Abstraction, summarization and code intelligence — built for both humans and LLMs

RepoGPT turns a source-tree into a consultable abstraction layer:
structured, queryable and ready for downstream indexing or RAG pipelines.

\[Collector] → \[Parser] → \[Processor] → \[Publisher]
\|            |              |             |
paths      CodeNode-trees  optional     JSON / NDJSON / stdout

• Languages – Python (.py) & Markdown (.md) out-of-the-box.
  Extendable via plug-in parsers.
• Outputs – hierarchical or flat, single-file JSON or streaming NDJSON.
• Logging – powered by structlog; fully STDOUT-safe.
• Fail-fast – abort immediately on the first parser error if you need strict runs.
• Ignore rules – .repogptignore (git-wildmatch) + sensible defaults (.git, node_modules, …).

git clone https://github.com/MrCabss69/RepoGPT.git
cd RepoGPT
python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"           # installs structlog, pathspec, pytest, ruff…

Este proyecto utiliza pre-commit para asegurar calidad automática:

• Linting y autoformato (black, ruff)
• Chequeo de tipado (mypy)
• Tests unitarios y cobertura ≥80% (pytest --cov)

¿Cómo contribuyo de forma segura?

1. Instala pre-commit (una vez):

   pip install pre-commit
   pre-commit install
   

2. Antes de commitear, ejecuta todos los checks:

   pre-commit run --all-files
   

Si algún check falla, arregla el código antes de push/PR.
El pipeline de CI es igual de estricto.

# analyse a codebase and emit a single JSON file
repogpt path-to-project/ -o report.json

# NDJSON one-line-per-file, streamed to stdout (great for pipes)
repogpt path-to-project/  --flatten node --format ndjson --stdout | jq 'select(.type=="Class")'

Use the same glob syntax as .gitignore to exclude paths or files in addition to built-ins such as .git/, node_modules/, __pycache__/, etc.

# ignore generated docs
docs/build/

# ignore big assets
*.png
*.pdf

[
  {
    "id": "…",
    "type": "Module",
    "name": "utils",
    "path": "src/repogpt/utils/text_processing.py",
    "lang": "py",
    "metrics": { "lines_of_code": 180, "blank_lines": 40 },
    …
  },
  { "id": "…", "type": "Function", "name": "extract_comments", … },
  …
]

{"id":"…","type":"Module", ... ,"path":"README.md","lang":"md"}
{"id":"…","type":"Module", ... ,"path":"src/repogpt/__init__.py","lang":"py"}

RepoGPT never mixes data and logs:

• Data → STDOUT (--stdout) or the output file.
• Logs → STDERR (via structlog).

Examples:

2025-05-17 18:12:07 [info ] starting run          format=ndjson repo=/path/to/repo
2025-05-17 18:12:07 [debug] skip                  path=tests/foo.py reason=ignored
2025-05-17 18:12:08 [error] aborting — fail-fast  first_error="SyntaxError: invalid syntax"

Capture with pytest’s caplog, or redirect STDERR to a file in CI.

ruff check .
mypy src/
pytest -q

src/repogpt/
├── adapters/
│   ├── collector/      # filesystem traversal & ignore logic
│   ├── parser/         # language-specific parsers → CodeNode trees
│   ├── pipeline/       # glue + processors
│   └── publisher/      # JSON/NDJSON writer
├── core/               # service + clean-architecture ports
├── utils/              # file & text helpers
└── app/cli.py          # entry-point

1. Create src/repogpt/adapters/parser/<lang>_parser.py implementing parse() → CodeNode.
2. Register it in adapters/parser/__init__.py.
3. Add extension to docs and tests.

pytest                                 # full unit/integration suite
pytest tests/test_phase3.py -q         # logging & fail-fast happy-path

The suite exercises:

• Collect / ignore rules
• Markdown & Python parsers (fixtures under tests/data/)
• NDJSON vs JSON writer
• Fail-fast & debug logging

• Phase 4 – caching by file-hash + parallel workers
• Phase 5 – CI (ruff + mypy + pytest), release to PyPI
• Phase 6 – plug-in entry-points for custom processors & new languages
• Phase 7 – optional HTML / graph visualizer

MIT

Happy hacking 💻