LLM/DocsGPT
1
0
Fork 0
mirror of https://github.com/arc53/DocsGPT.git synced 2025-11-29 16:43:16 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix tokens for header

Browse Source
This commit is contained in:
Alex
2023-07-24 16:14:08 +01:00
parent b2e86e105d
commit dac76a867f
2 changed files with 10 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
application/parser/token_func.py
View File

@@ -46,6 +46,10 @@ def split_documents(documents: List[Document], max_tokens: int) -> List[Document
docs.append(doc)
else:
header, body = separate_header_and_body(doc.text)
if len(tiktoken.get_encoding("cl100k_base").encode(header)) > max_tokens:
print("header too long, skipping", file=sys.stderr)
body = doc.text
header = ""
num_body_parts = ceil(token_length / max_tokens)
part_length = ceil(len(body) / num_body_parts)
body_parts = [body[i:i + part_length] for i in range(0, len(body), part_length)]

7
scripts/parser/token_func.py
View File

@@ -4,7 +4,7 @@ from typing import List
import tiktoken
from parser.schema.base import Document
import sys
def separate_header_and_body(text):
header_pattern = r"^(.*?\n){3}"
@@ -17,6 +17,7 @@ def separate_header_and_body(text):
def group_documents(documents: List[Document], min_tokens: int, max_tokens: int) -> List[Document]:
docs = []
current_group = None
print("Grouping", len(documents), "documents", file=sys.stderr)
for doc in documents:
doc_len = len(tiktoken.get_encoding("cl100k_base").encode(doc.text))
@@ -46,6 +47,9 @@ def split_documents(documents: List[Document], max_tokens: int) -> List[Document
docs.append(doc)
else:
header, body = separate_header_and_body(doc.text)
if len(tiktoken.get_encoding("cl100k_base").encode(header)) > max_tokens:
body = doc.text
header = ""
num_body_parts = ceil(token_length / max_tokens)
part_length = ceil(len(body) / num_body_parts)
body_parts = [body[i:i + part_length] for i in range(0, len(body), part_length)]
@@ -55,6 +59,7 @@ def split_documents(documents: List[Document], max_tokens: int) -> List[Document
embedding=doc.embedding,
extra_info=doc.extra_info)
docs.append(new_doc)
print("split into", len(docs), "documents", file=sys.stderr)
return docs