gmakstutis/cognee
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Fix paragraph_ids handling

Browse source
This commit is contained in:
Leon Luithlen 2024-11-14 16:47:51 +01:00
parent 7cf8c74cf9
commit 15420dd864
1 changed files with 7 additions and 9 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

16
cognee/tasks/chunks/chunk_by_paragraph.py
View file

@ -13,10 +13,8 @@ def chunk_by_paragraph(data: str, paragraph_length: int = 1024, batch_paragraphs
paragraph_ids = []
last_cut_type = None
for paragraph_id, _, sentence, word_count, end_type in chunk_by_sentence(data, maximum_length=paragraph_length):
for paragraph_id, sentence, word_count, end_type in chunk_by_sentence(data, maximum_length=paragraph_length):
# Check if this sentence would exceed length limit
paragraph_ids.append(paragraph_id)
if current_word_count > 0 and current_word_count + word_count > paragraph_length:
# Yield current chunk
chunk_dict = {
@ -32,13 +30,13 @@ def chunk_by_paragraph(data: str, paragraph_length: int = 1024, batch_paragraphs
# Start new chunk with current sentence
paragraph_ids = []
current_chunk = sentence
current_word_count = word_count
current_chunk = ""
current_word_count = 0
chunk_index += 1
else:
# Just concatenate directly - no space handling
current_chunk += sentence
current_word_count += word_count
paragraph_ids.append(paragraph_id)
current_chunk += sentence
current_word_count += word_count
# Handle end of paragraph
if end_type in ("paragraph_end", "sentence_cut") and not batch_paragraphs: