Find answers from the community

Updated 3 months ago

@Logan M If you wanted to do some sort

If you wanted to do some sort of entity resoltution. What would be the best way to delete and add entities back into the graph? Would it be defining a custom retriever to retrieve nodes you want to combine then coming up with a way to delete and upsert those nodes?
L
C
14 comments
yea you can delete and upsert manually.

Here's he delete method on the graph store for reference
Plain Text
def delete(
    self,
    entity_names: Optional[List[str]] = None,
    relation_names: Optional[List[str]] = None,
    properties: Optional[dict] = None,
    ids: Optional[List[str]] = None,
) -> None:
Hey @Logan M Using a Neo4jPGStore what do you think is the best way to implement "Organizations" or "Groups" inside of a multi-tenant version of this?
Like we have different user uploading different files would it make sense to store all user data in one graph and then parse on something or is there a way to easily scope to specific graphs or sub-graphs etc?
Mmm filtering on the properties I think.
can you do that when calling index.as_retriever(include_text=True).retrieve(query) ?
Hmm, it would be possible with the vector retriever, but I think I missed baking in filtering to the other retrievers.

You'd have to write some custom retrievers (which we have docs for!) or someone needs to make a PR
awesome thanks logan
Also one more thing? Do you know why page_label is refering to the document(chunk) index instead of the page_number when using simple directory reader for loading documents from PDFs?
pretty sure its the page number πŸ‘€
i uploaded a 3 page doc and my returned chunks are 13, 10, 12 which there was a total of 13 documents(chunks)
I'll do more testing just in case tho
if you can give a repro test case, that would be super helpful πŸ™
I am sanity checking rn and if its still an issue will let you know
Add a reply
Sign up and join the conversation on Discord