# tree-sitter-stack-graphs definition for Python This project defines tree-sitter-stack-graphs rules for Python using the [tree-sitter-python][] grammar. [tree-sitter-python]: https://crates.io/crates/tree-sitter-python - [API documentation](https://docs.rs/tree-sitter-stack-graphs-python/) - [Release notes](https://github.com/github/stack-graphs/blob/main/languages/tree-sitter-stack-graphs-python/CHANGELOG.md) ## Using the API To use this library, add the following to your `Cargo.toml`: ```toml [dependencies] tree-sitter-stack-graphs-python = "0.2" ``` Check out our [documentation](https://docs.rs/tree-sitter-stack-graphs-python/*/) for more details on how to use this library. ## Using the Command-line Program The command-line program for `tree-sitter-stack-graphs-python` lets you do stack graph based analysis and lookup from the command line. The CLI can be run as follows: 1. _(Installed)_ Install the CLI using Cargo as follows: ```sh cargo install --features cli tree-sitter-stack-graphs-python ``` After this, the CLI should be available as `tree-sitter-stack-graphs-python`. 2. _(From source)_ Instead of installing the CLI, it can also be run directly from the crate directory, as a replacement for a `tree-sitter-stack-graphs-python` invocation, as follows: ```sh cargo run --features cli -- ``` The basic CLI workflow for the command-line program is to index source code and issue queries against the resulting database: 1. Index a source folder as follows: ```sh tree-sitter-stack-graphs-python index SOURCE_DIR ``` _Indexing will skip any files that have already be indexed. To force a re-index, add the `-f` flag._ To check the status if a source folder, run: ```sh tree-sitter-stack-graphs-python status SOURCE_DIR ``` To clean the database and start with a clean slate, run: ```sh tree-sitter-stack-graphs-python clean ``` _Pass the `--delete` flag to not just empty the database, but also delete it. This is useful to resolve `unsupported database version` errors that may occur after a version update._ 2. Run a query to find the definition(s) for a reference on a given line and column, run: ```sh tree-sitter-stack-graphs-python query definition SOURCE_PATH:LINE:COLUMN ``` Resulting definitions are printed, including a source line if the source file is available. Discover all available commands and flags by passing the `-h` flag to the CLI directly, or to any of the subcommands. ## Development The project is written in Rust, and requires a recent version installed. Rust can be installed and updated using [rustup][]. [rustup]: https://rustup.rs/ The project is organized as follows: - The stack graph rules are defined in `src/stack-graphs.tsg`. - Builtins sources and configuration are defined in `src/builtins.it` and `builtins.cfg` respectively. - Tests are put into the `test` directory. ### Running Tests Run the tests as follows: ```sh cargo test ``` The project consists of a library and a CLI. By default, running `cargo` only applies to the library. To run `cargo` commands on the CLI as well, add `--features cli` or `--all-features`. Run the CLI from source as follows: ```sh cargo run --features cli -- ARGS ``` Sources are formatted using the standard Rust formatted, which is applied by running: ```sh cargo fmt ``` ### Writing TSG The stack graph rules are written in [tree-sitter-graph][]. Checkout the [examples][], which contain self-contained TSG rules for specific language features. A VSCode [extension][] is available that provides syntax highlighting for TSG files. [tree-sitter-graph]: https://github.com/tree-sitter/tree-sitter-graph [examples]: https://github.com/github/stack-graphs/blob/main/tree-sitter-stack-graphs/examples/ [extension]: https://marketplace.visualstudio.com/items?itemName=tree-sitter.tree-sitter-graph Parse and test a single file by executing the following commands: ```sh cargo run --features cli -- parse FILES... cargo run --features cli -- test TESTFILES... ``` Generate a visualization to debug failing tests by passing the `-V` flag: ```sh cargo run --features cli -- test -V TESTFILES... ``` To generate the visualization regardless of test outcome, execute: ```sh cargo run --features cli -- test -V --output-mode=always TESTFILES... ``` Go to for links to examples and documentation.