test(benchmarks): add RFC-0001 LLM surface corpus

[tingfeng347]

Summary

• add a provider-neutral bilingual natural-language to canonical OneRoll v2 surface corpus
• validate the corpus with a Draft 2020-12 JSON Schema and repository contract tests
• verify every canonical oracle against the accepted RFC-0001 target Pest grammar
• document the boundary between exact canonical match, grammar validity, and future semantic evaluation

Scope

This PR does not change RFC-0001, the target grammar, parser behavior, runtime semantics, package dependencies, or CI workflows. It establishes deterministic evidence infrastructure for the usability and LLM tool-call concern discussed in #51.

Validation

• uv sync --all-groups --frozen
• uv run --frozen python -m unittest tests.test_llm_surface_contract -v
• cargo test v2_target_grammar_parses_llm_surface_corpus -- --nocapture
• cargo fmt --all -- --check
• cargo test --all-targets --all-features
• cargo test --release checked_i64_
• cargo test --release rfc_0002_
• cargo clippy --all-targets --all-features -- -D warnings -A non-local-definitions
• uv run --frozen ruff check .
• uv run --frozen ruff format --check .
• uv run --frozen mypy --strict src/oneroll docs/rfcs/0004-target-api.pyi
• uv run --frozen maturin develop
• uv run --no-sync python -m unittest discover -s tests -v
• LC_ALL=C LANG=C uv run --no-sync sphinx-build -W --keep-going -b html docs/source docs/_build/html
• git diff --check

Related to #51.

Summary by Sourcery

Establish deterministic bilingual LLM surface evaluation evidence for RFC-0001 without changing language, parser, runtime, dependencies, or CI behavior.

New Features:

• Add a provider-neutral bilingual corpus mapping natural-language intents to canonical OneRoll v2 expressions.

Enhancements:

• Define clear boundaries between exact canonical matching, target-grammar validity, and future semantic evaluation.

Documentation:

• Document the corpus purpose, scope, evaluation metrics, and provider-runner contract.

Tests:

• Validate corpus structure, bilingual pairing, canonical formatting, RFC references, and feature coverage with Draft 2020-12 schema and repository contract tests.
• Verify every canonical corpus expression parses with the accepted RFC-0001 target grammar.

[sourcery-ai]

Reviewer's Guide

Introduces deterministic evidence infrastructure for RFC-0001 LLM-surface usability evaluation: a bilingual canonical corpus, schema and repository contract validation, and Rust target-grammar coverage, while explicitly avoiding changes to the RFC, runtime semantics, dependencies, or CI.

Flow diagram for RFC-0001 LLM surface corpus validation

flowchart LR
    Cases[cases.json] --> Schema[cases.schema.json]
    Cases --> PyContract[Python contract tests]
    Cases --> RustContract[Rust V2TargetParser test]
    Schema --> PyContract
    PyContract --> Structure[Structure and bilingual pairing valid]
    RustContract --> Grammar[Target grammar valid]
    Structure --> Evidence[Deterministic evaluation evidence]
    Grammar --> Evidence

Loading

File-Level Changes

Change	Details	Files
Add a provider-neutral bilingual corpus of natural-language intents paired with canonical RFC-0001 OneRoll v2 expressions.	Add 32+ English/Chinese paired cases covering multiple surface feature families. Document corpus purpose, provider-runner format, validation boundaries, and the distinction between exact matching, grammar validity, and semantic equivalence. Define the corpus structure and constraints with a Draft 2020-12 JSON Schema.	benchmarks/llm_surface/README.md benchmarks/llm_surface/cases.json benchmarks/llm_surface/cases.schema.json
Add repository-level contract checks to enforce corpus quality and metadata consistency.	Validate the schema and corpus contents. Check unique IDs, bilingual pairing, shared expected values and metadata, canonical formatting, RFC references, and feature-family coverage. Ignore strings and top-level comments when checking canonical spacing.	tests/test_llm_surface_contract.py
Verify every corpus oracle against the accepted RFC-0001 target grammar without changing parser or runtime behavior.	Deserialize the corpus in Rust and parse each canonical expression with the embedded V2TargetParser. Report the case ID and canonical expression when grammar validation fails.	src/conformance_tests.rs

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[sourcery-ai]

Hey - I've reviewed your changes and they look great!

Sourcery assessment

Approved.

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[fu050409]

cc @HsiangNianian plz take a look