Continuous unit testing with Pydev (Python and Eclipse)

Question:

Is there a way to integrate background unit tests with the Pydev Eclipse environment?

My unit tests run well, but I would like to integrate them to run in the background based on source file changes (e.g. with nose) and to integrate the result back to Eclipse (I’m thinking big red X when tests fail with a console and trace log view).

No, a command prompt running nose on the side does not count.

I have had this Eclipse integration when developing RoR stuff.

Thanks,

Tal.

EDIT: Check out the new Pydev (1.6.4) http://pydev.org/manual_adv_pyunit.html

Asked By: Tal Weiss

||

Answers:

Pydev does have some unit-test integration, but that’s only as a run configuration…so…

This is not a very elegant way, but if you:

  1. Enable Project->Build Automatically
  2. In your project properties, add a new builder of type Program
  3. Configure it to run your tests and select ‘during auto builds’

Then at least you will get something that outputs the test results to the console on resource saves.

Answered By: Henrik Gustafsson

I just realized that PyDev has rather powerful scripting support. Unfortunately I don’t have the time to do it all for you (but if you complete this, please post it here 🙂

If you create a file named pyedit_nose.py that looks like this in an otherwise empty folder :

assert cmd is not None
assert editor is not None

if cmd == 'onSave':
    from java.lang import Runtime
    from java.io import BufferedReader
    from java.io import InputStreamReader

    from org.eclipse.core.resources import ResourcesPlugin
    from org.eclipse.core.resources import IMarker
    from org.eclipse.core.resources import IResource

    proc = Runtime.getRuntime().exec('ls -al')
    extra_message = BufferedReader(InputStreamReader(proc.inputStream)).readLine()

    r = ResourcesPlugin.getWorkspace().getRoot()
    for marker in r.findMarkers(IMarker.PROBLEM, False, IResource.DEPTH_INFINITE):
        if marker.getAttribute(IMarker.MESSAGE).startsWith("Some test failed!"):
            marker.delete()

    for rr in r.getProjects():
        marker = rr.createMarker(IMarker.PROBLEM)
        marker.setAttribute(IMarker.MESSAGE, "Some test failed! " + extra_message)
        marker.setAttribute(IMarker.PRIORITY, IMarker.PRIORITY_HIGH)
        marker.setAttribute(IMarker.SEVERITY, IMarker.SEVERITY_ERROR)

and set up Preferences->PyDev->Scripting Pydev to point to this directory you will get all projects in your workspace marked with an error every time a file is saved.

By executing a script that returns the test results in some easy to parse format rather than ls and parsing the output you should be able to put your markers in the right places.

See this for some starting points:

  • Jython Scripting in Pydev
  • IMarker is what represents a marker.
  • IResource is what you attach your markers to. Can be workspaces, projects, files, directories etc. resource.createMarker(IMarker.PROBLEM) creates a problem marker.
  • IProject is a type of IResource that represents a project. Use the members() method to get the contents.
Answered By: Henrik Gustafsson

I use Nosy (available on pypi):

Run the nose test discovery and execution tool whenever a source file
is changed.

Answered By: leo

I enhanced the “nosy” script to automatically build documentation and runs tests
continuously. Nothing stellar, but gets the job done. Posting it here because the original
link went down. Unlike the original nosy script, this one scans the directory recursively
and allows looking for multiple patterns.

import os
import os.path
import sys
import stat
import time
import subprocess
from fnmatch import fnmatch


def match_patterns(pathname, patterns):
    """Returns True if the pathname matches any of the given patterns."""
    for pattern in patterns:
        if fnmatch(pathname, pattern):
            return True
    return False


def filter_paths(pathnames, patterns=["*"], ignore_patterns=[]):
    """Filters from a set of paths based on acceptable patterns and
    ignorable patterns."""
    result = []
    if patterns is None:
        patterns = []
    if ignore_patterns is None:
        ignore_patterns = []
    for path in pathnames:
        if match_patterns(path, patterns) and not match_patterns(path, ignore_patterns):
            result.append(path)
    return result


def absolute_walker(path, recursive):
    if recursive:
        walk = os.walk
    else:
        def walk(path):
            return os.walk(path).next()
    for root, directories, filenames in walk(path):
        yield root
        for directory in directories:
            yield os.path.abspath(os.path.join(root, directory))
        for filename in filenames:
            yield os.path.abspath(os.path.join(root, filename))


def glob_recursive(path, patterns=["*"], ignore_patterns=[]):
    full_paths = []
    for root, directories, filenames in os.walk(path):
        for filename in filenames:
            full_path = os.path.abspath(os.path.join(root, filename))
            full_paths.append(full_path)
    filepaths = filter_paths(full_paths, patterns, ignore_patterns)
    return filepaths


def check_sum(path='.', patterns=["*"], ignore_patterns=[]):
    sum = 0
    for f in glob_recursive(path, patterns, ignore_patterns):
        stats = os.stat(f)
        sum += stats[stat.ST_SIZE] + stats[stat.ST_MTIME]
    return sum


if __name__ == "__main__":
    if len(sys.argv) > 1:
        path = sys.argv[1]
    else:
        path = '.'

    if len(sys.argv) > 2:
        command = sys.argv[2]
    else:
        command = "make -C docs html; bin/python tests/run_tests.py"

    previous_checksum = 0
    while True:
        calculated_checksum = check_sum(path, patterns=['*.py', '*.rst', '*.rst.inc'])
        if calculated_checksum != previous_checksum:
            previous_checksum = calculated_checksum
            subprocess.Popen(command, shell=True)
            time.sleep(2)

Hope this helps.

=)

Answered By: GoraKhargosh

I run the test by hand the first time (Run > Run As > Python unit test). After that, I use
Ctrl+Shift+F9 to have the files saved and the tests executed, instead of saving with Ctrl+S and expecting some magic to happen.

The Ctrl+Shift+F9 key combination relaunches the last run configuration.

Disclaimer: I’m new to Eclipse and to PyDev, so I may be suggesting something silly/obvious/wrong

Answered By: xverges

This feature has been added to PyDev 2.0.1 with an option to relaunch the tests in the last test run whenever a python file change, with an additional option to rerun only the errors — although it’ll run the full test suite if no errors were found, as the idea is that you work through your errors and when all pass a final launch for the whole suite is done (then you can move on to another task).

The current nightly build has this feature incorporated.

Picture with new action

Answered By: Fabio Zadrozny
Categories: questions Tags: , ,
Answers are sorted by their score. The answer accepted by the question owner as the best is marked with
at the top-right corner.