2017 - Albert Sweigart - Logging and Testing

Description
Logging messages, unit tests, and using the debugger are best practices, but often left out of introductory programming tutorials. As a result, beginners come away with the idea that these time-saving features are somehow "advanced" or things to be learned later on. In this presentation, Al Sweigart dispels these notions by providing quickstart guides to Python's logging, doctest, and pdb modules.

Abstract
Logging messages, unit tests, and using the debugger are best practices, but often left out of introductory programming tutorials. As a result, beginners come away with the idea that these time-saving features are somehow "advanced" or things to be learned later on. In this presentation, Al Sweigart dispels these notions by providing quickstart guides to Python's logging, doctest, and pdb modules, including:

Why beginners (incorrectly) avoid these programming best practices.
The "if you're doing this" signs that you aren't using these best practices.
The 4 lines of code you need from the logging module.
The 4 lines of code you need from the doctest module.
The 2 lines of code you need from the pdf module.
How much time you'll save, not spend, by using these not-so-advanced modules.

Bio
Al Sweigart is the author of several programming books, including Automate the Boring Stuff with Python. He releases all of his books online for free under a Creative Commons license. He lives in San Francisco, and his cat weighs 12 pounds.

Video: https://youtu.be/TDx_Mv3MUw0

Transcript

1. LOGGING AND TESTING AND DEBUGGING, OH MY! @AlSweigart al@inventwithpython.com (last name rhymes with “why dirt”) bit.ly/ohmypy
2. Why?
3. LOGGING
4. LIVE DEMO!
5. This is terrible.
6. TERRIBLE Print Debugging
7. Why do logging instead of print? • Debug output vs. normal output • Easier to remove debug messages. • Easier to put debug messages back. • Fine grain control with logging levels.
8. Logging import logging logging.basicConfig( level=logging.DEBUG) logging.debug('The log message.') logging.disable(logging.CRITICAL) Back to the live demo… This is looks complicated.
10. Logging Levels • DEBUG • INFO • WARN • ERROR • CRITICAL • logging.basicConfig( level=logging.DEBUG) • logging.disable(logging.CRITICAL)
11. DEBUG • Low-level debugging information. • Variables values, etc. • You only look at when something has gone wrong. • Probably going to ignore this.
12. INFO • Report that a perfectly ordinary event happened.
13. WARN
14. ERROR CC BY-SA https://www.flickr.com/photos/74157931@N00/
15. CRITICAL
16. Logging Levels • logging.debug('message') • logging.info('message') • logging.warn('message') • logging.error('message') • logging.critical('message') • logging.basicConfig( level=logging.DEBUG) • logging.disable(logging.CRITICAL)
17. Logging to Screen and File logging.basicConfig( level=logging.DEBUG) logging.basicConfig( filename='logfile.txt', level=logging.DEBUG)
18. Following log files in real time tail –f <filename> (On Windows, install Cygwin.) Back to the live demo…
20. DEBUGGING
21. import pdb; pdb.set_trace() Back to the live demo…
22. Where am I? l . (List) w (Where)
23. Print Values p <variable or expression>
24. Stepping s (Step Into) n (Next / step over) r (Return / step out) c (continue running until next breakpoint)
25. Breakpoints • b (list all breakpoints & their numbers) • b <line number> • b <function name> • cl <breakpoint number> (clear bp) • import pdb; pdb.set_trace()
26. Quit the debugger q
27. import pdb; pdb.set_trace()
28. TESTING
30. import doctest doctest.testmod() Back to the live demo…
32. • Doctests are not a replacement for unit tests. • Doctests are unit tests for documentation. • Back to the live demo…
33. if __name__ == '__main__': import doctest doctest.testmod()
35. Logging import logging logging.basicConfig( level=logging.DEBUG) logging.debug('The log message.') logging.disable(logging.CRITICAL)
36. Debugging import pdb; pdb.set_trace()
37. Testing import doctest doctest.testmod()
39. LOGGING AND TESTING AND DEBUGGING, OH MY! @AlSweigart al@inventwithpython.com (last name rhymes with “why dirt”) bit.ly/ohmypy