Pytype

Warnings

• breaking Pytype is in maintenance mode; Google has discontinued active development, and Python 3.12 is the last supported version. No new features will be added, only bugfixes. Users are encouraged to explore alternative type checkers like Mypy, Pyright, ty, or Pyrefly.
  Fix: Consider migrating to actively developed type checkers if you need support for newer Python versions or actively evolving typing features.
• gotcha Pytype's default behavior is lenient compared to other type checkers like Mypy. It allows operations that succeed at runtime and don't contradict annotations, which might lead to different results, particularly with container types or `Optional` handling.
  Fix: Be aware of Pytype's leniency. If stricter checking is desired, review Pytype's error classes and configuration options, or consider a different type checker.
• gotcha For silencing specific errors, use `pytype: disable=error-class` on the line or block, or `type: ignore`. Disabling errors for an entire file can be done by placing a disable directive on its own line at the start of the file.
  Fix: Use inline directives like `# pytype: disable=attribute-error` or `# type: ignore` for specific suppression. Consult the error classes documentation for precise error names.
• gotcha Performance can degrade significantly on very large Python files (anecdotally, over ~1500 lines). Extensive inference on complex, unannotated code can be slow.
  Fix: Split large files into smaller modules. Add explicit type annotations to function parameters and return types. Simplify complex variable initializations or union types.
• gotcha Some experimental features, such as `...` as a top-level annotation for 'inferred type', are unique to Pytype and not supported by other type checkers (e.g., Mypy, Pyright). Using these can lead to errors when analyzing code with other tools.
  Fix: Avoid experimental Pytype-specific annotations in code intended for wider compatibility or analysis by other type checkers.
• gotcha Pytype is primarily developed and tested on Linux. While it supports macOS (10.7+ and Xcode 8+), Windows support typically requires the Windows Subsystem for Linux (WSL).
  Fix: For Windows users, it is recommended to run Pytype within a WSL environment.

Install

• pip install pytype Install Pytype

Quickstart

Pytype is primarily a command-line tool. To quickly get started, run `pytype` followed by the file or directory you wish to analyze. For project-level configuration, use a `pyproject.toml` file to specify inputs and settings.

# To check a single file or directory:
# pytype my_module.py
# pytype my_package/

# For package-wide configuration using pyproject.toml:
# 1. Create a pyproject.toml in your project root with the following:
# [tool.pytype]
# inputs = ['your_package_name']

# 2. Then run pytype without arguments:
# pytype

# Example Python file (example.py):
def greet(name):
    return "Hello, " + name

def add(a, b):
    return a + b

# Run pytype on this file from your terminal:
# pytype example.py
# Expected output for add function (e.g., if you call add(1, '2')):
# File "example.py", line 5, in add: unsupported operand type(s) for +: 'int' and 'str' [unsupported-operands]