Add table quantization option for pgvector

[lucagiac81]

This PR adds a table quantization option for pgvector.

• The table_quantization_type option controls quantization of data as it is loaded into the database.
• This is in addition to the existing quantization_type option which controls quantization of data in an index.

Quantizing data as it is loaded enables storage savings if the original float vectors do not need to be preserved for other purposes. For example, halfvec quantization could be used in both the table and index, reducing the size of the table and eliminating the need for conversions.

table_quantization_type supports the same quantization options as quantization type: none, halfvec, bit.
Most combinations of the two options are supported, with some exceptions.

• If table_quantization_type = bit, then quantization type is forced to bit as well, as it does not make sense to convert bit vectors in the table to vector or halfvec in an index.
• Reranking is only supported if quantization_type = bit and table_quantization_type is vector or halfvec.

[lucagiac81]

Rebased on latest main.

Below are the configurations verified with runs from cli.

• Case: Performance1536D500K
• Table size: pg_vector_collection size reported by psql \d+
• Index size: pgvector_index size reported by psql \di+

Table quantization	Index quantization	Table size (MB)	Index size (MB)
none (vector)	none (vector)	3908	3906
none (vector)	halfvec	3908	1953
none (vector)	bit	3908	239
halfvec	none (vector)	1954	3906
halfvec	halfvec	1954	1953
halfvec	bit	1954	239
bit	bit	115	239

[lucagiac81]

Formatting checks are failing for files not modified by this PR. Would you like me to fix them in this PR? If they are fixed in a separate PR, I will rebase on it.

PYTHONPATH=`pwd` python3 -m black vectordb_bench --check
would reformat /home/runner/work/VectorDBBench/VectorDBBench/vectordb_bench/backend/clients/milvus/config.py
would reformat /home/runner/work/VectorDBBench/VectorDBBench/vectordb_bench/backend/clients/milvus/cli.py

[lucagiac81]

/assign @XuanYang-cn

[sre-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: akashsha1, alwayslove2013, lucagiac81
To complete the pull request process, please assign xuanyang-cn after the PR has been reviewed.
You can assign the PR to them by writing /assign @xuanyang-cn in a comment when ready.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[alwayslove2013]

@lucagiac81 I will check it.

Formatting checks are failing for files not modified by this PR.