Python Code Coverage and Multiprocessing

Question

I use coveralls in combination with coverage.py to track python code coverage of my testing scripts. I use the following commands:

coverage run --parallel-mode --source=mysource --omit=*/stuff/idont/need.py ./mysource/tests/run_all_tests.py
coverage combine
coveralls --verbose

This works quite nicely with the exception of multiprocessing. Code executed by worker pools or child processes is not tracked.

Is there a possibility to also track multiprocessing code? Any particular option I am missing? Maybe adding wrappers to the multiprocessing library to start coverage every time a new process is spawned?

EDIT:

I (and jonrsharpe, also :-) found a monkey-patch for multiprocessing.

However, this does not work for me, my Tracis-CI build is killed almost right after the start. I checked the problem on my local machine and apparently adding the patch to multiprocessing busts my memory. Tests that take much less than 1GB of memory need more than 16GB with this fix.

EDIT2:

The monkey-patch does work after a small modification: Removing the config_file parsing (config_file=os.environ['COVERAGE_PROCESS_START']) did the trick. This solved the issue of the bloated memory. Accordingly, the corresponding line simply becomes:

cov = coverage(data_suffix=True)

Paule

Do you not test the code for those child processes directly? – Boy

Well, yes most of it I do. But there are certain parts that are only useful and are only executed in case multiprocessing is used (like wrapping data base access with locks or a multiprocessing queue to enforce serial data storage). And I know myself that this code is working due to the successful tests. It just would be nice if this would also show up on coveralls :-) – Paule

See bitbucket.org/ned/coveragepy/issue/117/…, via nedbatchelder.com/code/coverage/trouble.html – Boy

Thanks, I stumbled across this as well. However, the monkey-patch does not work for me. Adding this to my script kills my Travis-CI built almost instantly. I checked this on my local machine as well. Apparently, the monkey patch busts my memory. Coverage allocates more than 16GB of memory for tests that usually need much less than 1GB. – Paule

@Paule I'm very interested to hear about your experiences. The idea of removing the config file parsing seems very odd: I can't imagine how parsing the config file would radically change the memory footprint. What is the value of COVERAGE_PROCESS_START for you? Do you have a .coveragerc file? Drop me an email if you want to dig into it. – Armond

Hey, nope I am not using a coveragerc file. So I rechecked and, accordingly, the config_file=os.environ['COVERAGE_PROCESS_START'] throws a key error. However, the main program continues to run and spawns new processes. Maybe this was the reason why the memory was garbled? – Paule

Answer 1

Coverage 4.0 includes a command-line option --concurrency=multiprocessing to deal with this. You must use coverage combine afterward. For instance, if your tests are in regression_tests.py, then you would simply do this at the command line:

coverage run --concurrency=multiprocessing regression_tests.py
coverage combine

Passepartout

Thanks for pointing out the requirement to use coverage combine afterward. I had been spinning my wheels for a while trying to figure out why the concurrency=multiprocessing that I had in my .coveragerc file wasn't working. – Murchison

Could you add a simple basic example to this answer? It would help me out a lot. – Markley

If you're using the pytest-cov plugin you'll need specify concurrency = multiprocessing in one of the config location (pyproject.toml, .converagerc, etc) in the [run] section. – Generalissimo

The correct syntax is concurrency = ["multiprocessing"] – Elodiaelodie

Answer 2

I've had spent some time trying to make sure coverage works with multiprocessing.Pool, but it never worked.

I have finally made a fix that makes it work - would be happy if someone directed me if I am doing something wrong.

https://gist.github.com/andreycizov/ee59806a3ac6955c127e511c5e84d2b6

Epenthesis

I believe this link here might help you: pytest-cov.readthedocs.io/en/latest/… - it seems you need to call join in the threads after to avoid a few issues. – Presbyterial

Answer 3

One of the possible causes of missing coverage data from forked processes, even with concurrency=multiprocessing, is the way of multiprocessing.Pool shutdown. For example, with statement leads to terminate() call (see __exit__ here). As a consequence, pool workers have no time to save coverage data. I had to use close(), timed join() (in a thread), terminate sequence instead of with to get coverage results saved.

Exuviae

Yep, I had the same issue with using with. I had a blocking operation (starmap) so I just did a close then join inside the with block. – Hallie