Tests

The testing framework is Pytest with these plugins:

• pytest-qt: allows for writing tests for PyQt5, PyQt6, PySide2 and PySide6 applications.
• pytest-xvfb: allows tests to be run without windows popping up during GUI tests or on systems without a display (like a CI).
• pytest-expect: stores test expectations by saving the set of failing tests, allowing them to be marked as xfail when running them in the future.
• pytest-sugar: showing failures and errors instantly, adding a progress bar, and making the output look better.
• pytest-randomly: randomly order tests and control random seed.

Some default options are set in the pyproject.toml file under the [tool.pytest.ini_options] section.

Running tests

From the project root folder you can run tests with the command pytest, and it will automatically gather and run all the tests found in the tests directory.

You can restrict the tests to a specific file by appending the path(s) to the command, e.g:

pytest test/tests/recalc_test.py test/tests/review_test.py

You can also add "tags" to specific tests and run all the tests that have those tags. This is done by adding the decorator @pytest.mark.{tag}, e.g. @pytest.mark.external_morphemizers, and you can then run those tests with the command:

pytest -m external_morphemizers

Note: the "tag" also has to be added to the "markers" list in pyproject.toml because we run tests with the --strict-markers argument.

You can specify how many fails the tests should abort after with the --maxfail argument, e.g:

pytest --maxfail=10

The pytest-randomly plugin runs the tests in a random order based on a seed. You can specify the seed:

pytest --randomly-seed=84903385

You can also rerun the last used seed:

pytest --randomly-seed=last

Test coverage

Run: pytest --cov=ankimorphs --cov-report html and click on index.html in htmlcov/

Card collections

Note:

• all of these collections have all the extra fields selected
• the extra fields contain inflections unless stated otherwise
• morphs are evaluated based on inflections unless stated otherwise
• morphs priority is 'Collection Frequency' unless stated otherwise

Current card collections (test/data/card_collections):

• lemma_evaluation_lemma_extra_fields_collection.anki2
  • contains two "known" cards ("the", "man"), and then 9 cards with 4 lemmas and 9 inflections. this is used for testing if the inflections are given the scores of their respective lemmas and the inflections are skipped on review.
  • the extra fields contain lemmas
  • morphs are evaluated based on lemma
  • morphemizer: 'spaCy: en_core_web_sm'
• some_studied_lemmas_collection.anki2
  • duplicate of lemma_evaluation_lemma_extra_fields_collection.anki2, but one card for each lemma has been studied, so the other cards that have those lemmas should contain no unknowns.
  • the extra fields contain lemmas
  • morphs are evaluated based on lemma
  • morphemizer: 'spaCy: en_core_web_sm'
• big_japanese_collection.anki2
  • monolithic card collection (https://github.com/mortii/anki-decks)
  • morphemizer: 'Ankimorphs Japanese'
• ignore_names_txt_collection.anki2
  • Contains one card that has two names, one of which is found in names.txt
  • morphemizer: 'AnkiMorphs: Language w/ Spaces'
• known-morphs-collection.anki2
  • contains one card with 7 morphs, 6 of which are found in the 'known-morphs-valid' directory
  • morphemizer: 'AnkiMorphs: Language w/ Spaces'
• offset_new_cards_inflection_collection.anki2
  • Contains two cards, both with "hello".
  • recalc_offset_new_cards config option enabled
  • morphemizer: 'AnkiMorphs: Language w/ Spaces'
• offset_new_cards_lemma_collection.anki2
  • duplicate of lemma_evaluation_lemma_extra_fields_collection.anki2, but has uses the recalc_offset_new_cards config option enabled
  • the extra fields contain lemmas
  • morphs are evaluated based on lemma
  • morphemizer: 'spaCy: en_core_web_sm'
• some_studied_japanese_collection.anki2
  • contains three cards:
    • one that has the tag 'am-known-manually'
    • one with a learning interval of 1 day
    • one with a learning interval of 30+ days
  • morphemizer: 'Ankimorphs Japanese'
• max_morph_priority_collection.anki2
  • contains two cards, and uses the priority file ja_core_news_sm_freq_inflection_min_occurrence.csv. One card has a morph that is found in the priority file, the other card does not, which should give it a max morph priority value.
  • morphemizer: 'spaCy: ja_core_news_sm'
• card_handling_collection.anki2
  • contains 6 cards:
    • 3 known, not new cards
    • 1 fresh, not new card
    • 1 with only known or fresh morphs, new card
    • 1 with known, fresh, and one unknown morph, new card
  • creating an evergreen collection that has fresh morphs is surprisingly difficult, because the due date of the card actually depends on the timestamp of card creation. To make this problem go away you can set the date of your system back a couple of months and then create the cards.
• suspend_all_morphs_known.anki2
  • the base of this collection is card_handling_collection.anki2, that has been recalced with the config config_suspend_morphs_known
• suspend_morphs_known_or_fresh.anki2
  • the base of this collection is card_handling_collection.anki2, that has been recalced with the config config_suspend_morphs_known_or_fresh
• move_to_end_morphs_known.anki2
  • the base of this collection is card_handling_collection.anki2, that has been recalced with the config config_move_to_end_morphs_known
• move_to_end_morphs_known_or_fresh.anki2
  • the base of this collection is card_handling_collection.anki2, that has been recalced with the config config_move_to_end_morphs_known_or_fresh
• card_stability_collection.anki2
  • contains ~100 cards from a real and mature collection, same filters as big_japanese_collection,
• card_interval_collection.ank2
  • same as card_stability_collection.anki2 but recalced without USE_STABILITY_FOR_KNOWN_THRESHOLD

Engineering and adding collections

1. create a new anki profile
2. (optional) create a new deck
3. (optional) create a new note type
4. add cards and change the ankimorphs settings to fit the use case
5. recalc
6. exit anki/close the profile
7. extract the collection.anki2 file from the profile directory and rename it to something relevant to the use case and place the renamed collection file in test/data/card_collections
8. repeat this process for the ankimorphs.db if the tests require it, and move it to test/data/am_dbs
9. create a new fake config in test/fake_configs.py and make any necessary adjustments to have it match the settings used when you previously ran recalc.

Investigating collections

Basically reverse the steps in the section above, i.e:

1. extract the collection file you are interested in
2. rename the collection to collection.anki2
3. place the file in the profile folder of a new anki profile
4. open the anki profile and the cards should be available for viewing

Priority files

The priority files found in correct_outputs are generated with the following settings:

• all the file formats selected
• none of the preprocess options selected
• occurrences column is always added
• when comprehension is selected the target is always 90%
• when minimum occurrence is used the target is always 1