github/langchain
1
0
Fork 0
mirror of https://github.com/hwchase17/langchain.git synced 2025-06-22 06:39:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

for ChatVectorDBChain, add top_k_docs_for_context to allow control how many chunks of context will be retrieved (#1155)

Browse Source 
given that we allow user define chunk size, think it would be useful for
user to define how many chunks of context will be retrieved.
This commit is contained in:
Ji 2023-02-19 20:48:23 -08:00 committed by GitHub
parent 955c89fccb
commit ed37fbaeff
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 7 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
langchain/chains/chat_vector_db/base.py
View File

@ -32,6 +32,7 @@ class ChatVectorDBChain(Chain, BaseModel):
question_generator: LLMChain
output_key: str = "answer"
return_source_documents: bool = False
top_k_docs_for_context: int = 4
"""Return the source documents."""
@property
@ -88,7 +89,9 @@ class ChatVectorDBChain(Chain, BaseModel):
)
else:
new_question = question
docs = self.vectorstore.similarity_search(new_question, k=4, **vectordbkwargs)
docs = self.vectorstore.similarity_search(
new_question, k=self.top_k_docs_for_context, **vectordbkwargs
)
new_inputs = inputs.copy()
new_inputs["question"] = new_question
new_inputs["chat_history"] = chat_history_str
@ -109,7 +112,9 @@ class ChatVectorDBChain(Chain, BaseModel):
else:
new_question = question
# TODO: This blocks the event loop, but it's not clear how to avoid it.
docs = self.vectorstore.similarity_search(new_question, k=4, **vectordbkwargs)
docs = self.vectorstore.similarity_search(
new_question, k=self.top_k_docs_for_context, **vectordbkwargs
)
new_inputs = inputs.copy()
new_inputs["question"] = new_question
new_inputs["chat_history"] = chat_history_str