Add mtherieau's Cloze code to genanki #37

dvklopfenstein · 2019-08-31T20:02:40Z

Please consider this pull request which adds support for Cloze cards to genanki.
I added the code from @mtherieau's gist example, as it works great.

Thank you creating this repo and sharing it with everyone and thank you for taking the time to consider this pull request.

genanki/model.py

kerrickstaley · 2019-09-24T04:28:35Z

genanki/note.py

-      if op(self.fields[ord_] for ord_ in required_field_ords):
-        rv.append(Card(card_ord))
-    return rv
+    if self.model.model_type == 0:


Could you define separate _front_back_cards() and _cloze_cards() methods so that you can keep each method shorter and more readable?

Definitely easier to read now that it is addressed with 4aff1c6

kerrickstaley · 2019-09-24T04:38:10Z

genanki/note.py

+          rv.append(Card(card_ord))
+      return rv
+    # returns a Card with unique ord for each unique cloze reference
+    assert self.model.model_type == 1, self.model.model_type


Instead of an assertion, I would use an if-statement and throw an exception if model_type is not 0 or 1.

Assertions aren't really intended for checking the validity of arguments that someone else passed to you (coming from code that you didn't write). They're more intended for checking the internal validity of your own code.

Regular exceptions are better because you can throw an specific type of exception (assert can only raise AssertionError) and because they are always run (assert statements are disabled when running in "optimized" mode).

Done: 7d59da9

kerrickstaley · 2019-09-24T04:50:40Z

genanki/note.py

+    assert self.model.model_type == 1, self.model.model_type
+    card_ords = set()
+    # find cloze replacements in first template's qfmt, e.g "{{cloze::Text}}"
+    cloze_replacements = set(


Instead of using a regex it may be better to look at Anki's code to see how it does it (probably uses a real template engine to parse the template and some sort of trick to figure out what the cards should be; see _req in this file for a similar thing). Regexes can be fragile and hard to debug. And also hard to review.

That said, I'm OK with shipping this as a v1 and waiting for the bug reports to roll in :)

just a quick note that although I agree that regexes can be fragile and difficult to debug, these regexes were actually lifted from anki: https://github.com/dae/anki/blob/eef86bf37ebf8c4f9eb4cee50b10aedecfe647aa/anki/models.py#L578

@mtherieau's regex exactly matches anki's code, so I think this is okay as is.

Sounds good.

kerrickstaley · 2019-09-24T05:01:46Z

tests/test_cloze.py

@@ -0,0 +1,70 @@
+#!/usr/bin/env python


Thanks for writing tests!

Could you remove the #!/usr/bin/env python and unset the executable bit? I think it may cause confusion for people who just want to run the test. If you want to execute this file you can always do it by running python3 tests/test_cloze.py.

Also side note: you shouldn't just put python in a shebang; you should always specify python2 or python3 unless it's compatible with both. There is a PEP that I'm very familiar with that explains why :)

Thanks for the link to the PEP. Great information.
The top line is removed and the exe bit is unset with 2f18899

kerrickstaley · 2019-09-24T05:09:11Z

tests/test_cloze.py

+
+  fields = ['NOTE TWO: {{c1::1st deletion}} {{c2::2nd deletion}} {{c3::3rd deletion}}', '']
+  my_cloze_note = Note(model=MY_CLOZE_MODEL, fields=fields)
+  assert {card.ord for card in my_cloze_note.cards} == {0, 1, 2}


Probably better to assert the sorted list is equal, i.e.

assert sorted(card.ord for card in my_cloze_note.cards) == [0, 1, 2]

Otherwise there could be duplicate ord values in my_cloze_note.cards and this test wouldn't catch them.

Missed that. Thanks. This is addressed with: c61a8bc

kerrickstaley · 2019-09-24T05:11:25Z

tests/test_cloze.py

+  for note in notes:
+    deck.add_note(note)
+  fout_anki = '{NAME}.apkg'.format(NAME=deckname)
+  Package(deck).write_to_file(fout_anki)


Could you avoid writing this file and printing to the console by default? People won't want this behavior when they're running tests.

You could put the behavior behind a flag, e.g. add a write_to_test_apkg=False flag and put if write_to_test_apkg right before deckname = 'mtherieau', and then run test_cloze(True) in your if __name__ == '__main__': block.

Done: 03523a4

kerrickstaley · 2019-09-24T05:20:47Z

genanki/note.py

+      field_index = next((i for i, f in enumerate(self.model.fields) if f['name'] == field_name), -1)
+      field_value = self.fields[field_index] if field_index >= 0 else ""
+      # update card_ords with each cloze reference N, e.g. "{{cN::...}}"
+      card_ords.update([int(m)-1 for m in re.findall(r"{{c(\d+)::.+?}}", field_value) if int(m) > 0])


nit: You can remove the [] here.

Done: 0fa99f6

kerrickstaley · 2019-09-24T05:28:02Z

tests/test_cloze.py

+  assert {card.ord for card in my_cloze_note.cards} == {0, 1, 2}
+  notes.append(my_cloze_note)
+
+  fields = ['NOTE THREE: {{c1::1st deletion::C1-CLOZE}}', '']


Could you add another test case where there are multiple Cloze deletions with the same index, e.g.

'NOTE FOUR: {{c1:1st deletion}} foo {{c2:2nd deletion}} bar {{c1::3rd deletion}}'

This is documented in the manual:

You can also elide multiple sections on the same card. In the above example, if you change c2 to c1, only one card would be created, with both Canberra and 1913 hidden.

Good catch. I am working on it...

dvklopfenstein · 2019-09-24T13:33:34Z

Thank you for considering the pull request and thank you for the useful suggestions. I will take a look in a little bit as I am under a deadline right now.

#37 (comment)

…oze_cards() #37 (comment)

#37 (comment)

kerrickstaley · 2019-10-10T17:58:28Z

genanki/note.py

+  def _cloze_cards(self):
+    """returns a Card with unique ord for each unique cloze reference"""
+    if self.model.model_type != self.model.CLOZE:
+      raise RuntimeError('Expected model_type CLOZE or FRONT_BACK')


ValueError would be better than RuntimeError here. That's the preferred exception to use when an argument passed to a function does not have a valid value.

For example, Python raises a ValueError when you try to do math.sqrt(-1):

In [1]: import math In [2]: math.sqrt(-1) --------------------------------------------------------------------------- ValueError Traceback (most recent call last) <ipython-input-2-2037b8d41d70> in <module> ----> 1 math.sqrt(-1) ValueError: math domain error

Good idea. Done: #37 (comment)

dvklopfenstein · 2019-11-14T23:21:11Z

Thank you for merging this. Is there anything else I need to do for completeness? Work sidelined me for a while.

Thank you for your code review and comments and adding to that last repeated field test. Your comments were constructive. If anything is else is needed for this PR, I have time to work on it tomorrow.

kerrickstaley · 2019-12-05T06:04:29Z

@dvklopfenstein sorry for the delay. I've published your changes to PyPI. Version 0.8.0 of genanki has your changes.

dvklopfenstein added 2 commits August 31, 2019 15:58

Add mtherieau's Cloze code to genanki #3

c65ea0b

Add color to hidden Cloze words

94b655e

lins05 mentioned this pull request Sep 15, 2019

Support set model type #39

Closed

kerrickstaley requested changes Sep 24, 2019

View reviewed changes

dvklopfenstein added 8 commits October 10, 2019 11:15

Added constants, FRONT_BACK = 0 and CLOZE = 1 to the Model class.

4046a80

#37 (comment)

improved readability by defining separate _front_back_cards() and _cl…

4aff1c6

…oze_cards() #37 (comment)

Check for expected model_type values FRONT_BACK(0) or CLOZE(1)

7d59da9

#37 (comment)

remove '#!/usr/bin/env python' and unset executable bit.

2f18899

Use sorted list in assertion to be sure we don't see duplicate ords.

c61a8bc

#37 (comment)

Only write anki flashcards to a apkg file if requested by user.

03523a4

#37 (comment)

Removed unecessary square brackets, []

0fa99f6

re-run CI

7c055c2

kerrickstaley requested changes Oct 10, 2019

View reviewed changes

dvklopfenstein and others added 5 commits October 10, 2019 21:42

Change RuntimeError to ValueError if model_type value is unexpected

06803b8

Fix test for repeated field case

dae7043

Capital / period

c67818e

Move ValueError to more readable place

a7c3149

add a few more tests

adf2a63

kerrickstaley merged commit 906c84e into kerrickstaley:master Nov 8, 2019

This was referenced Nov 8, 2019

Support Cloze note type #3

Closed

Seeking clarification as to which is the 'real' master branch #41

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add mtherieau's Cloze code to genanki #37

Add mtherieau's Cloze code to genanki #37

dvklopfenstein commented Aug 31, 2019

kerrickstaley Sep 24, 2019

dvklopfenstein Oct 10, 2019

kerrickstaley Sep 24, 2019

dvklopfenstein Oct 10, 2019

kerrickstaley Sep 24, 2019

mtherieau Oct 2, 2019

dvklopfenstein Oct 10, 2019

kerrickstaley Oct 10, 2019

kerrickstaley Sep 24, 2019

dvklopfenstein Oct 10, 2019

kerrickstaley Sep 24, 2019

dvklopfenstein Oct 10, 2019

kerrickstaley Sep 24, 2019

dvklopfenstein Oct 10, 2019

kerrickstaley Sep 24, 2019

dvklopfenstein Oct 10, 2019

kerrickstaley Sep 24, 2019

dvklopfenstein Oct 11, 2019

dvklopfenstein commented Sep 24, 2019 •

edited

Loading

kerrickstaley Oct 10, 2019

dvklopfenstein Oct 11, 2019

dvklopfenstein commented Nov 14, 2019 •

edited

Loading

kerrickstaley commented Dec 5, 2019

Add mtherieau's Cloze code to genanki #37

Add mtherieau's Cloze code to genanki #37

Conversation

dvklopfenstein commented Aug 31, 2019

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dvklopfenstein commented Sep 24, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dvklopfenstein commented Nov 14, 2019 • edited Loading

kerrickstaley commented Dec 5, 2019

dvklopfenstein commented Sep 24, 2019 •

edited

Loading

dvklopfenstein commented Nov 14, 2019 •

edited

Loading