PySnooper: Trace Python Functions Without a Full Debugger

Overview

PySnooper is a Python debugging library that records execution details for selected code. Add a decorator to a function or use a context manager around a block to see which lines ran, when they ran, and when local variables changed.

It is useful when print statements are cumbersome and setting up a full debugger is more effort than the investigation warrants. The trace can go to standard error, a file, a stream, or a callable, making it practical in codebases where debugger setup is inconvenient.

Key Features

• Trace a function by adding the @pysnooper.snoop() decorator.
• Trace only a specific block with a with pysnooper.snoop(): context manager.
• Record executed lines, timestamps, and changes to local variables.
• Send trace output to standard error or specify a file, stream, or callable.
• Watch expressions beyond local variables, such as object attributes and indexed values.
• Increase trace depth to include functions called by the traced function.

Use Cases

• A Python developer investigating unexpected function output can quickly inspect execution and local state without placing print statements throughout the code.
• A maintainer working in a large codebase can trace a narrow function or block when configuring a full debugger is impractical.
• A developer debugging code whose standard error is hard to access can direct the trace to a log file or another supported output target.
• A developer checking how nested calls affect a result can increase trace depth to include called functions.

Getting Started

Install with pip:

pip install pysnooper

See the README for usage examples and links to advanced options.

Alternatives

• pudb: PuDB provides an interactive terminal debugger with breakpoints and stepping, while PySnooper logs execution and local-variable changes automatically.

Considerations

PySnooper provides execution traces, not the breakpoints and watch interface of a full-fledged debugger. Tracing can produce detailed logs, so use it on the code you need to inspect and choose an appropriate output destination. The repository is MIT licensed; see the README for installation options and advanced usage.