MB-71397: Fix the output of Size API of VectorIndex

[CascadingRadium]

• Implement a full recursive size calculation for the vector index, accounting for all owned sub-structures: the index params, the ID mapping, the exclude bitmap etc.
• The underlying FAISS index memory is excluded from the calculation since it is memory-mapped and not RAM-resident.
• Size reported in the VectorOptimize API of bleve :
  • Before:
    • 0.000152587890625 MB
  • After:
    • 171.66588592529297 MB
• pprof output:

$ go tool pprof go_mem.pprof
File: bleve_query
Build ID: 8f0bd7de29dc423c4830da78a30cbbc3ac5025fb
Type: inuse_space
Time: 2026-05-11 17:13:34 IST
Entering interactive mode (type "help" for commands, "o" for options)
(pprof) top
Showing nodes accounting for 174.07MB, 99.14% of 175.57MB total
Dropped 22 nodes (cum <= 0.88MB)
Showing top 10 nodes out of 18

[Copilot]

Pull request overview

This PR adjusts how vector-index memory usage is reported by VectorIndex.Size(), avoiding FAISS Size() calls that are misleading under mmap, and additionally accounting for wrapper-side structures like ID mappings and exclusion bitmaps.

Changes:

• Reworked vectorIndexWrapper.Size() to sum sizes from the FAISS index wrapper, idMapping, and exclusion bitmap, removing the previously cached vecIndexSize.
• Added approximate sizing helpers for idMapping and bitmap using reflect-based static sizing plus backing-slice contributions.
• Updated FAISS index wrapper size() implementations (float32, GPU float32, binary IVF) to avoid calling FAISS Size() under mmap and instead return a reflect-based static size.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
faiss_vector_wrapper.go	Adds reflect-based static sizing, introduces bitmap.size()/idMapping.size(), and updates VectorIndex.Size() to aggregate sizes.
faiss_vector_posting.go	Removes initialization of the cached vecIndexSize in InterpretVectorIndex().
faiss_vector_index_float32.go	Changes FAISS float32 index size() to return reflect-based static size instead of idx.Size().
faiss_vector_index_gpu_float32.go	Changes GPU float32 index size() to return reflect-based static size instead of cpuIdx.Size().
faiss_vector_index_bivf.go	Changes binary IVF index size() to return reflect-based static size instead of binary.Size()+backing.Size().

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 7 out of 7 changed files in this pull request and generated 3 comments.

[Copilot]

Pull request overview

Copilot reviewed 7 out of 7 changed files in this pull request and generated 4 comments.

[Thejas-bhat]

looks good, but can you attach the MB associated with this?

The base branch was changed.