feat(vchordrq): estimate MaxSim planner cost

[HuXinjing]

Summary

• persist the indexed vector/token count in unused MetaTuple padding without changing tuple offsets
• update the count after index build and bulk deletion
• add a MaxSim-specific planner model based on heap rows, indexed tokens, probes, query-token count, filters, and LIMIT
• preserve the existing planner path for every non-MaxSim opfamily

This is the first focused implementation PR proposed in #476. It does not add
an exact reranker, external tensor storage, a network protocol, or GPU code.

Compatibility

The statistic uses the existing 2-byte and 6-byte padding regions in the meta
tuple. Compile-time offset assertions and unit tests protect the on-disk layout.

• Old indexes expose the statistic as None.
• The planner falls back to a configurable average document-token count.
• Build and subsequent vacuum/bulk-delete operations publish an exact count.
• No REINDEX is required solely for compatibility.

The planner query-token and fallback document-token values are configurable
until expression/page-level statistics are available.

Validation

• cargo test -p vchordrq
• cargo test --locked --workspace --exclude vchord --no-fail-fast
• PostgreSQL 16 release extension build through the repository xtask build
  inside a clean PostgreSQL development container
• cargo fmt --all -- --check
• git diff --check

Review notes

The cost constants are intentionally conservative and hardware-independent.
The model prices the eager token search and page aggregation currently
performed by MaxSim. A future optional exact-rerank backend can add its own cost
without changing this candidate-generation model.

Discussion: #476

[github-actions]

All contributors have signed the CLA ✍️ ✅
_{Posted by the CLA Assistant Lite bot.}

[HuXinjing]

Thank you for your submission, we really appreciate it. Like many open-source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution. You can sign the CLA by just posting a Pull Request Comment same as the below format.

I have read the CLA Document and I hereby sign the CLA

You can retrigger this bot by commenting recheck in this Pull Request. Posted by the CLA Assistant Lite bot.

I have read the CLA Document and I hereby sign the CLA

[HuXinjing]

I have read the CLA Document and I hereby sign the CLA