docs/how_to/self_query/

[giscus[bot]]

docs/how_to/self_query/

Head to Integrations for documentation on vector stores with built-in support for self-querying.

https://python.langchain.com/docs/how_to/self_query/

[ChellShort]

How can we define our own allowed comparators for search?

[marcnaweb]

The answer for the composite filter

retriever.invoke("What's a highly rated (above 8.5) science fiction film?")

[Document(page_content='A psychologist / detective gets lost in a series of dreams within dreams within dreams and Inception reused the idea', metadata={'director': 'Satoshi Kon', 'rating': 8.6, 'year': 2006}),
 Document(page_content='Three men walk into the Zone, three men walk out of the Zone', metadata={'director': 'Andrei Tarkovsky', 'genre': 'thriller', 'rating': 9.9, 'year': 1979})]

is a bit unexpected: there is no highly rated movie that is a science fiction.

Even if we can, indeed, give the benefit of doubt on "A psychologist / detective gets lost in.. " the "Three men walk into the Zone" movie is recorded as "genre": "thriller".

As such, other LLMs might give an (expected) empty result for this query, since there is no "highly rated" movie that is also a "science fiction" --> "rating" > 8.5 AND "genre" == "science fiction" ".