repo-k-graph (rkg)

repo-k-graph (rkg) is a Rust project for deterministic repository knowledge graphs and context infrastructure for AI-assisted software engineering.

Current release: 1.0.12.

The current repository state is complete through Phase 17, Swift UI and macro depth, evaluation/benchmarking, FTS search, packaging/distribution support, and Kotlin Android integration:

Phase 0-2: Repository foundation, local SQLite knowledge store, and git-ignore-aware file ingestion.
Phase 3-4: AST-based Python symbol extraction and full relationship graphs.
Phase 5-6: Python test discovery, coverage integration, and documentation intelligence.
Phase 7-8: Graph traversal, impact analysis, and token-budgeted context packing.
Phase 9-10: MCP stdio server support and Git intelligence.
Phase 11-17: Python, Rust, F#, Mojo, Kotlin, Swift, Android, and benchmark intelligence.
Distribution: GitHub Release workflow, shell completions, generated command reference, release checklist, schema reference, and language adapter guide.

Project Documents

Workspace Layout

crates/
  rkg-cli          CLI binary entry point and command executors
  rkg-core         shared domain model types
  rkg-db           SQLite persistence layer
  rkg-indexer      repository ingestion and file discovery
  rkg-query        unified query logic layer
  rkg-lang-*       language adapters
  rkg-mcp          Model Context Protocol server
fixtures/
  sample-repos/    small repositories for indexing integration tests

Verification

Run these checks before finalizing code changes:

cargo fmt --all -- --check
cargo clippy --workspace --all-targets -- -D warnings
cargo test --workspace

For release packaging validation:

cargo package --workspace --allow-dirty --list

Versioning Policy

rkg uses a project-specific language-support release policy. Completed Python support is the 1.0.0 baseline. Each completed additional language adapter increments the patch version by 0.0.1: Rust (1.0.1), F# (1.0.2), Mojo (1.0.3), Kotlin (1.0.4), and Swift (1.0.5). Subsequent releases introduce feature enhancements, search backends, test coverage integration, and deeper language-specific modeling.

Development Principles