AutoMerging-Retriever: support more document stores

[davidsbatista]

• Pinecone doesn't support it because apparently, one cannot filter by id:
  • https://docs.pinecone.io/guides/data/query-data#querying-by-record-id
  • https://community.pinecone.io/t/does-pinecone-support-filtering-by-vector-id/3039/2
• Weaviate needs to be properly tested
• All the others document stores work, with the exception of Chroma which as a known BUG in the filtering mechanism

[davidsbatista]

• We decided to put PineCone fix on hold - we now the AutoMergingRetriever doesn't work with Pinecone and why, if there's ever such a user case we pick this issue again

[davidsbatista]

Weaviate as a similar issue, the python client as 3 methods to query/fetch:

• fetch_objects()
• fetch_object_by_id()
• fetch_objects_by_ids()
  However, Haystack implementation only uses fetch_objects()

Another note, it seems the original Document id is also stored in the metadata as _original_id, another possible hack would be to catch when the id is part of the filter, and map it to the _original_id in the metadata