The bot created this issue to inform you that pyup.io has been set up on this repo.
Once you have closed it, the bot will open pull requests for updates as soon as they are available.

It does not work. No errors, no warnings, no tests running.

 pytest --picked

Changed test files... 0. []
Changed test folders... 0. []
Test session starts (platform: darwin, Python 3.9.10, pytest 6.2.5, pytest-sugar 0.9.4)
django: settings: dmp.settings (from ini)
rootdir: /Users/hipertracker/dev/app, configfile: pytest.ini
plugins: Faker-11.3.0, picked-0.4.6, xdist-2.5.0, forked-1.4.0, html-3.1.1, django-3.10.0, sugar-0.9.4, metadata-1.11.0, notifier-1.0.4, testmon-1.2.2
collecting ...

Results (0.01s):

pip freeze > requirements.lock.txt

...
pytest==6.2.5
pytest-django==3.10.0
pytest-forked==1.4.0
pytest-html==3.1.1
pytest-metadata==1.11.0
pytest-notifier==1.0.4
pytest-picked==0.4.6
pytest-sugar==0.9.4
pytest-testmon==1.2.2
pytest-watch==4.2.0
pytest-xdist==2.5.0
python-dateutil==2.8.2

$ python -V
Python 3.9.10

This is a meta issue to organize issues and the repo for a beginner-friendly workshop for Pyladies Berlin.

Within the current implementation, it removes 8 leading chars from git diff --name-status result for each candidate. However on my Mac for some reason it is \t between the modification flag and the file name. For example, instead of
M <File Path>
it is actually M\t<File Path>
I guess it is something configurable but we'd better handle it since there are must be some people with same git settings as I do.

To reproduce it, run pytest in this repo:

tests/test_pytest_picked.py: 13 tests with warnings
  /home/ana/workspace/pytest-picked/venv/lib/python3.8/site-packages/pytest-5.4.1-py3.8.egg/_pytest/terminal.py:287: PytestDeprecationWarning: TerminalReporter.writer attribute is deprecated, use TerminalReporter._tw instead at your own risk.
  See https://docs.pytest.org/en/latest/deprecations.html#terminalreporter-writer for more information.
    warnings.warn(

-- Docs: https://docs.pytest.org/en/latest/warnings.html

Hi Ana, I'd be interested in taking over as a maintainer. Thanks!

I'm trying to run pytest in docker container. I always use pytest in my projects with docker.
Is the .git folder missing in the container?

This is my .dockerignore:

.*
!.coveragerc
!.env
!.pylintrc

This is traceback:

INTERNALERROR> Traceback (most recent call last):
INTERNALERROR>   File "/usr/local/lib/python3.6/site-packages/_pytest/main.py", line 174, in wrap_session
INTERNALERROR>     config._do_configure()
INTERNALERROR>   File "/usr/local/lib/python3.6/site-packages/_pytest/config/__init__.py", line 593, in _do_configure
INTERNALERROR>     self.hook.pytest_configure.call_historic(kwargs=dict(config=self))
INTERNALERROR>   File "/usr/local/lib/python3.6/site-packages/pluggy/hooks.py", line 306, in call_historic
INTERNALERROR>     res = self._hookexec(self, self.get_hookimpls(), kwargs)
INTERNALERROR>   File "/usr/local/lib/python3.6/site-packages/pluggy/manager.py", line 67, in _hookexec
INTERNALERROR>     return self._inner_hookexec(hook, methods, kwargs)
INTERNALERROR>   File "/usr/local/lib/python3.6/site-packages/pluggy/manager.py", line 61, in <lambda>
INTERNALERROR>     firstresult=hook.spec.opts.get("firstresult") if hook.spec else False,
INTERNALERROR>   File "/usr/local/lib/python3.6/site-packages/pluggy/callers.py", line 208, in _multicall
INTERNALERROR>     return outcome.get_result()
INTERNALERROR>   File "/usr/local/lib/python3.6/site-packages/pluggy/callers.py", line 80, in get_result
INTERNALERROR>     raise ex[1].with_traceback(ex[2])
INTERNALERROR>   File "/usr/local/lib/python3.6/site-packages/pluggy/callers.py", line 187, in _multicall
INTERNALERROR>     res = hook_impl.function(*args)
INTERNALERROR>   File "/usr/local/lib/python3.6/site-packages/pytest_picked/plugin.py", line 45, in pytest_configure
INTERNALERROR>     picked_files, picked_folders = mode.affected_tests()
INTERNALERROR>   File "/usr/local/lib/python3.6/site-packages/pytest_picked/modes.py", line 12, in affected_tests
INTERNALERROR>     raw_output = self.git_output()
INTERNALERROR>   File "/usr/local/lib/python3.6/site-packages/pytest_picked/modes.py", line 31, in git_output
INTERNALERROR>     output = subprocess.run(self.command(), stdout=subprocess.PIPE)
INTERNALERROR>   File "/usr/local/lib/python3.6/subprocess.py", line 403, in run
INTERNALERROR>     with Popen(*popenargs, **kwargs) as process:
INTERNALERROR>   File "/usr/local/lib/python3.6/subprocess.py", line 709, in __init__
INTERNALERROR>     restore_signals, start_new_session)
INTERNALERROR>   File "/usr/local/lib/python3.6/subprocess.py", line 1344, in _execute_child
INTERNALERROR>     raise child_exception_type(errno_num, err_msg, err_filename)
INTERNALERROR> FileNotFoundError: [Errno 2] No such file or directory: 'git': 'git'

There are certain scenarios that git is not available at the time of execution, so I guess it will be beneficial to have a new "File" mode, which means it reads change set from a text file instead of git diff.
For example, in my infra setup, tests are built into Docker images and deployed to K8s for execution. .git fold is dropped during image building. Even the .git folder is retained the K8s pod does not have permission to access git.
My current workaround is to run git diff at the time of image building and save the result to a text file in the image. If the plugin has the capability to read changes from that text file then I can directly use this plugin. Currently I have to write my own plugin.

Things that we need to do:

• How to configure the development environment and run the tests
• Explain usage of each mode

This can also be break down into different PRs.

Hi,
How can I run picked tests after commits ?
the scenario : commit on branch feature, and run tests pytest --picked --mode=branch --parent-branch=main

Thank you

pytest-picked looks great, but I often have 10s or even hundreds of tests in a single file.

It would be great therefore if this library could pick specific test rather than whole files.

I imagine this would be possible by introspecting git diff or similar.

Hi. I was just looking through the codebase to familiarize myself with how it works, and I found this bug- It collects files that have test in the name, including something with test in the middle of the word:

╰─ touch sample/intestine.py 
╰─ ls sample 
__init__.py  intestine.py  __pycache__  settings.py  urls.py  wsgi.py
╰─ pytest --picked 
======================================================== test session starts =========================================================
platform linux -- Python 3.6.5, pytest-3.6.0, py-1.5.3, pluggy-0.6.0
rootdir: /home/misterrios/Projects/sample/sample, inifile:
plugins: picked-0.1.0
collecting 0 items                                                                                                                   
Changed test files... 1. ['sample/intestine.py']
Changed test folders... 1. ['.pytest_cache/']
collected 0 items                                                                                                                    

==================================================== no tests ran in 0.05 seconds ====================================================

Pytest discovers tests by checking for test_*.py or *_test.py
See: https://docs.pytest.org/en/latest/goodpractices.html#test-discovery
So, a possible solution would be to use the built-in module pathlib to check the filenames and the file suffix combined with startswith and endswith. See: https://docs.python.org/3/library/pathlib.html#pathlib.PurePath.name

Being able to run pytest --picked tests/unit/.
There are many use cases when it could be handy, in my case it is being able to run each test suite separately, or part of them: unit, integration and etc.
Easy one, I will be happy to contribute but would like to get feedback on the feature before spending time.

When running with --cov and 100% coverage required, picked runs only a subset of tests, and coverage fails.

Long term it would be nice to make the coverage ignore all files not modified, so the coverage percentage is of the files being tested, however relationship between files modified and test coverage will make that very difficult.

But a quick fix is to have a mode that does run all of the tests, but reorders the tests to run the guessed modified test cases first, surfacing any errors sooner in the test suite.

Let's say you've changed house/serializers.py. Would be nice to run the tests of house/test_serializers.py or even the tests that were related.

To keep in mind:

• There is an option on pytest to specify test paths

Hello, thank you for making this library.
How can I set the exit code to 5 instead of 1 when no tests are collected exit-codes.
Thanks

I recently got interested in full-featured Test Impact Analysis in Python and I was thinking about creating a Python package python-tia which provides TIA functionality as new fun project. pytest is my Python test framework of choice as well. But I think TIA has a lot of potential which justifies the overhead of putting the core functionality of TIA into a separate package. This would enable integration of TIA with other test runners than pytest, other version control systems than git and probably even other coverage analysis tools than coverage. Are you interested in collaborating? If yes: Properly structured core functionality of pytest-picked could be refactored into python-tia leaving just the wrapping part in here...

Things to accomplish here:

• Remove "TODO" session
• Add gif with the demonstration of how to use it
• Add some examples with git status output
• Add badge for pypi
• Replace _ picked_ for pytest-picked on LICENSE session

See what is needed here.

Tools:

• https://github.com/twisted/towncrier

As an extension to #4 , I would also suggest to get the correct pytest naming conventions (pytest_classes , pytest_function and pytest_files) and check with the correct name instead of assuming test

Pytest check should be accessible from the fixture - https://docs.pytest.org/en/latest/reference.html#pytestconfig

Hi. It seems that at the moment this (very cool!) project isn't usable when run with --mode=branch using GitHub Actions checkout action (the start of pretty much every GHA workflow). If I attempt to use the something like the following

on:
  push:

jobs:
  test:

    runs-on: ubuntu-latest

    steps:
    - uses: actions/checkout@v2
      with:
        fetch-depth: 0
    - name: Set up Python 3.8
      uses: actions/setup-python@v2
      with:
        python-version: 3.8
    - name: Install dependencies
      run: |
        python -m pip install --upgrade pip setuptools wheel
        python -m pip install --ignore-installed -U -q --no-cache-dir -e .
        python -m pip list
    - name: Test with pytest
      run: |
        python -m pytest -r sx --picked=first --mode=branch

I'll get

fatal: ambiguous argument 'master': unknown revision or path not in the working tree.
Use '--' to separate paths from revisions, like this:
'git <command> [<revision>...] -- [<file>...]'

in the printed pytest logs. This is because the checkout action will checkout the current branch and then put the project under remotes/origin/. So adding in

    - name: Check what branches are available
      run: |
        git branch -a

would show something like

* name-of-branch-being-tested-in-pr
  remotes/origin/master

Would it be possible to look for some pattern that could match this if master isn't detected? Or allow for a branch name to given as an argument to test relative to?

Hi, love the package, it's super useful!

I do have one feature proposal - at my org we've migrated all our repositories to use main as the primary branch (as opposed to master), meaning that by default using --mode=branch doesn't work, one has to always specify --parent-branch=main, which gets a bit old. I would love if there was a way to permanently set the default parent branch, maybe through an environment variable or configuration in pytest.ini or similar?

Thanks

Let's say you are working on a feature branch and would like to run all the tests that were changed when compared with the master. To check it, you can run: git diff --name-only master.

Would be nice to have this option on picked. I'm wondering if something like pytest --picked=branch would be the best way to go.

After merging a PR from dependencies or docs updates, for instance, the workflow "Publish to Test PyPi" will be triggered and it will fail because the version on setup.py is not bumped. This workflow is nice to have because it will generate a beta version so we can test and avoid surprises when the package is published to PyPi in production. Would be nice to skip the workflow when the setup.py is not updated.