Starting to create more tests

[Jean-Baptiste-Camps]

No description provided.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (d8e8c9c) 28.70% compared to head (b005a17) 55.23%.
Report is 1 commits behind head on master.

Additional details and impacted files

@@             Coverage Diff             @@
##           master      #43       +/-   ##
===========================================
+ Coverage   28.70%   55.23%   +26.52%     
===========================================
  Files           2        3        +1     
  Lines         108      172       +64     
===========================================
+ Hits           31       95       +64     
  Misses         77       77               

Flag	Coverage Δ
unittests	55.23% <100.00%> (+26.52%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[florian-gandon]

I see you are doing some testing lately.
There is an usual way to format it which is given-when-then: https://martinfowler.com/bliki/GivenWhenThen.html
(my 2 cents)
Florian

[Jean-Baptiste-Camps]

Hi @florian-gandon ! Thanks for the input !
As I understand it, it is a way to document the contents of the test. For instance, if I take one of my existing tests,

# FEATURE: text normalisation before further processing
def test_normalise(self):
    # SCENARIO: I want to normalise a string of text with defaults settings
    # GIVEN the following string
    text = " Hello,  Mr. 𓀁, how are §§ you; doing?"
    # WHEN I apply normalisation
    result = superstyl.preproc.tuyau.normalise(text)
    # THEN I expect the following output in lower case and without punctuation and symbols
    expected_default = "hello mr how are you doing"
    self.assertEqual(result, expected_default)

Or do I miss something ?

[florian-gandon]

Yes, for instance.
You could also put it in the method name so when failing (or succeed) it is easier to read in the console or in the IDE. Like that:

def givenTextWithUpperCasePunctuationAndSymbols_whenNormalizing_thenTextIsLowerCaseWithoutPunctuation(self):
    # GIVEN
    text = " Hello,  Mr. 𓀁, how are §§ you; doing?"
    # WHEN
    result = superstyl.preproc.tuyau.normalise(text)
    # THEN
    expected_default = "hello mr how are you doing"
    self.assertEqual(result, expected_default)

(or optionally without comments as there is only one line for each part)
The method could be written in a file or a class which describes the feature. I'm not sure what is the usual manner in Python.

[Jean-Baptiste-Camps]

Nice ! I'm not using a lot of class oriented stuff right now, but for the tests, I'm using a class by feature of the main entry points (so a class for data loading, a class for training, etc.), in which I have a test function definition for each function of the source library.

[Jean-Baptiste-Camps]

Merging and closing this one as I need to implement and test a completely new function, but please give feedback whenever you want, it is much appreciated :)