Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Reformat test structure and imports #50

Closed
wants to merge 1 commit into from

Conversation

TylerYep
Copy link

@TylerYep TylerYep commented Sep 27, 2020

Hi, I was interested in using this project, but ran into some issues with pip installing the package. I ended up renovating the project to fit my needs, and thought there was a chance you might be interested in using the changes.

Please feel free to take these changes or throw them away, I'm mainly just leaving this PR here as a suggestion of possible improvements!

Changes included:

  • Run isort .
  • Run black .
  • Rename didyoumean files to make imports easier
  • Add a separate tests/ folder to avoid mixing source files with test files
  • Resolve several deprecation warnings that are raised when running pytest

@SylvainDe
Copy link
Owner

Thank you so much for your pull request!
I remember that I struggled trying to make this installable with pip for a very limited success ( #3 ).

I'll have a look at the details and integrate it as soon as possible.

Copy link
Owner

@SylvainDe SylvainDe left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thank you so much for the hard work you've put inthis contribution!

I have not gone through the whole review yet.

My first concern would be to ensure that the CI runs smoothly on all Python versions.
Currently, the pre-checks are failing because of the line length - the best option is to change the pycodestyle limit in the travis configuration.
Also, I suspect other tests would fail at runtime (at least on the older Python versions)

If you decide to spend more time on this, I would appreciate if things could be split into pull-requests with a somehow clearer scope. For instance, one for black (and typo), one for the reorganisation, etc. If you can't be bothered, that's fine as well.

Let me know if there's anything I may help you with

All the best!

@@ -110,6 +113,3 @@ def didyoumean_disablehook():
"""Function to set hooks to their normal value."""
sys.excepthook = sys.__excepthook__
set_ipython_custom_exc(None)

# NOTE: It could be funny to have a magic command in Python
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do you know why this got removed ?

self.assertRegexpMatches(name, regex)
for name in ['1a']:
self.assertNotRegexpMatches(name, regex)
# def test_var_name(self):
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sorry for that one - I guess it caused you troubles.
Would you remember which name led to problems by any chance ?

@@ -31,7 +32,7 @@ def re_matches(self, text, regexp, results):
and groupdict().
"""
groups, named_groups = results
self.assertRegexpMatches(text, regexp) # does pretty printing
self.assertRegex(text, regexp) # does pretty printing
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The reasoning for this was to ensure the lib was usable and testable under all Python version >= 2.6.
Indeed, somes of the reasons caught and handled are somehow related to running a piece of Python code with the wrong version of the interpreter (not as common now but a few years back, using 2.7 instead of 3.0+ or the other way round was a common gotcha).
Would it be possible to revert that part of the patch ?

r"^(?:local|free) variable '(?P<name>{0})' "
r"referenced before assignment(?: in enclosing scope)?$".format(VAR_NAME)
)
NAMENOTDEFINED_RE = r"^(?:global )?name '(?P<name>{0})' " r"is not defined$".format(
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Now, this line (and a few others) are considered too long by the call to pycodestyle on travis.
Would it be possible to either:

  • tweak the travis config to disable the corresponding warning ( E501 )
  • run black with a different line length limit
    (I think option 1 is both an easier and cleaner option)

https://travis-ci.org/github/SylvainDe/DidYouMean-Python/jobs/730786126

didyoumean/internal.py Outdated Show resolved Hide resolved
@SylvainDe
Copy link
Owner

Note: I don't know why this does not appear more explicitly here but you can find the Travis runs for your Pull Request here: https://travis-ci.org/github/SylvainDe/DidYouMean-Python/builds/730786103 .

@TylerYep
Copy link
Author

Are you dedicated to supporting Python 2.6 at this point? All versions of up to Python 3.5 have reached end-of-life by now (https://endoflife.date/python), so it seems wise to move along with this trend - especially if this project is not currently used by many people yet.

It would be pretty easy to split this into multiple changes - I can do it sometime this week. Thanks for reviewing the changes!

@TylerYep TylerYep force-pushed the master branch 5 times, most recently from 24ebcef to 7aba306 Compare September 28, 2020 19:51
@TylerYep TylerYep changed the title Run Black + isort, reformat test structure Reformat test structure and imports Sep 28, 2020
@SylvainDe
Copy link
Owner

This looks great! Thanks for the additional work you've gone through.

Unfortunately, I can see that the CI is failing - https://travis-ci.org/github/SylvainDe/DidYouMean-Python/pull_requests . Would you have any idea about what happened ?

As for the support for older Python version, I'd rather keep it until it does bring a lot more troubles while moving forward. If it is only for DeprecationWarning, I'd rather add a few wrappers around the unit-test methods to ensure that everything works smoothly in all cases.

SylvainDe added a commit that referenced this pull request Oct 4, 2020
Correct spelling for "dictionnary" is "dictionary".
This is one of the things hilighted in:
#50
@SylvainDe
Copy link
Owner

Hello @TylerYep , I've performed a few changes which will unfortunately lead to merge issues.
On the positive side: the DeprecationWarnings are now fixed and I've removed a few pep8 errors so that Black can be used safely.

To you mind resubmitting your code organisation PR ? (If you want, I can perform it myself based on the work you've done and try to give you the proper attribution). Similarly, I can perform the call to Black in a different commit if you want.

Please let me know how you want to handle these and thanks for your work!

@TylerYep
Copy link
Author

TylerYep commented Oct 4, 2020

Resubmitted the PR. I can go ahead and create the PR for the black formatting, but could you fix the travis CI integration first?
I believe you just need to:

  • Add the Travis-CI app in the Github Marketplace
  • Add the DidYouMean project to TravisCI (I believe this is already done)
  • Remove the scrutinizer app from this project

As a side note, I would also really recommend upgrading to a newer toolchain for linting + testing. I believe the more common standard now is pytest + flake8 instead of unittest + pep8/pycodestyle/pyflakes.

Pytest should be able to execute all unittests without any modifications, but with a better test-discovery. The following commands should do everything correctly and concisely in the .travis.yml:

isort .
black .
flake8 .
pytest
codecov

@SylvainDe
Copy link
Owner

Thanks for all the suggestions ! It looks like I have not followed the latest trends on a few aspects indeed.
I've disabled scrutinizer for all PRs.
As for Travis, a few things changed (apps, migration from org to com) and I did not quite manage to make it work yet. I'll try again later on.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants