refactor(core): decode SCIP into flat records

[fishmingyu]

Summary

Separate provider-neutral decoded SCIP records from graph materialization so later transports and indexes can consume one stable decode boundary.

Closes #559.

Changes

• add DecodedRecords as the flat C++ decode result
• make SCIPDecoderBase decode records before optionally materializing CodeGraph
• preserve merge behavior and Python post-processing at the records boundary
• add structured decode profiling without unconditional stdout output
• validate indexed-edge endpoints and cover records/graph parity
• document the refactor(core): decode SCIP into provider-neutral flat records #559 boundary and defer FactBatch and query indexing to tracker(core): land the versioned FactBatchBuffer transport #560 and perf(core): add a graph-free FactQueryIndex for symbol queries #561

Type of Change

• Bug fix
• New feature
• Breaking change
• Documentation update
• Refactoring
• Performance improvement
• Tests

Testing

• Tests pass locally
• Added new tests for the changes
• pre-commit run --files <11 changed files>
• make core-test — 7 passed, 4 skipped
• python -m mkdocs build --strict

Checklist

• My code follows the project style guidelines
• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• My changes generate no new warnings
• Any dependent changes have been merged and published

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f1e2f140fe

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Sort vertices before assigning indexed IDs

When the same index is decoded with different standard-library implementations or builds, iteration over merged_nodes is not stable because it is an unordered_map, so both the returned vertex order and every indexed edge endpoint can change. This contradicts the new decode_records() contract in docs/core_cpp.md that promises deterministic vertex ordering and can make downstream serialized fact batches or cache keys non-reproducible; sort the nodes or preserve deterministic insertion order before constructing vertex_ids.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Replace the hard-coded root node type

The new record path introduces another literal graph node type, so a future vocabulary or schema change can update CodeGraph::add_root_node without updating decoded records, producing different root attributes depending on the entry point. Add the root type to the centralized node-type vocabulary and use a named constant here instead of "root".

AGENTS.md reference: AGENTS.md:L122-L123

Useful? React with 👍 / 👎.