feat: add scorer-kit

[ovitrif]

Context

This PR moves the offline scorer-file tooling previously introduced via synonymdev/ldk-node#79 into the scorer repo as scorer-kit and enhances its functionality with merging and decoding capabilities.

The tool is intended for local/operator workflows around serialized LDK ChannelLiquidities files: inspect them, decode them to JSON, compare two snapshots, validate bytes, and merge multiple scorer files into a new scorer binary.

The scorer-kit binary uses the scorer merge/diagnostics API exposed in synonymdev/rust-lightning#2 via an aliased dependency pinned to the commit in the branch of that branch, so the existing scorer node/prober dependency graph remains on its current LDK crate set.

What Changed

• Adds src/bin/scorer-kit.rs with subcommands:
  • inspect
  • decode
  • compare
  • merge
  • validate
• Adds policy-driven duplicate handling for merge:
  • default: richer-history
  • optional: prefer-first, prefer-last, combine, newer
• Adds merge audit reports with duplicate decisions and summary counts.
• Adds sanitized decoded JSON fixtures under test_data/scorer-kit/.
• Documents common scorer-kit flows in the README.

Current Sample Result

Using the current two source samples, scorer-kit merge produced a 16,609-entry merged scorer file. The audit report recorded 1,184 duplicate short-channel-id decisions: 389 kept from the existing input, 795 replaced by the incoming input, and 0 blindly combined under the default richer-history policy.

Notes

The default merge policy avoids blindly combining duplicate short-channel-id entries. Unique entries are preserved, and overlaps choose the entry with stronger historical signal unless another policy is explicitly requested.

Test Plan

• cargo check --bin scorer-kit
• cargo check
• cargo test
• scorer-kit validate <source-a.bin> <source-b.bin>
• scorer-kit inspect <source-a.bin> --label source-a --top 5
• scorer-kit inspect <source-b.bin> --label source-b --top 5
• scorer-kit compare <source-a.bin> <source-b.bin> --left-label source-a --right-label source-b
• scorer-kit compare <source-a.bin> <source-b.bin> --left-label source-a --right-label source-b --output csv
• scorer-kit merge <source-b.bin> <source-a.bin> --label source-b --label source-a --output target/scorer-kit/merged.bin --report target/scorer-kit/merge-report.json --policy richer-history
• scorer-kit validate target/scorer-kit/merged.bin
• sensitive term scan over staged files and PR body