How do I run the tests in goodreads?

[steinarb]

I don't have goodreads account to export bookshelves from.

Therefore I am writing this project https://github.com/steinarb/bokbase which is a simple book database that creates goodreads.csv files that can be imported into bookwyrm.

Unfortunately the import of CSV files produced by bokbase into bookwyrm fails.

So then my idea was to replace the goodreads.csv in bookwyrm/bookwyrm/test/data/ with my own file, and use that to create the correct format for bookwyrm.

Unfortunately running the test is hard. After installing all dependencies required for the tests, I still get the following error:

bookwyrm/tests/views/preferences/test_delete_user.py:10: in <module>
    from bookwyrm import forms, models, views
bookwyrm/forms.py:11: in <module>
    from bookwyrm import models
bookwyrm/models/__init__.py:5: in <module>
    from .book import Book, Work, Edition, BookDataModel
bookwyrm/models/book.py:12: in <module>
    from imagekit.models import ImageSpecField
/usr/lib/python3/dist-packages/imagekit/__init__.py:2: in <module>
    from . import conf
/usr/lib/python3/dist-packages/imagekit/conf.py:6: in <module>
    class ImageKitConf(AppConf):
/usr/lib/python3/dist-packages/appconf/base.py:75: in __new__
    new_class._configure()
/usr/lib/python3/dist-packages/appconf/base.py:101: in _configure
    value = getattr(obj._meta.holder, prefixed_name, default_value)
/usr/local/lib/python3.9/dist-packages/django/conf/__init__.py:84: in __getattr__
    self._setup(name)
/usr/local/lib/python3.9/dist-packages/django/conf/__init__.py:65: in _setup
    raise ImproperlyConfigured(
E   django.core.exceptions.ImproperlyConfigured: Requested setting IMAGEKIT_CACHEFILE_NAMER, but settings are not configured. You must either define the environment variable DJANGO_SETTINGS_MODULE or call settings.configure() before accessing settings.

Any ideas what I need to do to be able to run "pytest" without errors?

Thanks!

[mouse-reeve]

It looks like there's a missing dependency -- what command did you use to run the test?

[steinarb]

>>>> Mouse Reeve ***@***.***>: It looks like there's a missing dependency -- what command did you use to run the test?

I have tried two: pytest and python -m pytest Both give the same error messages. This is on a debian stable with some python dependencies installed with apt and some (not found as debian packages), installed with pip3 run as root. I have used apt or pip3 to fill a lot of missing dependencies warnings.

[steinarb]

Thanks, this command worked (added a "web" between "--rm" and "pytest"):

docker-compose run --rm web pytest bookwyrm/tests/importers/test_goodreads_import.py

Where are the file in the test picked up from?

I.e. if I copy my own file into bookwyrm/test/data/goodreads.csv and re-run the command, will the test be run with my csv file or the one already in the doker image?

Also, is there a way to see the console output from the test run?

Thanks again!

[mouse-reeve]

oh yes sorry! it should definitely have web in it. The tests open the files manually using relative paths, in the goodreads tests, it's datafile = pathlib.Path(__file__).parent.joinpath("../data/goodreads.csv"), so yes, replacing the goodreads.csv file should mean that it's working from the new file.

The console output seems to only show up when the tests fail, and is suppressed when they pass. There is probably a way around that, but I haven't found it; I usually just add raise Exception to the end of the test so that the test fails, which is janky but does get me the print statements. You can also add pdb statements to the test if you prefer a proper debugger.

[steinarb]

Hm... the tests fail on my .csv file, but the print statements don't appear in the output.

What's output for each failing test, is the code of the failing test, with a ">" pointing at the failing line, e.g. like so:

______________________________________________________________ GoodreadsImport.test_create_job _______________________________________________________________

self = <bookwyrm.tests.importers.test_goodreads_import.GoodreadsImport testMethod=test_create_job>
_ = (<MagicMock name='delay' id='140603341207392'>, <MagicMock name='delay' id='140603338347040'>, <MagicMock name='delay' id='140603338503504'>)
import_job = <ImportJob: ImportJob object (1)>, import_items = <QuerySet []>

    def test_create_job(self, *_):
        """creates the import job entry and checks csv"""
        import_job = self.importer.create_job(
            self.local_user, self.csv, False, "public"
        )

        import_items = models.ImportItem.objects.filter(job=import_job).all()
>       self.assertEqual(len(import_items), 3)
E       AssertionError: 0 != 3

[steinarb]

The reason nothing was output, was that the code I had put the debug trace in was never called, and the reason that code was never called, was that my test file had 0 rows in it.

So: my bad!

To get that far I discovered that adding the -s flag to pytest, would output all debug trace, interleaved with the pytest output:

docker-compose run --rm web pytest -s bookwyrm/tests/importers/test_goodreads_import.py

And then I traced slowly from the test into the code until I came to the point that failed.

[mouse-reeve]

I see - thank you for debugging, I was totally flummoxed as to why it wouldn't be printing