« Is your code a novel?Overheard at lunch »

Code Coverage with CherryPy 2.1


Permalink 12:05:47 pm, by fumanchu Email , 785 words   English (US)
Categories: Python, CherryPy

Code Coverage with CherryPy 2.1

CherryPy1 helps with both the collection and the analysis of coverage data (for a good introduction to code coverage, see bullseye.com). Now, I'm a visual learner, so I'm going to skip right to the screenshot and explain it in detail afterward. This is a browser session with two frames: a menu frame on the left and a file frame on the right. Clicking on one of the filenames in the menu will show you that file, annotated with coverage data, in the right-hand frame. This stats-browser is included with CherryPy, and can be used for any application, not just CherryPy or CP apps.

1 All of this is present in CherryPy 2.1 beta, revision 543. Get it via SVN

coverage stats browser session

Collection of coverage statistics

You need to start by obtaining the coverage.py module, either the original from Gareth Rees, or Ned Batchelder's updated version. Drop it in site-packages.

Covering CherryPy

If you're collecting coverage statistics for CherryPy itself, just run the test suite with the --cover option. Coverage data will be collected in cherrypy/lib/coverage.cache. Example:

mp5:/usr/lib/python2.3/site-packages# python cherrypy/test/test.py --cover

Covering CherryPy applications

If you write a test suite for your own applications, build it on top of the tools present in cherrypy/test. Here's a minimal example:

import os, sys
localDir = os.path.dirname(__file__)
dbpath = os.path.join(localDir, "db")

from cherrypy.test import test

if __name__ == '__main__':
    # Place our current directory's parent (myapp/) at the beginning
    # of sys.path, so that all imports are from our current directory.
    curpath = os.path.normpath(os.path.join(os.getcwd(), localDir))
    sys.path.insert(0, os.path.normpath(os.path.join(curpath, '../../')))

    testList = ["test_directory",
    testConf = os.path.join(localDir, "test.conf")

By using the TestHarness from CherryPy's test suite, you automatically get access to the --cover command-line arg (and --profile and all the others, too, but that's for another day). Again, coverage data will be collected in cherrypy/lib/coverage.cache by default.

Covering Other Applications

You can use the stats-browser even if you don't use the CherryPy framework to develop your applications. Just use coverage.py as it was originally intended:

coverage.py -x yourapp.py

The coverage data, in this case, will be collected by default into a .coverage file. You need to tell the stats-server where this file is (see below). Note that successive manual calls to coverage.py will accumulate stats; the CherryPy test suite, in contrast, erases the data on each run.

Analysis of coverage statistics

Once you've got coverage data sitting around in a file somewhere, it's a snap to have CherryPy serve it in your browser. If you're covering the CherryPy test suite, or your own CP app using CP's TestHarness (see above), just execute:

mp5:/usr/lib/python2.3/site-packages# python cherrypy/lib/covercp.py

Then, point your browser to http://localhost:8080, and you should see an image similar to the above.

By default, the server reads coverage data from cherrypy/lib/coverage.cache, the same file our collector wrote to by default. If you covered your own application and collected the data in another file, you can supply that path as a command-line arg:

# python cherrypy/lib/covercp.py /path/to/.coverage 8088

If you supply a second arg, as in this example, it will change the port for you (from the default of 8080).

You need to stop (Ctrl-C) and restart the server if you recollect coverage data.

The interface

Each file in the menu has coverage stats, and is a hyperlink; click on one, and the file frame will show you the file contents, annotated with coverage data. Lines that start with ">" were touched, and those that start with "!" were not.

Click the "Show %" button to show a "percent covered" figure for each file. This can take a long time if you have lots of files, so it's best to first restrict your view using the directory links. Each directory is a hyperlink; click on one to restrict the menu to that folder only. Percentages below the "threshold" value will be shown in red. The "Show %" feature isn't "sticky", by the way; that is, if you click on a different directory link, or refresh the page, the figures will disappear. That's a necessary evil due to the slowness of generating percentages for many files. Just hit the "Show %" button again as needed.

As you can see from the screenshot, I've got some more tests to write! Hope you find this tool as useful as I do. :)

No feedback yet

Leave a comment

Your email address will not be revealed on this site.

Your URL will be displayed.

Please enter the phrase "I am a real human." in the textbox above.
(Line breaks become <br />)
(Name, email & website)
(Allow users to contact you through a message form (your email will not be revealed.)
January 2015
Sun Mon Tue Wed Thu Fri Sat
 << <   > >>
        1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30 31


The requested Blog doesn't exist any more!

XML Feeds

powered by b2evolution