pydebug

Decorators for debugging Python

40
3
Python

PyDebug

This is a simple set of utilities which makes it easier to debug Python objects.
Pydebug respects Django’s config. It checks if the decorator is running within a Django
project and DEBUG is set to True. If it’s set to False, it simply returns the function.

Installing

You can install by cloning this repository and running pip install . within this directory, or
you can install through Pypi, by running pip install python-pydebug.

Requirements

Although Pydebug uses Python’s standard library, it tried to import Django to make sure you are not
within a Django project.

Library Defaults
Cython -
IPDB PDB
Line-profiler -

Line-profiler requires Cython to work, if IPDB is not found, it fallsback to Python’s standard PBD library.

NOTE: At the moment, it is not possible to add rkern/line_profiler as a requirements due to their install process.
Check their README for installation instructions.

Usage

PDBDebugger

from pydebug import PDBDebugger

@PDBDebugger()
def hello_world(a, b):
    x = a + b
    return (a, b)

You may tell PDBdebugger to drop a IPDB shell only when an error occurs

from pydebug import PDBDebugger

@PDBDebugger(on_error=True)
def hello_world(a, b):
    x = a + b
    return (a, b)

DisassembleDebug

You may disassemble a Python function by using the Disassembledebug decorator.

from pydebug import Disassembledebug

@DisassembleDebug()
def hello_world(a, b):
    x = a + b
    return (a, b)

Profilerdebug

Profilerdebug uses Python’s default cProfile library.

from pydebug import ProfilerDebug

@ProfilerDebug()
def hello_world(a, b):
    x = a + b
    return (a, b)

LineProfilerDebug

LineProfilerDebug requires line-profile library to work, which also requires Cython.

from pydebug import LineProfilerDebug

@LineProfilerDebug()
def hello_world(a, b):
    x = a + b
    return (a, b)

ObjectInfoDebug

ObjectInfoDebug prints the arguments which were passed to the function, the performace counter (in minutes), the
process time (in minutes) and the returned objects by the function.

from pydebug import ObjectInfoDebug

@ObjectInfoDebug()
def hello_world(a, b):
    x = a + b
    return (a, b)

TODOs

  1. Allow developer to use multiple decorators instead of one decorating the other
  2. Show better profile metrics
  3. Plot metrics to an image