Skip to content

Only situate if chunk char length is less than specified #1289

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Only situate if chunk char length is less than specified #1289

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
0a66c0b
only situate if chunk char length is less
ahuang11 Jun 17, 2025
dec7ab2
Merge branch 'main' into situate_character_length
ahuang11 Jun 17, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
25 changes: 19 additions & 6 deletions lumen/ai/vector_store.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -82,11 +82,16 @@ class VectorStore(LLMUser):
doc="List of metadata keys to exclude when creating the embeddings.",
)

situate = param.Boolean(
situate = param.ClassSelector(
class_=(int, bool),
default=False,
doc="""
Whether to insert a `llm_context` key in the metadata containing
contextual about the chunks.""",
Controls whether to add a `llm_context` key to the metadata
with contextual information about the chunks.
If set to True, all chunks will be situated.
If set to an integer, only chunks with a character
count less than or equal to this value will be situated.
If set to False, no chunks will be situated.""",
Comment on lines +89 to +94
Copy link
Collaborator

@amaloney amaloney Jun 19, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would like an example in the doc, because I do not fully understand what is being added

)

def __init__(self, **params):
Expand Down Expand Up @@ -204,14 +209,16 @@ def _chunk_text(self, text: str, chunk_size: int | None = None, chunk_func: Call

return chunks

async def should_situate_chunk(self, chunk: str) -> bool:
async def should_situate_chunk(self, chunk: str, max_characters: int | None = None) -> bool:
"""
Determine whether a chunk should be situated based on its content.

Parameters
----------
chunk: str
The chunk text to evaluate
max_characters: int | None
The maximum character count for situating chunks

Returns
-------
Expand All @@ -220,6 +227,8 @@ async def should_situate_chunk(self, chunk: str) -> bool:
"""
if not self.llm:
return self.situate
elif max_characters is not None and len(chunk) > max_characters:
return False

try:
# Use a user message to avoid conflicts with system instructions
Expand All @@ -235,7 +244,7 @@ async def add(
self,
items: list[dict],
force_ids: list[int] | None = None,
situate: bool | None = None,
situate: int | bool | None = None,
) -> list[int]:
"""
Add items to the vector store.
Expand All @@ -260,6 +269,10 @@ async def add(

# Use the provided situate parameter or fall back to the class default
use_situate = self.situate if situate is None else situate
if isinstance(use_situate, int) and use_situate > 0:
max_characters = use_situate
else:
max_characters = None

for item in items:
text = item["text"]
Expand All @@ -281,7 +294,7 @@ async def add(
if should_situate and self.llm:
previous_context = None # Start with no previous context
for chunk in content_chunks:
needs_context = await self.should_situate_chunk(chunk)
needs_context = await self.should_situate_chunk(chunk, max_characters)
if not needs_context:
continue

Expand Down
Loading