Skip to content

Latest commit

 

History

History
299 lines (230 loc) · 9.33 KB

README.md

File metadata and controls

299 lines (230 loc) · 9.33 KB

push-test pypi-publish

한국어

pyftrace

Introduction

pyftrace is a lightweight Python function tracing tool designed to monitor and report on function calls within Python scripts. It leverages sys.setprofile(Python 3.8 ~ 3.11), sys.monitoring(Python 3.12 ~) to monitor Python events and trace functions based on the results. With pyftrace, you can trace function calls across multiple modules, visualize call hierarchies, and generate execution time reports.

pyftrace-demo

Key features of pyftrace include:

  • Function Call/Return Tracing: Monitor calls/returns to functions in your Python script and imported modules.
  • Built-in Function Tracing: Optionally trace built-in functions like print using the --verbose flag.
  • Multiple Module Support: Trace functions across multiple files within your project.
  • Execution Reports: Generate reports detailing function execution times and call counts with the --report flag.
  • Path Tracing: Trace the path of traced Python file using the --path flag.
  • TUI Mode: Run pyftrace in a Text User Interface (TUI) mode using the tui command.
$ pyftrace --help
usage: pyftrace [options] [tui] script [script_args ...]

pyftrace: Python function tracing tool.

positional arguments:
  script         Path to the script to run and trace. Specify 'tui' before the script path to run in TUI mode.

options:
  -h, --help     show this help message and exit
  -V, --version  Show the version of pyftrace and exit
  -v, --verbose  Enable built-in and third-party function tracing
  -p, --path     Show file paths in tracing output
  -r, --report   Generate a report of function execution times

Usage

Requirements

  • Python Version: pyftrace requires Python 3.8+.
$ pyftrace [options] /path/to/python/script.py

Installation

$ git clone https://github.com/kangtegong/pyftrace.git
$ cd pyftrace
$ pip install -e .

or

$ pip install pyftrace

note: on Windows, windows-curses is required.

Command-Line Options

  • --report or -r: Generate a report of function execution times and call counts at the end of the script's execution.
  • --verbose or -v: Enable tracing of built-in functions (e.g., print, len). Without this flag, pyftrace only traces user-defined functions.
  • --path or -p: Include file paths in the tracing output.
  • --help or -h: Display help information about pyftrace and its options.
  • --version or -V: Display help information about pyftrace and its options.

TUI

To run pyftrace in TUI(Text User Interface) mode, use the tui command before the script path.

pyftrace [options] tui path/to/your_script.py

Key Bindings

  • or : Scroll up or down one line at a time.
  • PgUp or PgDn: Scroll up or down one page at a time.
  • Home or End: Jump to the beginning or end of the trace.
  • or : Scroll left or right horizontally.
  • q: Quit the TUI mode.

tui-demo

Examples

The examples/ directory contains a variety of Python files that can be traced using pyftrace.

In this example, main_script.py imports and calls functions from module_a.py and module_b.py. We'll use pyftrace to trace these function calls and understand the flow of execution.

# module_a.py
  1 def function_a():
  2     print("Function A is called.")
  3     return "ret_a"
# module_b.py
  1 def function_b():
  2     print("Function B is called.")
  3     return "ret_b"
# main_script.py
  1 from module_a import function_a
  2 from module_b import function_b
  3
  4 def main():
  5     result_a = function_a()
  6     result_b = function_b()
  7     print(f"Results: {result_a}, {result_b}")
  8
  9 if __name__ == "__main__":
 10     main()

Basic Tracing

To trace function calls in main_script.py without including built-in functions or file paths:

$ pyftrace examples/module_trace/main_script.py

output:

Running script: examples/module_trace/main_script.py
Called main from line 10
    Called function_a from line 5
Function A is called.
    Returning function_a-> ret_a
    Called function_b from line 6
Function B is called.
    Returning function_b-> ret_b
Results: ret_a, ret_b
Returning main-> None
Returning <module>-> None

Trace Built-in Functions with --verbose

$ pyftrace --verbose examples/module_trace/main_script.py

output:

Running script: examples/module_trace/main_script.py
Called main from line 10
    Called function_a from line 5
        Called print from line 2
Function A is called.
        Returning print
    Returning function_a-> ret_a
    Called function_b from line 6
        Called print from line 2
Function B is called.
        Returning print
    Returning function_b-> ret_b
    Called print from line 7
Results: ret_a, ret_b
    Returning print
Returning main-> None
Returning <module>-> None

Showing File Paths with --path

$ pyftrace --path examples/module_trace/main_script.py

In this case, When a function is called, pyftrace displays it in the following format:

Called {function} @ {defined file path}:{defined line} from {called file path}:{called line}
  • {function}: name of the function being called
  • {defined file path}: file path where the function is defined (enabled with --path option)
  • {defined line}" line number in the defined file
  • {called line} line number in the calling file
  • {called file path} path to the file that contains the calling function (enabled with --path option)

output:

Running script: examples/module_trace/main_script.py
Called main@examples/module_trace/main_script.py:4 from examples/module_trace/main_script.py:10
    Called function_a@examples/module_trace/module_a.py:1 from examples/module_trace/main_script.py:5
Function A is called.
    Returning function_a-> ret_a @ examples/module_trace/module_a.py
    Called function_b@examples/module_trace/module_b.py:1 from examples/module_trace/main_script.py:6
Function B is called.
    Returning function_b-> ret_b @ examples/module_trace/module_b.py
Results: ret_a, ret_b
Returning main-> None @ examples/module_trace/main_script.py
Returning <module>-> None @ examples/module_trace/main_script.py

Generating an Execution Report

To generate a summary report of function execution times and call counts:

$ pyftrace --report examples/module_trace/main_script.py

output:

Running script: examples/module_trace/main_script.py
Function A is called.
Function B is called.
Results: ret_a, ret_b

Function Name	| Total Execution Time	| Call Count
---------------------------------------------------------
main           	| 0.000082 seconds	| 1
function_a     	| 0.000022 seconds	| 1
function_b     	| 0.000008 seconds	| 1

Combining --verbose and --path

To trace built-in functions and include file paths:

$ pyftrace --verbose --path examples/module_trace/main_script.py
$ pyftrace -vp examples/module_trace/main_script.py

output:

Running script: examples/module_trace/main_script.py
Called main@examples/module_trace/main_script.py:4 from examples/module_trace/main_script.py:10
    Called function_a@examples/module_trace/module_a.py:1 from examples/module_trace/main_script.py:5
        Called print@builtins from examples/module_trace/module_a.py:2
Function A is called.
        Returning print @ examples/module_trace/module_a.py
    Returning function_a-> ret_a @ examples/module_trace/module_a.py
    Called function_b@examples/module_trace/module_b.py:1 from examples/module_trace/main_script.py:6
        Called print@builtins from examples/module_trace/module_b.py:2
Function B is called.
        Returning print @ examples/module_trace/module_b.py
    Returning function_b-> ret_b @ examples/module_trace/module_b.py
    Called print@builtins from examples/module_trace/main_script.py:7
Results: ret_a, ret_b
    Returning print @ examples/module_trace/main_script.py
Returning main-> None @ examples/module_trace/main_script.py
Returning <module>-> None @ examples/module_trace/main_script.py

Combining --verbose and --report

$ pyftrace --verbose --report examples/module_trace/main_script.py
$ pyftrace -vr examples/module_trace/main_script.py

output:

Running script: examples/module_trace/main_script.py
Function A is called.
Function B is called.
Results: ret_a, ret_b

Function Name	| Total Execution Time	| Call Count
---------------------------------------------------------
main           	| 0.000127 seconds	| 1
print          	| 0.000041 seconds	| 3
function_a     	| 0.000021 seconds	| 1
function_b     	| 0.000016 seconds	| 1

Notes

  • simple-pyftrace.py is a simplified pyftrace script for the Pycon Korea 2024 presentation. It is about 100 lines of code, but has limited functionality.

LICENESE

MIT

See LICENSE for more infomation

See Also

pyftrace is heavily inspired by:

  • ftrace: Ftrace is an internal tracer for linux kernel.
  • uftrace: uftrace is a function call graph tracer for C, C++, Rust and Python programs.