pprobe: Python Hook Tool

Welcome to the pprobe repository, a powerful Python hook tool designed to enhance your development and debugging experience. This tool provides a variety of features that can help you trace function calls, manage device conversions, and much more. Below is an overview of the project, its installation process, usage, and contribution guidelines.

# pprobe: Python Hook Tool

Welcome to the pprobe repository, a powerful Python hook tool designed to enhance your development and debugging experience. This tool provides a variety of features that can help you trace function calls, manage device conversions, and much more. Below is an overview of the project, its installation process, usage, and contribution guidelines.

## Overview

pprobe is a versatile Python package that allows you to:

- Trace and log function calls within your Python applications.
- Detect and log device conversions for tensors in PyTorch.
- Utilize a command-line interface to enable or disable specific options.

The tool is particularly useful for debugging complex Python applications, especially those involving deep learning frameworks like PyTorch.

## Installation

To install pprobe, clone the repository and use pip to install the package and its dependencies:

```sh
git clone https://github.com/your-username/pprobe.git
cd pprobe
pip install -r requirements.txt
python setup.py install

Usage

pprobe can be used in various ways depending on your needs:

Basic Usage

To start using pprobe in your Python script, you can import and initialize it as follows:

from pprobe.bootstrap import TorchFunctionContext

context = TorchFunctionContext()
context.__enter__()

# Your code here

context.__exit__()

Command Line Interface

pprobe comes with a command-line interface that allows you to enable, disable, list, and reset options. Here are some examples:

List current status:

PPROBE --list

Enable specific options:

PPROBE --enable option1,option2

Disable specific options:

PPROBE --disable option1,option2

Reset options to default:

PPROBE --reset

PyTorch Specific Hooks

For PyTorch users, pprobe provides specific hooks to trace tensor operations:

from pprobe.bootstrap import torch_hook_fn

torch_hook_fn('torch', torch)

Contribution

Contributions to pprobe are welcome! To contribute, please follow these steps:

Fork the repository and clone it to your local machine.
Create a new branch for your changes.
Make your changes and ensure that tests pass.
Submit a pull request with a detailed description of your changes.

Testing

To run tests, use tox:

tox

PPROBE CLI

PPROBE --list
=================================================
██████╗ ██████╗ ██████╗  ██████╗ ██████╗ ███████╗
██╔══██╗██╔══██╗██╔══██╗██╔═══██╗██╔══██╗██╔════╝
██████╔╝██████╔╝██████╔╝██║   ██║██████╔╝█████╗  
██╔═══╝ ██╔═══╝ ██╔══██╗██║   ██║██╔══██╗██╔══╝  
██║     ██║     ██║  ██║╚██████╔╝██████╔╝███████╗
╚═╝     ╚═╝     ╚═╝  ╚═╝ ╚═════╝ ╚═════╝ ╚══════╝                                                                                                                                  
=================================================


+------------------------+--------+---------+
| TOGGLE-NAMES           | STATUS | DEFAULT |
+------------------------+--------+---------+
| PPROBE_ENABLE          |  True  |  False  |
| TORCH_REPRODUCE        |  True  |  True   |
| TORCH_CATCH_STEP       | False  |  False  |
| TORCH_CATCH_LOSS       | False  |  False  |
| TORCH_CATCH_LR         | False  |  False  |
| TORCH_DUMP_OP          |  True  |  False  |
| TORCH_DUMP_MODULE      | False  |  False  |
| TORCH_DUMP_DIST        | False  |  False  |
| TORCH_DUMP_MEMORY      | False  |  False  |
| TORCH_TEST_DUMP_OP     | False  |  False  |
| TORCH_TEST_DUMP_MODULE | False  |  False  |
| TORCH_TEST_DUMP_DIST   | False  |  False  |
| TORCH_PERF_ISSUE       | False  |  False  |
| TORCH_TRACE_FILE       | False  |  False  |
+------------------------+--------+---------+

License

pprobe is released under the MIT License.

Contact

For any questions or feedback, please reach out to clemente0620@gmail.com.

Feel free to explore the code and examples provided in this repository. We hope that pprobe will be a valuable addition to your Python development toolkit. Happy coding!

clemente0731/pprobe