sidetrack 2.0.1

Simple debug tracing package for Python

Sidetrack

Sidetrack provides a simple interface for writing log messages in Python programs. Calls to the log functions can be left in your code to let users produce debug logs in the field; if performance matters, using certain coding idioms and turning on Python optimization will cause log statements to be compiled out.

Table of contents

• Introduction
• Installation
• Usage
• Getting help
• Contributing
• License
• Authors and history
• Acknowledgments

Introduction

IDEs are great for debugging and tracing execution of your code, but they can't be used in all situations. For example, if your code is executing on multiple remote computers, or you have released a program to general users and you would like them to send you a debug log/trace of execution, using an IDE at run time may be impractical or impossible. Logging packages such as logging are made for these situations; you can insert logging statements in your code and use the output to understand what is happening as well as for software telemetry and other purposes. However, setting up Python logging or most similar packages is (IMHO) complicated and verbose if you don't need all its features.

Sidetrack (Simple debug tracing package) offers a simple API that lets you turn on logging, set the output destination (which can be stdout), and sprinkle log(f'my message and my {variable} value') throughout your code. Moreover, it is carefully written so that you can cause the log calls to be optimized out completely if your run Python with the -O option and you prefix your log calls with if __debug__. This leads to the following style of using Sidetrack:

...
for item in item_list:
    if __debug__: log(f'getting data for {item}')
    ...

When running with -O, the log statement in the loop will not simply be a no-op function call: Python will completely discard the conditional block, as if the code did not exist. This is as optimal as possible, and means that you do not have to worry about the performance costs of using log or evaluating its arguments.

Installation

The instructions below assume you have a Python interpreter installed on your computer; if that's not the case, please first install Python version 3 and familiarize yourself with running Python programs on your system.

On Linux, macOS, and Windows operating systems, you should be able to install sidetrack with pip. To install sidetrack from the Python package repository (PyPI), run the following command:

python3 -m pip install sidetrack

As an alternative to getting it from PyPI, you can use pip to install sidetrack directly from GitHub, like this:

python3 -m pip install git+https://github.com/caltechlibrary/sidetrack.git

Usage

There are just four functions in the sidetrack package:

• set_debug: turn logging on/off, set the output destination, and configure options
• log: logs a message as-is.
• loglist: takes a list of messages and logs them individually, as if applying log to each one in turn.
• logf: logs a message with optional arguments; the message string can contain embedded format directives.

How to import Sidetrack

To take advantage of Python's optimization behavior, make sure to conditionalize all references to Sidetrack functions on the Python built-in symbol __debug__. This includes the import statement for Sidetrack:

if __debug__:
    from sidetrack import set_debug, log, logf, loglist

The fragment above illustrates another tip: to make calls to the log functions as short as possible in your code, import set_debug, log, logf, and loglist directly using the from sidetrack ... approach instead of doing a plain import sidetrack, so that you can write log(...) instead of sidetrack.log(...). Believe me, your fingers and eyes will thank you!

How to turn on debug logging

To turn on logging, call set_debug(...) at least once in your code. Often, this will most convenient if combined with a command-line argument to your program, so that debug tracing can be enabled or disabled at run-time. The following example shows the basic usage.

if __debug__:
    if ...some condition of your choosing...:
        set_debug(True)
else:
    print('Python -O is in effect, so debug logging is not available.')

The above will turn on debug logging and send output to the default destination, which is the standard error stream (sys.stderr). To send the output to a different destination, use the optional second argument debug_output, which can be either a file name, a stream, or the dash symbol (-); the latter indicates the destination should be the default (i.e., sys.stderr). Here is an example:

if __debug__:
    set_debug(True, dest = '/tmp/debug.txt')

The function set_debug also accepts another optional argument, show_package, that causes each log, loglist, and logf message to be prefixed with the name of the Python package containing the source file where the call to the log function is used. This is very helpful when Sidetrack is used in multiple packages.

if __debug__:
    set_debug(True, show_package = True)

Finally, the function set_debug(...) also accepts one more optional argument, extra, that lets you prefix every output line with extra text of your choosing. The extra text string can contain Python logging system % formatting strings. For example, the process ID can be inserted by passing '%(process)d' as in the following example:

if __debug__:
    set_debug(True, debug_output, extra = '%(process)d')

If your program uses threads, you may find the use of extra = "%(threadName)s" helpful.

How to call log, loglist, and logf

The function log is the most basic function in Sidetrack. Call it with a single argument, the message to be printed:

if __debug__: log("I'm right here!")

If you want to print multiple strings, it's certainly possible to call log multiple times consecutively, but in some situations, it may be more convenient to call loglist – especially if the strings are generated dynamically as in the following example:

if __debug__: loglist(f'{var} = {value}' for var, value in settings())

Finally, there are situations where f-strings cannot be used due to how they are evaluated at run time or due to certain inherent limitations. For those situations, Sidetrack provides the logf function. It accepts one argument, a string, and any number of optional arguments. Here's an example:

if __debug__: logf('exception (failure #{}): {}', failures, str(ex))

Internally, logf applies format to the string and passes any remaining arguments as the arguments to format. In other words, it is essentially the following pseudocode:

def logf(msg, *other_args):
    final_text = msg.format(*other_args)
    write_log(final_text)

When using logf, beware of including references to variables that might expand at run time to contain characters that have special meaning to Python's format command, such as the { character.

Understanding the output

In all cases, each line of the output has the following form:

<package> extra  filename:lineno  function() -- message

where package and extra are optional and controlled by the arguments show_package and extra, respectively, to set_debug(...), and the remaining values are always printed: the file name, line number and function where the call to the log, loglist, or logf was made, and the message. Examples are shown in the next section.

Tips for using Sidetrack

Throughout the rest of your code, in places where it's useful, add calls to the log functions. Here's a simple contrived example, taken from the demonstration program supplied with Sidetrack:

if __debug__: log('=== demo program starting ===')

print('Looping my loopy loop:')
for i in range(0, 3):
    if __debug__: log(f'loop value {i}')
    print('Another go-around the loop')
print('Done looping.')

if __debug__: log('=== demo program stopping ===')

With the code above, if debugging is not turned on, or the program is running with Python optimization turned on, the output will be:

Looping my loopy loop:
Another go-around the loop
Another go-around the loop
Another go-around the loop
Done looping.

With debugging turned on and the destination set to -, the output becomes:

demo_debug.py:32 main() -- === demo program starting ===
Looping my loopy loop:
demo_debug.py:36 main() -- loop value 0
Another go-around the loop
demo_debug.py:36 main() -- loop value 1
Another go-around the loop
demo_debug.py:36 main() -- loop value 2
Another go-around the loop
Done looping.
demo_debug.py:40 main() -- === demo program stopping ===

Being able to send the debug output to a file becomes useful when dealing with longer and more complicated programs – it makes it possible to store a detailed trace without cluttering the output as it is in the sample above. File output can also be useful for deployed code: you can leave the debug functionality in your code and instruct your users to turn on debugging with output directed to a file, then send you the file so you can debug problems more easily.

How to run the demo program

In the tests subdirectory, there is a simple demonstration program illustrating the use of Sidetrack. To run it, on Linux and macOS systems, you can start a terminal shell and run the following commands:

python3 tests/demo_debug.py -h

To run it with debug logging enabled, use the -d command-line option (where the output in this example is given as -, which means to send the output to the terminal):

python3 tests/demo_debug.py -d -

To see the difference when Python optimization is active, add the -O option to the Python interpreter:

python3 -O tests/demo_debug.py -d -

Getting help

If you find an issue, please submit it in the GitHub issue tracker for this repository.

Contributing

We would be happy to receive your help and participation with enhancing sidetrack! Please visit the guidelines for contributing for some tips on getting started.

License

Software produced by the Caltech Library is Copyright (C) 2020, Caltech. This software is freely distributed under a BSD/MIT type license. Please see the LICENSE file for more information.

Authors and history

I developed the first version of this code while implementing Spiral. I started using the code in essentially every Python software package I have written since then, first by copy-pasting the code (which was initially very short) and eventually creating a single-file module (named debug.py). This was obviously a suboptimal approach. Finally, in 2020, I decided it was time to break it out into a proper self-contained Python package.

Acknowledgments

This work was funded by the California Institute of Technology Library.

The vector artwork of a document with a line break, used as the icon for this repository, was created by iconixar from the Noun Project. It is licensed under the Creative Commons CC-BY 3.0 license.