Added latest PGVector features

[danielfornarini]

• Added IVFFlat and HNSW index
• Added Iterative Scan for HNSW and IVFFlat index
• Added Iterative Scan max_scan_tuples and scan_mem_multiplier options for HNSW index
• Added ef_contruction and m options for HNSW index creation
• Added binary quantization
• Added partitioning by collection_id
• Added full text search

[danielfornarini]

@eyurtsev I also fixed all test errors

[kakao-bart-lee]

Please pay attention to this PR @danielfornarini @eyurtsev

[Freezaa9]

Thanks @danielfornarini for this PR we need it.

[adintegra]

FWIW, as a temporary workaround I've been using this PR directly via pip and have encountered no issues so far:

pip install git+https://github.com/langchain-ai/langchain-postgres@c32f6beb108e37aad615ee3cbd4c6bd4a693a76d

[Freezaa9]

I've been testing this pr for a project, and I don't see issues. It feels like this repo is a bit dead ?

[danielfornarini]

I've been testing this pr for a project, and I don't see issues. It feels like this repo is a bit dead ?

Yes, seems like no one maintains it anymore

[amotl]

Hi. I think the project is not hibernating, see GH-149, but I agree it will be sweet if maintainers could pick up this PR. Thanks a stack for your work on this, @danielfornarini!

@ccurme: We are also in dear need for an update, at least to be able to install pgvector 0.3, see GH-156. 🙏

[m-ibraheem-tkxel]

When will this branch be merged and become part of the package?

[Freezaa9]

I'm losing hope, meanwhile cloud sql postgres langchain implementation did implement some of the latest feature of pgvector: https://github.com/googleapis/langchain-google-cloud-sql-pg-python