GitScope is a local-first, offline Git repository analytics platform for developers and engineering teams. It analyzes a local Git repository with the Git CLI, normalizes data into SQLite, and generates a static HTML + JSON report.
The current implementation covers the six documented MVP analysis domains:
- Overview
- Commits
- Contributors
- Branches
- Files
- Timeline
The report output contract is:
report/
├── index.html
├── report.json
└── assets/
Install the package in editable mode:
python3 -m pip install -e .Generate a report for the current repository:
gitscope analyze .
gitscope analyze . --output report
gitscope analyze . --branch main --since 2025-01-01 --until 2025-12-31Run the test suite:
python3 -m pytest| Topic | Current implementation |
|---|---|
| Language | Python 3.11+ |
| CLI | Typer |
| Git data source | Local Git CLI |
| Modeling | Pydantic |
| Storage | In-memory SQLite via SQLAlchemy during each analysis run |
| HTML report | Jinja2 shell + local JavaScript/SVG renderer |
| Testing | pytest |
The HTML report is driven by the same unified data model written to report.json. For local file:// viewing, index.html embeds the same payload so the static report works without a web server.
- Local First: run locally by default
- Offline First: analysis does not depend on external platforms
- Source Safe: no source upload and no hosting-platform authorization
- CLI First: analysis starts from the command line
- Report Driven: HTML and JSON are the primary outputs
- Extensible: the module layout is ready for future analyzers
| Document | Purpose |
|---|---|
doc/requirement.md |
MVP goals, scope, requirements, non-functional constraints, and acceptance criteria |
doc/technology_design.md |
Stack choices, architecture, data model, execution flow, and extension strategy |
doc/ui_design.md |
Report information architecture, page layout, and UI behavior |
When implementation changes affect documented behavior, update the related docs in the same task so code and documentation stay aligned.
