{"owner":"feast-dev","repo":"feast","hasSkills":true,"hasMcp":false,"mcpConfig":null,"found":["AGENTS.md"],"skills":{"AGENTS.md":"# Feast - Agent Instructions\n\nThis file provides instructions for AI coding agents (GitHub Copilot, Claude Code, OpenAI Codex, etc.) working in this repository.\n\n## Project Overview\n\nFeast is an open source feature store for machine learning that helps ML platform teams manage features consistently for training and serving.\n\nFeast (Feature Store) is a Python-based project that provides:\n- **Offline Store**: Process historical data for batch scoring or model training\n- **Online Store**: Power real-time predictions with low-latency features\n- **Feature Server**: Serve pre-computed features online\n- **Point-in-time correctness**: Prevent data leakage during model training\n- **Data infrastructure abstraction**: Decouple ML from data infrastructure\n\n## Development Commands\n\n### Setup\n```bash\nmake install-python-dependencies-dev\nmake install-python-dependencies-minimal\n```\n\n### Code Quality\n```bash\n# Format Python code (entire codebase)\nmake format-python\n\n# Lint Python code (entire codebase)\nmake lint-python\n\n# Full type check (entire codebase)\nuv run bash -c \"cd sdk/python && mypy feast\"\n\n# Full type check including tests\nmake mypy-full\n```\n\n#### Single-file lint and type-check\nWhen working on a specific file, run checks scoped to that file to get fast feedback:\n```bash\nuv run ruff check sdk/python/feast/path/to/file.py          # lint\nuv run ruff check --fix sdk/python/feast/path/to/file.py    # lint + auto-fix\nuv run ruff format sdk/python/feast/path/to/file.py         # format\nuv run bash -c \"cd sdk/python && mypy feast/path/to/file.py\" # type-check\n```\n\n### Testing\n```bash\nmake test-python-unit                  # unit tests\nmake test-python-integration-local    # integration tests (local)\nmake test-python-integration          # integration tests (CI)\nmake test-python-universal            # all Python tests\n```\n\n### Protobuf Compilation\n```bash\nmake compile-protos-python   # Python protobufs\nmake protos                  # all protos\n```\n\n### Go Development\n```bash\nmake build-go && make test-go && make format-go && make lint-go\n```\n\n## Key Technologies\n\n- **Languages**: Python (primary), Go\n- **Dependencies**: pandas, pyarrow, SQLAlchemy, FastAPI, protobuf\n- **Data Sources**: BigQuery, Snowflake, Redshift, Parquet, Postgres, Spark\n- **Online Stores**: Redis, DynamoDB, Bigtable, Snowflake, SQLite, Postgres\n- **Offline Stores**: BigQuery, Snowflake, Redshift, Spark, Dask, DuckDB\n- **Cloud Providers**: AWS, GCP, Azure\n\n## Agent Skills\n\nThe `skills/` directory contains tool-agnostic skills (compatible with Claude Code, OpenAI Codex, and other agent tools):\n\n| Skill | Path | Use when… |\n|---|---|---|\n| **feast-user-guide** | `skills/feast-user-guide/SKILL.md` | Working with Feast as a user: defining features, retrieval, CLI, RAG |\n| **feast-dev** | `skills/feast-dev/SKILL.md` | Contributing to Feast: setup, tests, Docker, docs, PR workflow |\n| **feast-architecture** | `skills/feast-architecture/SKILL.md` | Understanding how each component works: registry, materialization, feature server, data flows |\n| **feast-testing** | `skills/feast-testing/SKILL.md` | Writing tests, running targeted tests, debugging registry/online store issues |\n\nReference docs: `skills/references/` — feature definitions, configuration, retrieval & RAG.\n\nArchitecture & design intent: `docs/getting-started/architecture/` (overview, write patterns, RBAC), `docs/getting-started/components/` (per-component pages), `docs/adr/` (design decisions).\n\n## Code Style\n\n- Use type hints on all Python function signatures\n- Follow existing patterns in the module you are modifying\n- PR titles must follow conventional commit conventions with a lowercase type and a capitalized subject after the colon: `feat: Add ...`, `fix: Correct ...`, `ci: Update ...`, `chore: Refresh ...`, `docs: Add ...`\n- Sign off commits with `git commit -s` (DCO requirement)\n- Uses `ruff` for Python linting and formatting; Go uses standard `gofmt`\n- Recompile protos after making changes to `.proto` files (`make protos`)\n- Changes to core functionality may require updates across both Python and Go SDKs\n\n## Documentation and Blog Posts\n\n- **Blog posts must be placed in `/infra/website/docs/blog/`** — do NOT create blog posts under `docs/blog/` or any other location.\n- Blog post files must include YAML frontmatter with `title`, `description`, `date`, and `authors` fields, following the format of existing posts in that directory.\n- All other reference documentation goes under `docs/`.\n\n## Contributing\n\n1. Follow the [contribution guide](docs/project/contributing.md)\n2. Set up your development environment\n3. Run relevant tests before submitting PRs\n4. Ensure code passes linting and type checking\n"},"files":{"AGENTS.md":"# Feast - Agent Instructions\n\nThis file provides instructions for AI coding agents (GitHub Copilot, Claude Code, OpenAI Codex, etc.) working in this repository.\n\n## Project Overview\n\nFeast is an open source feature store for machine learning that helps ML platform teams manage features consistently for training and serving.\n\nFeast (Feature Store) is a Python-based project that provides:\n- **Offline Store**: Process historical data for batch scoring or model training\n- **Online Store**: Power real-time predictions with low-latency features\n- **Feature Server**: Serve pre-computed features online\n- **Point-in-time correctness**: Prevent data leakage during model training\n- **Data infrastructure abstraction**: Decouple ML from data infrastructure\n\n## Development Commands\n\n### Setup\n```bash\nmake install-python-dependencies-dev\nmake install-python-dependencies-minimal\n```\n\n### Code Quality\n```bash\n# Format Python code (entire codebase)\nmake format-python\n\n# Lint Python code (entire codebase)\nmake lint-python\n\n# Full type check (entire codebase)\nuv run bash -c \"cd sdk/python && mypy feast\"\n\n# Full type check including tests\nmake mypy-full\n```\n\n#### Single-file lint and type-check\nWhen working on a specific file, run checks scoped to that file to get fast feedback:\n```bash\nuv run ruff check sdk/python/feast/path/to/file.py          # lint\nuv run ruff check --fix sdk/python/feast/path/to/file.py    # lint + auto-fix\nuv run ruff format sdk/python/feast/path/to/file.py         # format\nuv run bash -c \"cd sdk/python && mypy feast/path/to/file.py\" # type-check\n```\n\n### Testing\n```bash\nmake test-python-unit                  # unit tests\nmake test-python-integration-local    # integration tests (local)\nmake test-python-integration          # integration tests (CI)\nmake test-python-universal            # all Python tests\n```\n\n### Protobuf Compilation\n```bash\nmake compile-protos-python   # Python protobufs\nmake protos                  # all protos\n```\n\n### Go Development\n```bash\nmake build-go && make test-go && make format-go && make lint-go\n```\n\n## Key Technologies\n\n- **Languages**: Python (primary), Go\n- **Dependencies**: pandas, pyarrow, SQLAlchemy, FastAPI, protobuf\n- **Data Sources**: BigQuery, Snowflake, Redshift, Parquet, Postgres, Spark\n- **Online Stores**: Redis, DynamoDB, Bigtable, Snowflake, SQLite, Postgres\n- **Offline Stores**: BigQuery, Snowflake, Redshift, Spark, Dask, DuckDB\n- **Cloud Providers**: AWS, GCP, Azure\n\n## Agent Skills\n\nThe `skills/` directory contains tool-agnostic skills (compatible with Claude Code, OpenAI Codex, and other agent tools):\n\n| Skill | Path | Use when… |\n|---|---|---|\n| **feast-user-guide** | `skills/feast-user-guide/SKILL.md` | Working with Feast as a user: defining features, retrieval, CLI, RAG |\n| **feast-dev** | `skills/feast-dev/SKILL.md` | Contributing to Feast: setup, tests, Docker, docs, PR workflow |\n| **feast-architecture** | `skills/feast-architecture/SKILL.md` | Understanding how each component works: registry, materialization, feature server, data flows |\n| **feast-testing** | `skills/feast-testing/SKILL.md` | Writing tests, running targeted tests, debugging registry/online store issues |\n\nReference docs: `skills/references/` — feature definitions, configuration, retrieval & RAG.\n\nArchitecture & design intent: `docs/getting-started/architecture/` (overview, write patterns, RBAC), `docs/getting-started/components/` (per-component pages), `docs/adr/` (design decisions).\n\n## Code Style\n\n- Use type hints on all Python function signatures\n- Follow existing patterns in the module you are modifying\n- PR titles must follow conventional commit conventions with a lowercase type and a capitalized subject after the colon: `feat: Add ...`, `fix: Correct ...`, `ci: Update ...`, `chore: Refresh ...`, `docs: Add ...`\n- Sign off commits with `git commit -s` (DCO requirement)\n- Uses `ruff` for Python linting and formatting; Go uses standard `gofmt`\n- Recompile protos after making changes to `.proto` files (`make protos`)\n- Changes to core functionality may require updates across both Python and Go SDKs\n\n## Documentation and Blog Posts\n\n- **Blog posts must be placed in `/infra/website/docs/blog/`** — do NOT create blog posts under `docs/blog/` or any other location.\n- Blog post files must include YAML frontmatter with `title`, `description`, `date`, and `authors` fields, following the format of existing posts in that directory.\n- All other reference documentation goes under `docs/`.\n\n## Contributing\n\n1. Follow the [contribution guide](docs/project/contributing.md)\n2. Set up your development environment\n3. Run relevant tests before submitting PRs\n4. Ensure code passes linting and type checking\n"},"items":[{"name":"AGENTS.md","path":"AGENTS.md","title":"AGENTS.md","content":"# Feast - Agent Instructions\n\nThis file provides instructions for AI coding agents (GitHub Copilot, Claude Code, OpenAI Codex, etc.) working in this repository.\n\n## Project Overview\n\nFeast is an open source feature store for machine learning that helps ML platform teams manage features consistently for training and serving.\n\nFeast (Feature Store) is a Python-based project that provides:\n- **Offline Store**: Process historical data for batch scoring or model training\n- **Online Store**: Power real-time predictions with low-latency features\n- **Feature Server**: Serve pre-computed features online\n- **Point-in-time correctness**: Prevent data leakage during model training\n- **Data infrastructure abstraction**: Decouple ML from data infrastructure\n\n## Development Commands\n\n### Setup\n```bash\nmake install-python-dependencies-dev\nmake install-python-dependencies-minimal\n```\n\n### Code Quality\n```bash\n# Format Python code (entire codebase)\nmake format-python\n\n# Lint Python code (entire codebase)\nmake lint-python\n\n# Full type check (entire codebase)\nuv run bash -c \"cd sdk/python && mypy feast\"\n\n# Full type check including tests\nmake mypy-full\n```\n\n#### Single-file lint and type-check\nWhen working on a specific file, run checks scoped to that file to get fast feedback:\n```bash\nuv run ruff check sdk/python/feast/path/to/file.py          # lint\nuv run ruff check --fix sdk/python/feast/path/to/file.py    # lint + auto-fix\nuv run ruff format sdk/python/feast/path/to/file.py         # format\nuv run bash -c \"cd sdk/python && mypy feast/path/to/file.py\" # type-check\n```\n\n### Testing\n```bash\nmake test-python-unit                  # unit tests\nmake test-python-integration-local    # integration tests (local)\nmake test-python-integration          # integration tests (CI)\nmake test-python-universal            # all Python tests\n```\n\n### Protobuf Compilation\n```bash\nmake compile-protos-python   # Python protobufs\nmake protos                  # all protos\n```\n\n### Go Development\n```bash\nmake build-go && make test-go && make format-go && make lint-go\n```\n\n## Key Technologies\n\n- **Languages**: Python (primary), Go\n- **Dependencies**: pandas, pyarrow, SQLAlchemy, FastAPI, protobuf\n- **Data Sources**: BigQuery, Snowflake, Redshift, Parquet, Postgres, Spark\n- **Online Stores**: Redis, DynamoDB, Bigtable, Snowflake, SQLite, Postgres\n- **Offline Stores**: BigQuery, Snowflake, Redshift, Spark, Dask, DuckDB\n- **Cloud Providers**: AWS, GCP, Azure\n\n## Agent Skills\n\nThe `skills/` directory contains tool-agnostic skills (compatible with Claude Code, OpenAI Codex, and other agent tools):\n\n| Skill | Path | Use when… |\n|---|---|---|\n| **feast-user-guide** | `skills/feast-user-guide/SKILL.md` | Working with Feast as a user: defining features, retrieval, CLI, RAG |\n| **feast-dev** | `skills/feast-dev/SKILL.md` | Contributing to Feast: setup, tests, Docker, docs, PR workflow |\n| **feast-architecture** | `skills/feast-architecture/SKILL.md` | Understanding how each component works: registry, materialization, feature server, data flows |\n| **feast-testing** | `skills/feast-testing/SKILL.md` | Writing tests, running targeted tests, debugging registry/online store issues |\n\nReference docs: `skills/references/` — feature definitions, configuration, retrieval & RAG.\n\nArchitecture & design intent: `docs/getting-started/architecture/` (overview, write patterns, RBAC), `docs/getting-started/components/` (per-component pages), `docs/adr/` (design decisions).\n\n## Code Style\n\n- Use type hints on all Python function signatures\n- Follow existing patterns in the module you are modifying\n- PR titles must follow conventional commit conventions with a lowercase type and a capitalized subject after the colon: `feat: Add ...`, `fix: Correct ...`, `ci: Update ...`, `chore: Refresh ...`, `docs: Add ...`\n- Sign off commits with `git commit -s` (DCO requirement)\n- Uses `ruff` for Python linting and formatting; Go uses standard `gofmt`\n- Recompile protos after making changes to `.proto` files (`make protos`)\n- Changes to core functionality may require updates across both Python and Go SDKs\n\n## Documentation and Blog Posts\n\n- **Blog posts must be placed in `/infra/website/docs/blog/`** — do NOT create blog posts under `docs/blog/` or any other location.\n- Blog post files must include YAML frontmatter with `title`, `description`, `date`, and `authors` fields, following the format of existing posts in that directory.\n- All other reference documentation goes under `docs/`.\n\n## Contributing\n\n1. Follow the [contribution guide](docs/project/contributing.md)\n2. Set up your development environment\n3. Run relevant tests before submitting PRs\n4. Ensure code passes linting and type checking\n","category":"root","tokens":1175}]}