path = attachment.get('path') + if file_path: + mime_type = mimetypes.guess_type(file_path)[0] or 'application/octet-stream' + + if mime_type in self.get_supported_attachment_types(): + try: + file_uri = self._upload_file_to_google(attachment) + logging.info(f"GoogleLLM: Successfully uploaded file, got URI: {file_uri}") + files.append({"file_uri": file_uri, "mime_type": mime_type}) + except Exception as e: + logging.error(f"GoogleLLM: Error uploading file: {e}") + if 'content' in attachment: + prepared_messages[user_message_index]["content"].append({ + "type": "text", + "text": f"[File could not be processed: {attachment.get('path', 'unknown')}]" + }) + + if files: + logging.info(f"GoogleLLM: Adding {len(files)} files to message") + prepared_messages[user_message_index]["content"].append({ + "files": files + }) + + return prepared_messages + + def _upload_file_to_google(self, attachment): + """ + Upload a file to Google AI and return the file URI. + + Args: + attachment (dict): Attachment dictionary with path and metadata. + + Returns: + str: Google AI file URI for the uploaded file. + """ + if 'google_file_uri' in attachment: + return attachment['google_file_uri'] + + file_path = attachment.get('path') + if not file_path: + raise ValueError("No file path provided in attachment") + + if not os.path.isabs(file_path): + current_dir = os.path.dirname( + os.path.dirname(os.path.dirname(os.path.abspath(__file__))) + ) + file_path = os.path.join(current_dir, "application", file_path) + + if not os.path.exists(file_path): + raise FileNotFoundError(f"File not found: {file_path}") + + mime_type = attachment.get('mime_type') + if not mime_type: + mime_type = mimetypes.guess_type(file_path)[0] or 'application/octet-stream' + + try: + response = self.client.files.upload(file=file_path) + + file_uri = response.uri + + from application.core.mongo_db import MongoDB + mongo = MongoDB.get_client() + db = mongo["docsgpt"] + attachments_collection = db["attachments"] + if '_id' in attachment: + attachments_collection.update_one( + {"_id": attachment['_id']}, + {"$set": {"google_file_uri": file_uri}} + ) + + return file_uri + except Exception as e: + logging.error(f"Error uploading file to Google AI: {e}") + raise def _clean_messages_google(self, messages): cleaned_messages = [] @@ -26,7 +162,7 @@ class GoogleLLM(BaseLLM): elif isinstance(content, list): for item in content: if "text" in item: - parts.append(types.Part.from_text(item["text"])) + parts.append(types.Part.from_text(text=item["text"])) elif "function_call" in item: parts.append( types.Part.from_function_call( @@ -41,6 +177,14 @@ class GoogleLLM(BaseLLM): response=item["function_response"]["response"], ) ) + elif "files" in item: + for file_data in item["files"]: + parts.append( + types.Part.from_uri( + file_uri=file_data["file_uri"], + mime_type=file_data["mime_type"] + ) + ) else: raise ValueError( f"Unexpected content dictionary format:{item}" @@ -145,12 +289,26 @@ class GoogleLLM(BaseLLM): if tools: cleaned_tools = self._clean_tools_format(tools) config.tools = cleaned_tools - + + # Check if we have both tools and file attachments + has_attachments = False + for message in messages: + for part in message.parts: + if hasattr(part, 'file_data') and part.file_data is not None: + has_attachments = True + break + if has_attachments: + break + + logging.info(f"GoogleLLM: Starting stream generation. Model: {model}, Messages: {json.dumps(messages, default=str)}, Has attachments: {has_attachments}") + response = client.models.generate_content_stream( model=model, contents=messages, config=config, ) + + for chunk in response: if hasattr(chunk, "candidates") and chunk.candidates: for candidate in chunk.candidates: diff --git a/application/llm/openai.py b/application/llm/openai.py index 5e11a072..75bd37e0 100644 --- a/application/llm/openai.py +++ b/application/llm/openai.py @@ -1,4 +1,8 @@ import json +import base64 +import os +import mimetypes +import logging from application.core.settings import settings from application.llm.base import BaseLLM @@ -65,6 +69,15 @@ class OpenAILLM(BaseLLM): ), } ) + elif isinstance(item, dict): + content_parts = [] + if "text" in item: + content_parts.append({"type": "text", "text": item["text"]}) + elif "type" in item and item["type"] == "text" and "text" in item: + content_parts.append(item) + elif "type" in item and item["type"] == "file" and "file" in item: + content_parts.append(item) + cleaned_messages.append({"role": role, "content": content_parts}) else: raise ValueError( f"Unexpected content dictionary format: {item}" @@ -133,6 +146,183 @@ class OpenAILLM(BaseLLM): def _supports_tools(self): return True + def get_supported_attachment_types(self): + """ + Return a list of MIME types supported by OpenAI for file uploads. + + Returns: + list: List of supported MIME types + """ + return [ + 'application/pdf', + 'image/png', + 'image/jpeg', + 'image/jpg', + 'image/webp', + 'image/gif' + ] + + def prepare_messages_with_attachments(self, messages, attachments=None): + """ + Process attachments using OpenAI's file API for more efficient handling. + + Args: + messages (list): List of message dictionaries. + attachments (list): List of attachment dictionaries with content and metadata. + + Returns: + list: Messages formatted with file references for OpenAI API. + """ + if not attachments: + return messages + + prepared_messages = messages.copy() + + # Find the user message to attach file_id to the last one + user_message_index = None + for i in range(len(prepared_messages) - 1, -1, -1): + if prepared_messages[i].get("role") == "user": + user_message_index = i + break + + if user_message_index is None: + user_message = {"role": "user", "content": []} + prepared_messages.append(user_message) + user_message_index = len(prepared_messages) - 1 + + if isinstance(prepared_messages[user_message_index].get("content"), str): + text_content = prepared_messages[user_message_index]["content"] + prepared_messages[user_message_index]["content"] = [ + {"type": "text", "text": text_content} + ] + elif not isinstance(prepared_messages[user_message_index].get("content"), list): + prepared_messages[user_message_index]["content"] = [] + + for attachment in attachments: + mime_type = attachment.get('mime_type') + if not mime_type: + file_path = attachment.get('path') + if file_path: + mime_type = mimetypes.guess_type(file_path)[0] or 'application/octet-stream' + + if mime_type and mime_type.startswith('image/'): + try: + base64_image = self._get_base64_image(attachment) + prepared_messages[user_message_index]["content"].append({ + "type": "image_url", + "image_url": { + "url": f"data:{mime_type};base64,{base64_image}" + } + }) + except Exception as e: + logging.error(f"Error processing image attachment: {e}") + if 'content' in attachment: + prepared_messages[user_message_index]["content"].append({ + "type": "text", + "text": f"[Image could not be processed: {attachment.get('path', 'unknown')}]" + }) + # Handle PDFs using the file API + elif mime_type == 'application/pdf': + try: + file_id = self._upload_file_to_openai(attachment) + + prepared_messages[user_message_index]["content"].append({ + "type": "file", + "file": {"file_id": file_id} + }) + except Exception as e: + logging.error(f"Error uploading PDF to OpenAI: {e}") + if 'content' in attachment: + prepared_messages[user_message_index]["content"].append({ + "type": "text", + "text": f"File content:\n\n{attachment['content']}" + }) + + return prepared_messages + + def _get_base64_image(self, attachment): + """ + Convert an image file to base64 encoding. + + Args: + attachment (dict): Attachment dictionary with path and metadata. + + Returns: + str: Base64-encoded image data. + """ + file_path = attachment.get('path') + if not file_path: + raise ValueError("No file path provided in attachment") + + if not os.path.isabs(file_path): + current_dir = os.path.dirname( + os.path.dirname(os.path.dirname(os.path.abspath(__file__))) + ) + file_path = os.path.join(current_dir, "application", file_path) + + if not os.path.exists(file_path): + raise FileNotFoundError(f"File not found: {file_path}") + + with open(file_path, "rb") as image_file: + return base64.b64encode(image_file.read()).decode('utf-8') + + def _upload_file_to_openai(self, attachment): ##pdfs + """ + Upload a file to OpenAI and return the file_id. + + Args: + attachment (dict): Attachment dictionary with path and metadata. + Expected keys: + - path: Path to the file + - id: Optional MongoDB ID for caching + + Returns: + str: OpenAI file_id for the uploaded file. + """ + import os + import logging + + if 'openai_file_id' in attachment: + return attachment['openai_file_id'] + + file_path = attachment.get('path') + if not file_path: + raise ValueError("No file path provided in attachment") + + if not os.path.isabs(file_path): + current_dir = os.path.dirname( + os.path.dirname(os.path.dirname(os.path.abspath(__file__))) + ) + file_path = os.path.join(current_dir,"application", file_path) + + if not os.path.exists(file_path): + raise FileNotFoundError(f"File not found: {file_path}") + + + try: + with open(file_path, 'rb') as file: + response = self.client.files.create( + file=file, + purpose="assistants" + ) + + file_id = response.id + + from application.core.mongo_db import MongoDB + mongo = MongoDB.get_client() + db = mongo["docsgpt"] + attachments_collection = db["attachments"] + if '_id' in attachment: + attachments_collection.update_one( + {"_id": attachment['_id']}, + {"$set": {"openai_file_id": file_id}} + ) + + return file_id + except Exception as e: + logging.error(f"Error uploading file to OpenAI: {e}") + raise + class AzureOpenAILLM(OpenAILLM): diff --git a/application/parser/file/tabular_parser.py b/application/parser/file/tabular_parser.py index b2dbd193..40971b3c 100644 --- a/application/parser/file/tabular_parser.py +++ b/application/parser/file/tabular_parser.py @@ -73,7 +73,13 @@ class PandasCSVParser(BaseParser): for more information. Set to empty dict by default, this means pandas will try to figure out the separators, table head, etc. on its own. - + + header_period (int): Controls how headers are included in output: + - 0: Headers only at the beginning + - 1: Headers in every row + - N > 1: Headers every N rows + + header_prefix (str): Prefix for header rows. Default is "HEADERS: ". """ def __init__( @@ -83,6 +89,8 @@ class PandasCSVParser(BaseParser): col_joiner: str = ", ", row_joiner: str = "\n", pandas_config: dict = {}, + header_period: int = 20, + header_prefix: str = "HEADERS: ", **kwargs: Any ) -> None: """Init params.""" @@ -91,6 +99,8 @@ class PandasCSVParser(BaseParser): self._col_joiner = col_joiner self._row_joiner = row_joiner self._pandas_config = pandas_config + self._header_period = header_period + self._header_prefix = header_prefix def _init_parser(self) -> Dict: """Init parser.""" @@ -104,15 +114,26 @@ class PandasCSVParser(BaseParser): raise ValueError("pandas module is required to read CSV files.") df = pd.read_csv(file, **self._pandas_config) + headers = df.columns.tolist() + header_row = f"{self._header_prefix}{self._col_joiner.join(headers)}" - text_list = df.apply( - lambda row: (self._col_joiner).join(row.astype(str).tolist()), axis=1 - ).tolist() + if not self._concat_rows: + return df.apply( + lambda row: (self._col_joiner).join(row.astype(str).tolist()), axis=1 + ).tolist() + + text_list = [] + if self._header_period != 1: + text_list.append(header_row) + + for i, row in df.iterrows(): + if (self._header_period > 1 and i > 0 and i % self._header_period == 0): + text_list.append(header_row) + text_list.append(self._col_joiner.join(row.astype(str).tolist())) + if self._header_period == 1 and i < len(df) - 1: + text_list.append(header_row) - if self._concat_rows: - return (self._row_joiner).join(text_list) - else: - return text_list + return self._row_joiner.join(text_list) class ExcelParser(BaseParser): @@ -138,7 +159,13 @@ class ExcelParser(BaseParser): for more information. Set to empty dict by default, this means pandas will try to figure out the table structure on its own. - + + header_period (int): Controls how headers are included in output: + - 0: Headers only at the beginning (default) + - 1: Headers in every row + - N > 1: Headers every N rows + + header_prefix (str): Prefix for header rows. Default is "HEADERS: ". """ def __init__( @@ -148,6 +175,8 @@ class ExcelParser(BaseParser): col_joiner: str = ", ", row_joiner: str = "\n", pandas_config: dict = {}, + header_period: int = 20, + header_prefix: str = "HEADERS: ", **kwargs: Any ) -> None: """Init params.""" @@ -156,6 +185,8 @@ class ExcelParser(BaseParser): self._col_joiner = col_joiner self._row_joiner = row_joiner self._pandas_config = pandas_config + self._header_period = header_period + self._header_prefix = header_prefix def _init_parser(self) -> Dict: """Init parser.""" @@ -169,12 +200,22 @@ class ExcelParser(BaseParser): raise ValueError("pandas module is required to read Excel files.") df = pd.read_excel(file, **self._pandas_config) + headers = df.columns.tolist() + header_row = f"{self._header_prefix}{self._col_joiner.join(headers)}" + + if not self._concat_rows: + return df.apply( + lambda row: (self._col_joiner).join(row.astype(str).tolist()), axis=1 + ).tolist() + + text_list = [] + if self._header_period != 1: + text_list.append(header_row) - text_list = df.apply( - lambda row: (self._col_joiner).join(row.astype(str).tolist()), axis=1 - ).tolist() - - if self._concat_rows: - return (self._row_joiner).join(text_list) - else: - return text_list \ No newline at end of file + for i, row in df.iterrows(): + if (self._header_period > 1 and i > 0 and i % self._header_period == 0): + text_list.append(header_row) + text_list.append(self._col_joiner.join(row.astype(str).tolist())) + if self._header_period == 1 and i < len(df) - 1: + text_list.append(header_row) + return self._row_joiner.join(text_list) \ No newline at end of file diff --git a/application/requirements.txt b/application/requirements.txt index 3a7f6051..760078ae 100644 --- a/application/requirements.txt +++ b/application/requirements.txt @@ -41,7 +41,7 @@ lxml==5.3.1 markupsafe==3.0.2 marshmallow==3.26.1 mpmath==1.3.0 -multidict==6.1.0 +multidict==6.3.2 mypy-extensions==1.0.0 networkx==3.4.2 numpy==2.2.1 diff --git a/application/worker.py b/application/worker.py index 23ff0422..bbd422ac 100755 --- a/application/worker.py +++ b/application/worker.py @@ -328,34 +328,30 @@ def attachment_worker(self, directory, file_info, user): """ import datetime import os + import mimetypes from application.utils import num_tokens_from_string mongo = MongoDB.get_client() db = mongo["docsgpt"] attachments_collection = db["attachments"] - job_name = file_info["folder"] - logging.info(f"Processing attachment: {job_name}", extra={"user": user, "job": job_name}) + filename = file_info["filename"] + attachment_id = file_info["attachment_id"] + + logging.info(f"Processing attachment: {attachment_id}/{filename}", extra={"user": user}) self.update_state(state="PROGRESS", meta={"current": 10}) - folder_name = file_info["folder"] - filename = file_info["filename"] - file_path = os.path.join(directory, filename) - - logging.info(f"Processing file: {file_path}", extra={"user": user, "job": job_name}) - if not os.path.exists(file_path): - logging.warning(f"File not found: {file_path}", extra={"user": user, "job": job_name}) - return {"error": "File not found"} + logging.warning(f"File not found: {file_path}", extra={"user": user}) + raise FileNotFoundError(f"File not found: {file_path}") try: reader = SimpleDirectoryReader( input_files=[file_path] ) - documents = reader.load_data() self.update_state(state="PROGRESS", meta={"current": 50}) @@ -364,33 +360,37 @@ def attachment_worker(self, directory, file_info, user): content = documents[0].text token_count = num_tokens_from_string(content) - file_path_relative = f"{user}/attachments/{folder_name}/{filename}" + file_path_relative = f"{settings.UPLOAD_FOLDER}/{user}/attachments/{attachment_id}/{filename}" - attachment_id = attachments_collection.insert_one({ + mime_type = mimetypes.guess_type(file_path)[0] or 'application/octet-stream' + + doc_id = ObjectId(attachment_id) + attachments_collection.insert_one({ + "_id": doc_id, "user": user, "path": file_path_relative, "content": content, "token_count": token_count, + "mime_type": mime_type, "date": datetime.datetime.now(), - }).inserted_id + }) logging.info(f"Stored attachment with ID: {attachment_id}", - extra={"user": user, "job": job_name}) + extra={"user": user}) self.update_state(state="PROGRESS", meta={"current": 100}) return { - "attachment_id": str(attachment_id), "filename": filename, - "folder": folder_name, "path": file_path_relative, - "token_count": token_count + "token_count": token_count, + "attachment_id": attachment_id, + "mime_type": mime_type } else: logging.warning("No content was extracted from the file", - extra={"user": user, "job": job_name}) - return {"error": "No content was extracted from the file"} + extra={"user": user}) + raise ValueError("No content was extracted from the file") except Exception as e: - logging.error(f"Error processing file {filename}: {e}", - extra={"user": user, "job": job_name}, exc_info=True) - return {"error": f"Error processing file: {str(e)}"} + logging.error(f"Error processing file {filename}: {e}", extra={"user": user}, exc_info=True) + raise diff --git a/docs/pages/quickstart.mdx b/docs/pages/quickstart.mdx index cef1cd68..11a17895 100644 --- a/docs/pages/quickstart.mdx +++ b/docs/pages/quickstart.mdx @@ -73,9 +73,44 @@ The easiest way to launch DocsGPT is using the provided `setup.sh` script. This ## Launching DocsGPT (Windows) -For Windows users, we recommend following the Docker deployment guide for detailed instructions. Please refer to the [Docker Deployment documentation](/Deploying/Docker-Deploying) for step-by-step instructions on setting up DocsGPT on Windows using Docker. +For Windows users, we provide a PowerShell script that offers the same functionality as the macOS/Linux setup script. -**Important for Windows:** Ensure Docker Desktop is installed and running correctly on your Windows system before proceeding. +**Steps:** + +1. **Download the DocsGPT Repository:** + + First, you need to download the DocsGPT repository to your local machine. You can do this using Git: + + ```powershell + git clone https://github.com/arc53/DocsGPT.git + cd DocsGPT + ``` + +2. **Run the `setup.ps1` script:** + + Execute the PowerShell setup script: + + ```powershell + PowerShell -ExecutionPolicy Bypass -File .\setup.ps1 + ``` + +3. **Follow the interactive setup:** + + Just like the Linux/macOS script, the PowerShell script will guide you through setting DocsGPT. + The script will handle environment configuration and start DocsGPT based on your selections. + +4. **Access DocsGPT in your browser:** + + Once the setup is complete and Docker containers are running, navigate to [http://localhost:5173/](http://localhost:5173/) in your web browser to access the DocsGPT web application. + +5. **Stopping DocsGPT:** + + To stop DocsGPT run the Docker Compose down command displayed at the end of the setup script's execution. + +**Important for Windows:** Ensure Docker Desktop is installed and running correctly on your Windows system before proceeding. The script will attempt to start Docker if it's not running, but you may need to start it manually if there are issues. + +**Alternative Method:** +If you prefer a more manual approach, you can follow our [Docker Deployment documentation](/Deploying/Docker-Deploying) for detailed instructions on setting up DocsGPT on Windows using Docker commands directly. ## Advanced Configuration diff --git a/frontend/src/components/MessageInput.tsx b/frontend/src/components/MessageInput.tsx index c6a2d524..e7ef7f9d 100644 --- a/frontend/src/components/MessageInput.tsx +++ b/frontend/src/components/MessageInput.tsx @@ -4,13 +4,20 @@ import { useDispatch, useSelector } from 'react-redux'; import endpoints from '../api/endpoints'; import userService from '../api/services/userService'; +import AlertIcon from '../assets/alert.svg'; import ClipIcon from '../assets/clip.svg'; +import ExitIcon from '../assets/exit.svg'; import PaperPlane from '../assets/paper_plane.svg'; import SourceIcon from '../assets/source.svg'; import SpinnerDark from '../assets/spinner-dark.svg'; import Spinner from '../assets/spinner.svg'; import ToolIcon from '../assets/tool.svg'; -import { setAttachments } from '../conversation/conversationSlice'; +import { + addAttachment, + removeAttachment, + selectAttachments, + updateAttachment, +} from '../conversation/conversationSlice'; import { useDarkTheme } from '../hooks'; import { ActiveState } from '../models/misc'; import { @@ -18,6 +25,7 @@ import { selectToken, } from '../preferences/preferenceSlice'; import Upload from '../upload/Upload'; +import { getOS, isTouchDevice } from '../utils/browserUtils'; import SourcesPopup from './SourcesPopup'; import ToolsPopup from './ToolsPopup'; @@ -56,13 +64,35 @@ export default function MessageInput({ const [isToolsPopupOpen, setIsToolsPopupOpen] = useState(false); const [uploadModalState, setUploadModalState] = useState('INACTIVE'); - const [uploads, setUploads] = useState([]); const selectedDocs = useSelector(selectSelectedDocs); const token = useSelector(selectToken); + const attachments = useSelector(selectAttachments); const dispatch = useDispatch(); + const browserOS = getOS(); + const isTouch = isTouchDevice(); + + useEffect(() => { + const handleKeyDown = (event: KeyboardEvent) => { + if ( + ((browserOS === 'win' || browserOS === 'linux') && + event.ctrlKey && + event.key === 'k') || + (browserOS === 'mac' && event.metaKey && event.key === 'k') + ) { + event.preventDefault(); + setIsSourcesPopupOpen(!isSourcesPopupOpen); + } + }; + + document.addEventListener('keydown', handleKeyDown); + return () => { + document.removeEventListener('keydown', handleKeyDown); + }; + }, [browserOS]); + const handleFileAttachment = (e: React.ChangeEvent) => { if (!e.target.files || e.target.files.length === 0) return; @@ -73,23 +103,23 @@ export default function MessageInput({ const apiHost = import.meta.env.VITE_API_HOST; const xhr = new XMLHttpRequest(); - const uploadState: UploadState = { - taskId: '', + const newAttachment = { fileName: file.name, progress: 0, - status: 'uploading', + status: 'uploading' as const, + taskId: '', }; - setUploads((prev) => [...prev, uploadState]); - const uploadIndex = uploads.length; + dispatch(addAttachment(newAttachment)); xhr.upload.addEventListener('progress', (event) => { if (event.lengthComputable) { const progress = Math.round((event.loaded / event.total) * 100); - setUploads((prev) => - prev.map((upload, index) => - index === uploadIndex ? { ...upload, progress } : upload, - ), + dispatch( + updateAttachment({ + taskId: newAttachment.taskId, + updates: { progress }, + }), ); } }); @@ -97,34 +127,35 @@ export default function MessageInput({ xhr.onload = () => { if (xhr.status === 200) { const response = JSON.parse(xhr.responseText); - console.log('File uploaded successfully:', response); - if (response.task_id) { - setUploads((prev) => - prev.map((upload, index) => - index === uploadIndex - ? { ...upload, taskId: response.task_id, status: 'processing' } - : upload, - ), + dispatch( + updateAttachment({ + taskId: newAttachment.taskId, + updates: { + taskId: response.task_id, + status: 'processing', + progress: 10, + }, + }), ); } } else { - setUploads((prev) => - prev.map((upload, index) => - index === uploadIndex ? { ...upload, status: 'failed' } : upload, - ), + dispatch( + updateAttachment({ + taskId: newAttachment.taskId, + updates: { status: 'failed' }, + }), ); - console.error('Error uploading file:', xhr.responseText); } }; xhr.onerror = () => { - setUploads((prev) => - prev.map((upload, index) => - index === uploadIndex ? { ...upload, status: 'failed' } : upload, - ), + dispatch( + updateAttachment({ + taskId: newAttachment.taskId, + updates: { status: 'failed' }, + }), ); - console.error('Network error during file upload'); }; xhr.open('POST', `${apiHost}${endpoints.USER.STORE_ATTACHMENT}`); @@ -134,69 +165,63 @@ export default function MessageInput({ }; useEffect(() => { - let timeoutIds: number[] = []; - const checkTaskStatus = () => { - const processingUploads = uploads.filter( - (upload) => upload.status === 'processing' && upload.taskId, + const processingAttachments = attachments.filter( + (att) => att.status === 'processing' && att.taskId, ); - processingUploads.forEach((upload) => { + processingAttachments.forEach((attachment) => { userService - .getTaskStatus(upload.taskId, null) + .getTaskStatus(attachment.taskId!, null) .then((data) => data.json()) .then((data) => { - console.log('Task status:', data); - - setUploads((prev) => - prev.map((u) => { - if (u.taskId !== upload.taskId) return u; - - if (data.status === 'SUCCESS') { - return { - ...u, + if (data.status === 'SUCCESS') { + dispatch( + updateAttachment({ + taskId: attachment.taskId!, + updates: { status: 'completed', progress: 100, - attachment_id: data.result?.attachment_id, + id: data.result?.attachment_id, token_count: data.result?.token_count, - }; - } else if (data.status === 'FAILURE') { - return { ...u, status: 'failed' }; - } else if (data.status === 'PROGRESS' && data.result?.current) { - return { ...u, progress: data.result.current }; - } - return u; - }), - ); - - if (data.status !== 'SUCCESS' && data.status !== 'FAILURE') { - const timeoutId = window.setTimeout( - () => checkTaskStatus(), - 2000, + }, + }), + ); + } else if (data.status === 'FAILURE') { + dispatch( + updateAttachment({ + taskId: attachment.taskId!, + updates: { status: 'failed' }, + }), + ); + } else if (data.status === 'PROGRESS' && data.result?.current) { + dispatch( + updateAttachment({ + taskId: attachment.taskId!, + updates: { progress: data.result.current }, + }), ); - timeoutIds.push(timeoutId); } }) - .catch((error) => { - console.error('Error checking task status:', error); - setUploads((prev) => - prev.map((u) => - u.taskId === upload.taskId ? { ...u, status: 'failed' } : u, - ), + .catch(() => { + dispatch( + updateAttachment({ + taskId: attachment.taskId!, + updates: { status: 'failed' }, + }), ); }); }); }; - if (uploads.some((upload) => upload.status === 'processing')) { - const timeoutId = window.setTimeout(checkTaskStatus, 2000); - timeoutIds.push(timeoutId); - } + const interval = setInterval(() => { + if (attachments.some((att) => att.status === 'processing')) { + checkTaskStatus(); + } + }, 2000); - return () => { - timeoutIds.forEach((id) => clearTimeout(id)); - }; - }, [uploads]); + return () => clearInterval(interval); + }, [attachments, dispatch]); const handleInput = () => { if (inputRef.current) { @@ -230,42 +255,56 @@ export default function MessageInput({ }; const handleSubmit = () => { - const completedAttachments = uploads - .filter((upload) => upload.status === 'completed' && upload.attachment_id) - .map((upload) => ({ - fileName: upload.fileName, - id: upload.attachment_id as string, - })); - - dispatch(setAttachments(completedAttachments)); - onSubmit(); }; return (
- {uploads.map((upload, index) => ( + {attachments.map((attachment, index) => (
- {upload.fileName} + {attachment.fileName} - {upload.status === 'completed' && ( - ✓ + {attachment.status === 'completed' && ( + )} - {upload.status === 'failed' && ( - ✗ + {attachment.status === 'failed' && ( + Upload failed )} - {(upload.status === 'uploading' || - upload.status === 'processing') && ( + {(attachment.status === 'uploading' || + attachment.status === 'processing') && (
+ {/* Background circle */} @@ -317,19 +356,29 @@ export default function MessageInput({ {showSourceButton && ( )} diff --git a/frontend/src/components/SourcesPopup.tsx b/frontend/src/components/SourcesPopup.tsx index 9f8ce8ce..088b3eb1 100644 --- a/frontend/src/components/SourcesPopup.tsx +++ b/frontend/src/components/SourcesPopup.tsx @@ -207,7 +207,7 @@ export default function SourcesPopup({ )} -
+ ); -} \ No newline at end of file +} diff --git a/frontend/src/conversation/ConversationBubble.tsx b/frontend/src/conversation/ConversationBubble.tsx index ec244087..a241b2d3 100644 --- a/frontend/src/conversation/ConversationBubble.tsx +++ b/frontend/src/conversation/ConversationBubble.tsx @@ -57,7 +57,6 @@ const ConversationBubble = forwardRef< updated?: boolean, index?: number, ) => void; - attachments?: { fileName: string; id: string }[]; } >(function ConversationBubble( { @@ -72,7 +71,6 @@ const ConversationBubble = forwardRef< retryBtn, questionNumber, handleUpdatedQuestionSubmission, - attachments, }, ref, ) { @@ -99,36 +97,6 @@ const ConversationBubble = forwardRef< handleUpdatedQuestionSubmission?.(editInputBox, true, questionNumber); }; let bubble; - const renderAttachments = () => { - if (!attachments || attachments.length === 0) return null; - - return ( -
- {attachments.map((attachment, index) => ( -
- - - - {attachment.fileName} -
- ))} -
- ); - }; if (type === 'QUESTION') { bubble = (
{message}
- {renderAttachments()}