[Bug]: ImportError - cannot import name 'Vector' from 'azure.search.documents.models'

Bug Description

Hello! I had to update my azure-search-documents to the latest stable version, and unfortunately, the cogsearch.py has compatibility issues.

The problem is caused because in azure-search-documents v11.4.0b10 (2023-10-10) they renamed Vector as VectorizedQuery.

I needed to do two things to make it compatible with the latest version. Firstly, I substituted each Vector call within the cogsearch.py with the following:

# -> line 627
class AzureQueryResultSearchDefault(AzureQueryResultSearchBase):
    def _create_query_vector(self) -> Optional[List[Any]]:
        """Query vector store."""
        from azure.search.documents.models import VectorizedQuery

        if not self._query.query_embedding:
            raise ValueError("Query missing embedding")

        vector = VectorizedQuery(
            vector=self._query.query_embedding,
            k_nearest_neighbors=self._query.similarity_top_k,
            fields=self._field_mapping["embedding"],
        )
        vectors = [vector]
        logger.info("Vector search with supplied embedding")
        return vectors

# -> line 666
class AzureQueryResultSearchSemanticHybrid(AzureQueryResultSearchHybrid):
    def _create_query_vector(self) -> Optional[List[Any]]:
        """Query vector store."""
        from azure.search.documents.models import VectorizedQuery

        if not self._query.query_embedding:
            raise ValueError("Query missing embedding")
        # k is set to 50 to align with the number of accept document in azure semantic reranking model.
        # https://learn.microsoft.com/en-us/azure/search/semantic-search-overview
        vector = VectorizedQuery(
            vector=self._query.query_embedding,
            k_nearest_neighbors=50,
            fields=self._field_mapping["embedding"],
        )
        vectors = [vector]
        logger.info("Vector search with supplied embedding")
        return vectors

Moreover, accordingly to azure-search-documents changelog you need to modify the parameter vector within the _search_client.search into vector_queries (There are two calls within the cogsearch.py code).

I was able to then run it without problems :)

I hope it helps!

Version

0.9.35

Steps to Reproduce

install azure-search-index v 11.4.0.
use llamaindex to access and query a cognitive search index.

Relevant Logs/Tracbacks

No response

run-llama / llama_index