Updated type checker version

Pipfile

@@ -4,50 +4,49 @@ url = "https://pypi.org/simple"
 verify_ssl = true
 
 [dev-packages]
-black = ">=19.10b0"
-pytest = "==7.4.4"
-pytest-django = "==4.4.0"
+black = "*"
+pytest = "*"
+pytest-django = "*"
 pytest-cov = "*"
 pytest-datadir = "*"
 pytest-mypy = "*"
 hypothesis = {version = "*", extras = ["django"]}
-codecov = "*"
 pysnooper = "*"
 python-levenshtein = "*"
-django-debug-toolbar = "~=3.6"
+django-debug-toolbar = "*"
 isort = "*"
 toml = "*"
-mypy = "*"
+mypy = "~=1.7"
+django-stubs = "~=4.2"
 pytest-env = "*"
 jupyterlab = "*"
 appnope = "*"
-nb_black = "==1.0.5"
 pandas = "*"
 statsmodels = "*"
 pandas-stubs = "*"
 sphinx = "*"
 myst-parser = "*"
 
 [packages]
-colorama = "~=0.4"
+colorama = "*"
 coverage = "*"
 cree-sro-syllabics = ">=2020.6.23"
 Django = "~=4.2"
-typing-extensions = "~=4.0"
-attrs = "~=23.2"
-django-js-reverse = "~=0.9"
+typing-extensions = "*"
+attrs = "*"
+django-js-reverse = "*"
 secure = "*"
 snowballstemmer = "*"
-dawg2 = "~=0.12"
+dawg2 = "*"
 environs = {extras = ["django"], version = "*"}
 hfst-optimized-lookup = "*"
 tqdm = "*"
 whitenoise = "*"
 foma = {subdirectory = "foma/python", git = "https://github.com/andrewdotn/foma"}
 uwsgi = "*"
 gensim = "*"
-scipy = "==1.10.1"
-more-itertools = "~=8.7.0"
+scipy = "*"
+more-itertools = "*"
 types-requests = "*"
 requests = "*"
 
@@ -62,7 +61,3 @@ ensuretestdb = "env USE_TEST_DB=true ./crkeng-manage ensuretestdb"
 
 [requires]
 python_version = "3.10"
-
-[pipenv]
-# this allows "black>=19.10b0" to work
-allow_prereleases = true

mypy.ini

@@ -1,4 +1,9 @@
 [mypy]
+plugins = mypy_django_plugin.main
+mypy_path = ./src
+
+[mypy.plugins.django-stubs]
+django_settings_module = crkeng.site.settings
 
 [mypy-CreeDictionary.API.migrations.*]
 # Skip checking Django-auto-generated migration code
@@ -43,7 +48,7 @@ ignore_missing_imports = True
 # “It is the view of the Django Technical Board that inline type
 # annotations should not be added to Django at the current time … the
 # barrier for further inline changes will be high.”
-ignore_missing_imports = True
+#ignore_missing_imports = True
 
 [mypy-django_js_reverse.*]
 ignore_missing_imports = True

src/CreeDictionary/API/search/presentation.py

@@ -205,7 +205,11 @@ def serialize(self) -> SerializedPresentationResult:
             "definitions": serialize_definitions(
                 # given changes in django 4.1, the previous approach can fail.
                 # check first that the wordform actually exists before trying to get the reverse manager.
-                self.wordform.definitions.all() if not self.wordform._state.adding else [],
+                (
+                    self.wordform.definitions.all()
+                    if not self.wordform._state.adding
+                    else []
+                ),
                 # This is the only place include_auto_definitions is used,
                 # because we only auto-translate non-lemmas, and this is the
                 # only place where a non-lemma search result appears.
@@ -359,7 +363,7 @@ def serialize_wordform(
         if key not in result:
             result[key] = wordform.linguist_info[key]
 
-    return result
+    return cast(SerializedWordform, result)
 
 
 def serialize_definitions(
@@ -452,7 +456,7 @@ def emoji_for_value(choice: str) -> str:
 
 
 def get_lexical_info(
-    result_analysis: RichAnalysis,
+    result_analysis: Optional[RichAnalysis],
     animate_emoji: str,
     show_emoji: str,
     dict_source: list,

src/CreeDictionary/CreeDictionary/test_views.py

@@ -1,7 +1,7 @@
 import logging
 import re
 from http import HTTPStatus
-from typing import Dict, Optional
+from typing import Dict, Optional, cast
 
 import pytest
 from django.http import (
@@ -147,10 +147,13 @@ def test_paradigm_from_full_page_and_api(client: Client):
     # Get fragment from API request:
     response = client.get(
         reverse("cree-dictionary-paradigm-detail"),
-        {
-            "lemma-id": wordform.id,
-            "paradigm-size": paradigm_size,
-        },
+        cast(
+            Dict[str, int | str],
+            {
+                "lemma-id": wordform.id,
+                "paradigm-size": paradigm_size,
+            },
+        ),
     )
     assert response.status_code == HTTPStatus.OK
     html_fragment = response.content.decode("UTF-8")

src/CreeDictionary/CreeDictionary/views.py

@@ -48,13 +48,13 @@ def entry_details(request, slug: str):
 
     :param slug: the stable unique ID of the lemma
     """
-    lemma = Wordform.objects.filter(slug=slug, is_lemma=True)
+    lemmas = Wordform.objects.filter(slug=slug, is_lemma=True)
 
-    if lemma.count() != 1:
+    if lemmas.count() != 1:
         # The result is either empty or ambiguous; either way, do a search!
         return redirect(url_for_query(slug.split("@")[0] or ""))
 
-    lemma = lemma.get()
+    lemma = lemmas.get()
 
     if rich_analysis := lemma.analysis:
         morphemes = rich_analysis.generate_with_morphemes(lemma.text)
@@ -81,8 +81,7 @@ def entry_details(request, slug: str):
             if size not in sizes:
                 size = default_size
 
-        paradigm = paradigm_for(lemma, size)
-        paradigm = get_recordings_from_paradigm(paradigm, request)
+        paradigm = get_recordings_from_paradigm(paradigm_for(lemma, size), request)
 
         paradigm_context.update(
             paradigm=paradigm, paradigm_size=size, paradigm_sizes=sizes
@@ -306,7 +305,7 @@ def query_help(request):  # pragma: no cover
     )
 
 
-@staff_member_required()
+@staff_member_required
 def fst_tool(request):
     context = {}
 
@@ -405,10 +404,10 @@ def paradigm_for(wordform: Wordform, paradigm_size: str) -> Optional[Paradigm]:
     manager = default_paradigm_manager()
 
     if name := wordform.paradigm:
-        fst_lemma = wordform.lemma.text
+        fst_lemma = wordform.lemma.text if wordform.lemma else None
 
         if settings.MORPHODICT_ENABLE_FST_LEMMA_SUPPORT:
-            fst_lemma = wordform.lemma.fst_lemma
+            fst_lemma = wordform.lemma.fst_lemma if wordform.lemma else None
 
         if paradigm := manager.paradigm_for(name, fst_lemma, paradigm_size):
             return paradigm

src/CreeDictionary/cvd/definition_keys.py

@@ -33,7 +33,7 @@ def definition_to_cvd_key(d: Definition) -> CvdKey:
         json.dumps(
             [
                 # Every lemma has a stable unique identifier
-                d.wordform.lemma.slug,
+                getattr(d.wordform.lemma, "slug"),
                 # These next two should fields should identify non-lemma wordforms
                 d.wordform.text,
                 d.wordform.raw_analysis,
@@ -71,5 +71,5 @@ def wordform_query_matches(query: WordformQuery, wordform: Wordform):
             ("raw_analysis" in query and wordform.raw_analysis == query["raw_analysis"])
             or ("raw_analysis__isnull" in query and wordform.raw_analysis is None)
         )
-        and wordform.lemma.slug == query["lemma__slug"]
+        and getattr(wordform.lemma, "slug") == query["lemma__slug"]
     )

src/CreeDictionary/phrase_translate/translate.py

@@ -127,9 +127,13 @@ def translate_and_print_wordforms(wordforms: Iterable[Wordform]):
         print(f"wordform: {wordform.text} {wordform.analysis}")
 
         lemma = wordform.lemma
+
+        if lemma is None:
+            continue
+
         print(f"  lemma: {lemma.analysis}")
 
-        for d in wordform.lemma.definitions.all():
+        for d in lemma.definitions.all():
             # Don’t try to re-translate already-translated items
             if d.auto_translation_source_id is not None:
                 continue

src/CreeDictionary/search_quality/management/commands/featuredump.py

@@ -59,7 +59,9 @@ def handle(self, *args, **options) -> None:
                     ret = r.features()
                     ret["query"] = query
                     ret["wordform_text"] = r.wordform.text
-                    ret["lemma_wordform_text"] = r.wordform.lemma.text
+                    ret["lemma_wordform_text"] = (
+                        r.wordform.lemma.text if r.wordform.lemma else None
+                    )
                     ret["definitions"] = [
                         [d.text, ", ".join(c.abbrv for c in d.citations.all())]
                         for d in r.wordform.definitions.all()

src/crkeng/app/preferences.py

@@ -20,7 +20,7 @@ class DisplayMode(Preference):
         # (Short) linguistic labels; e.g., 1Sg → 2Sg, Present Tense
         "linguistic": "Linguistic labels",
         # nêhiyawêwin labels; e.g., niya → kiya, mêkwâc
-        "source_language": settings.MORPHODICT_LANGUAGE_ENDONYM + " labels",
+        "source_language": str(settings.MORPHODICT_LANGUAGE_ENDONYM) + " labels",
     }
     default = "english"
 

src/crkeng/site/settings.py

@@ -61,15 +61,15 @@
 #
 # Morphodict assumes that the `text` of all Wordform are written in the default
 # orthography.
-MORPHODICT_ORTHOGRAPHY = {
+MORPHODICT_ORTHOGRAPHY: Orthographies = {
     # All entries in Wordform should be written in SRO (êîôâ)
     "default": "Latn-y",
     "available": {
         # 'Latn' is Okimāsis/Wolvegrey's SRO
         "Latn-y": {
             "name": "SRO (êîôâ) with y",
-            "converter": "CreeDictionary.CreeDictionary.orthography.to_y"
-            },
+            "converter": "CreeDictionary.CreeDictionary.orthography.to_y",
+        },
         "Latn": {"name": "SRO (êîôâ) with ý"},
         "Latn-x-macron-y": {
             "name": "SRO (ēīōā) with y",

src/morphodict/lexicon/importer/test_importer.py

@@ -3,7 +3,8 @@
 import pytest
 from django.core.management import call_command
 from django.db import IntegrityError
-from pytest_django.fixtures import _django_db_fixture_helper
+from pytest_django.fixtures import _django_db_helper
+from pytest_django.plugin import DjangoDbBlocker
 
 from morphodict.lexicon.models import (
     Wordform,
@@ -89,7 +90,7 @@ def parametrize_incremental(f):
 
 
 @pytest.fixture(scope="class")
-def class_scoped_db(request, django_db_blocker):
+def class_scoped_db(request: pytest.FixtureRequest, django_db_blocker: DjangoDbBlocker):
     """A class-scoped DB fixture
 
     The normal pytest-django `db` fixture is function-scoped, meaning that it
@@ -99,7 +100,9 @@ def class_scoped_db(request, django_db_blocker):
     with multiple tests in which the same DB transaction is used for all tests
     in the class.
     """
-    _django_db_fixture_helper(request=request, django_db_blocker=django_db_blocker)
+    _django_db_helper(
+        request=request, django_db_setup=None, django_db_blocker=django_db_blocker
+    )
 
 
 # Tests

src/morphodict/lexicon/management/commands/importjsondict.py

@@ -369,7 +369,7 @@ def run(self):
             # Avoid dupes for this wordform
             seen_source_language_keywords: set[str] = set()
 
-            slug_base = wf.slug.split("@")[0]
+            slug_base = wf.slug.split("@")[0] if wf.slug else None
             if wf.text != slug_base and slug_base:
                 self.add_source_language_keyword(
                     wf, slug_base, seen_source_language_keywords

src/morphodict/lexicon/models.py

@@ -33,7 +33,7 @@ def get_queryset(self):
 
 
 # This type is the int pk for a saved wordform, or (text, analysis) for an unsaved one.
-WordformKey = Union[int, tuple[str, str]]
+WordformKey = Union[str, int, tuple[str, str]]
 
 
 class DiacriticPreservingJsonEncoder(DjangoJSONEncoder):

src/morphodict/preference/views.py

@@ -35,6 +35,10 @@ def change_preference(request: HttpRequest, name: str):
     if value not in preference.choices:
         return HttpResponse(status=HTTPStatus.BAD_REQUEST)
 
+    # Make the typechecker see that value is not None after the previous check:
+    if value is None:
+        return HttpResponse(status=HTTPStatus.BAD_REQUEST)
+
     if who_asked_us := request.headers.get("Referer"):
         # Force the browser to refresh the page that issued this request.
         response = HttpResponse(status=HTTPStatus.SEE_OTHER)

src/morphodict/site/checks.py

@@ -1,7 +1,10 @@
 from django.conf import settings
 from django.core.checks import Error, register
+from typing import Optional
 
-_MORPHODICT_REQUIRED_SETTING_SENTINEL = object()
+_MORPHODICT_REQUIRED_SETTING_SENTINEL = None
+
+RequiredString = Optional[str]
 
 
 @register()

src/morphodict/site/settings.py

@@ -14,12 +14,12 @@
 
 import os
 import secrets
-from typing import Optional
+from typing import Optional, TypedDict, Dict
 
 from environs import Env
 
 from . import base_dir_setup
-from .checks import _MORPHODICT_REQUIRED_SETTING_SENTINEL
+from .checks import _MORPHODICT_REQUIRED_SETTING_SENTINEL, RequiredString
 from .hostutils import HOSTNAME
 from .save_secret_key import save_secret_key
 
@@ -267,7 +267,6 @@ def defaultDatabasePath():
 
 USE_I18N = True
 
-USE_L10N = True
 
 USE_TZ = True
 
@@ -385,18 +384,24 @@ def defaultDatabasePath():
 
 # The name of the source language, written in the target language. Used in
 # default templates to describe what language the dictionary is for.
-MORPHODICT_SOURCE_LANGUAGE_NAME = _MORPHODICT_REQUIRED_SETTING_SENTINEL
+MORPHODICT_SOURCE_LANGUAGE_NAME: RequiredString = _MORPHODICT_REQUIRED_SETTING_SENTINEL
 
 # An optional, shorter name for the language. Currently only used in the search
 # bar placeholder, to show “Search in Cree” instead of “Search in Plains Cree”
 MORPHODICT_SOURCE_LANGUAGE_SHORT_NAME: Optional[str] = None
 
 # The name of the language, in the language itself, e.g., ‘nêhiyawêwin’
-MORPHODICT_LANGUAGE_ENDONYM = _MORPHODICT_REQUIRED_SETTING_SENTINEL
+MORPHODICT_LANGUAGE_ENDONYM: RequiredString = _MORPHODICT_REQUIRED_SETTING_SENTINEL
 
 
 # The marketing / brand / public-facing name of the dictionary
-MORPHODICT_DICTIONARY_NAME = _MORPHODICT_REQUIRED_SETTING_SENTINEL
+MORPHODICT_DICTIONARY_NAME: RequiredString = _MORPHODICT_REQUIRED_SETTING_SENTINEL
 
 # Used for the bulk search API
 SPEECH_DB_EQ = ["_"]
+
+Orthography = TypedDict("Orthography", {"name": str, "converter": str}, total=False)
+
+Orthographies = TypedDict(
+    "Orthographies", {"default": str, "available": Dict[str, Orthography]}
+)