Skip to content

armgit / wger

Public Imported from GitHub
0

Self hosted FLOSS fitness/workout, nutrition and weight tracker

Simplify and consolidate serializer behaviour

Roland Geider73df8abbcc5b
diff --git a/wger/core/api/serializers.py b/wger/core/api/serializers.py
index 392e7cbaf..0ffa73377 100644
--- a/wger/core/api/serializers.py
+++ b/wger/core/api/serializers.py
@@ -272,7 +272,6 @@ class LanguageCheckSerializer(serializers.Serializer):
# Try to detect the language
detector = (
LanguageDetectorBuilder.from_all_languages()
- .with_low_accuracy_mode()
.with_preloaded_language_models()
.build()
)
diff --git a/wger/exercises/api/serializers.py b/wger/exercises/api/serializers.py
index 6c2fa04a0..b9b421f8e 100644
--- a/wger/exercises/api/serializers.py
+++ b/wger/exercises/api/serializers.py
@@ -24,14 +24,11 @@ from django.db.models import Q
# Third Party
from actstream import action as actstream_action
-from lingua import LanguageDetectorBuilder
from rest_framework import serializers
# wger
-from wger.core.models import (
- Language,
- License,
-)
+from wger.core.models import License
+from wger.exercises.api.validators import validate_language_matches
from wger.exercises.models import (
Alias,
DeletionLog,
@@ -225,56 +222,12 @@ class ExerciseCommentSerializer(serializers.ModelSerializer):
'comment',
)
-
-class ExerciseCommentSubmissionSerializer(serializers.ModelSerializer):
- """
- ExerciseComment submission serializer
- """
-
- class Meta:
- model = ExerciseComment
- fields = ('comment',)
-
- def create(self, validated_data, **kwargs):
- """
- Custom create-method to handle the 'translation' keyword argument
- and set the foreign key relationship.
- """
- translation: Translation = kwargs.get('translation')
- if not translation:
- raise serializers.ValidationError(
- 'A translation object is required for creating a comment.'
- )
-
- # Validate the language of the description
- # -> This is done here instead of in the serializer's validate method
- # because the language is not available in the serializer's initial_data
- detector = (
- LanguageDetectorBuilder.from_all_languages()
- .with_low_accuracy_mode()
- .with_preloaded_language_models()
- .build()
- )
- language = translation.language
-
- # Try to detect the language
- detected_language = detector.detect_language_of(validated_data['comment'])
- detected_language_code = detected_language.iso_code_639_1.name.lower()
- if detected_language_code != language.short_name.lower():
- raise serializers.ValidationError(
- {
- 'language': f'The detected language of the comment is "{detected_language.name.capitalize()}" '
- f'({detected_language_code}), which does not match your selected language: '
- f'"{language.full_name.capitalize()}" ({language.short_name}). If you believe '
- f'this is incorrect, try adding more content or rephrasing your text, as '
- f'language detection works better with longer or more complete sentences.'
- }
- )
-
- # Create the comment with the parent translation
- comment = ExerciseComment.objects.create(translation=translation, **validated_data)
- _log_action_creation(self, comment)
- return comment
+ def validate(self, data):
+ comment = data.get('comment')
+ translation = data.get('translation') or (self.instance and self.instance.translation)
+ if comment and translation:
+ validate_language_matches(comment, translation.language, 'comment')
+ return super().validate(data)
class ExerciseAliasSerializer(serializers.ModelSerializer):
@@ -292,32 +245,6 @@ class ExerciseAliasSerializer(serializers.ModelSerializer):
)
-class ExerciseAliasSubmissionSerializer(serializers.ModelSerializer):
- """
- ExerciseAlias submission serializer
- """
-
- class Meta:
- model = Alias
- fields = ('alias',)
-
- def create(self, validated_data, **kwargs):
- """
- Custom create-method to handle the 'translation' keyword argument
- and set the foreign key relationship.
- """
- translation = kwargs.get('translation')
- if not translation:
- raise serializers.ValidationError(
- 'A translation object is required for creating an alias.'
- )
-
- # Create the Alias with the parent translation
- alias = Alias.objects.create(translation=translation, **validated_data)
- _log_action_creation(self, alias)
- return alias
-
-
class ExerciseInfoAliasSerializer(serializers.ModelSerializer):
"""
Exercise alias serializer for info endpoint
@@ -422,87 +349,6 @@ class ExerciseTranslationInfoSerializer(serializers.ModelSerializer):
)
-class ExerciseTranslationSubmissionSerializer(serializers.ModelSerializer):
- """
- Exercise translation submission serializer
- """
-
- language = serializers.PrimaryKeyRelatedField(queryset=Language.objects.all())
- aliases = ExerciseAliasSubmissionSerializer(many=True, required=False)
- comments = ExerciseCommentSubmissionSerializer(many=True, required=False)
-
- class Meta:
- model = Translation
- fields = (
- 'name',
- 'description_source',
- 'language',
- 'aliases',
- 'comments',
- 'license_author',
- )
-
- def validate(self, data):
- """
- Custom validator to ensure the detected language of the description corresponds with the
- provided language.
- """
- detector = (
- LanguageDetectorBuilder.from_all_languages()
- .with_low_accuracy_mode()
- .with_preloaded_language_models()
- .build()
- )
-
- language = data.get('language')
- description = data.get('description_source')
-
- # Try to detect the language
- detected_language = detector.detect_language_of(description)
- detected_language_code = detected_language.iso_code_639_1.name.lower()
- if detected_language_code != language.short_name.lower():
- raise serializers.ValidationError(
- {
- 'language': f'The detected language of the description is "{detected_language.name.capitalize()}" '
- f'({detected_language_code}), which does not match your selected language: '
- f'"{language.full_name.capitalize()}" ({language.short_name}). If you believe '
- f'this is incorrect, try adding more content or rephrasing your text, as '
- f'language detection works better with longer or more complete sentences.'
- }
- )
-
- return super().validate(data)
-
- def create(self, validated_data, **kwargs):
- """
- Custom create-method to handle the 'exercise' keyword argument
- and set the foreign key relationship.
- """
- aliases_data = validated_data.pop('aliases', [])
- comments_data = validated_data.pop('comments', [])
- exercise = kwargs.get('exercise')
- if not exercise:
- raise serializers.ValidationError(
- 'Translation object is required for creating an Alias'
- )
-
- # Create the translation with the parent exercise
- translation = Translation.objects.create(exercise=exercise, **validated_data)
- _log_action_creation(self, translation)
-
- # Create the individual aliases
- for alias_data in aliases_data:
- alias_serializer = self.fields['aliases'].child
- alias_serializer.create(validated_data=alias_data, translation=translation)
-
- # Create the individual comments
- for comment_data in comments_data:
- comment_serializer = self.fields['comments'].child
- comment_serializer.create(validated_data=comment_data, translation=translation)
-
- return translation
-
-
class ExerciseTranslationSerializer(serializers.ModelSerializer):
"""
Exercise translation serializer
@@ -533,8 +379,10 @@ class ExerciseTranslationSerializer(serializers.ModelSerializer):
def validate(self, value):
"""
- Check that there is only one language per exercise
+ Check that there is only one language per exercise, and that the
+ detected language of the description matches the declared language.
"""
+
if value.get('language'):
# Editing an existing object
# -> Check if the language already exists, excluding the current object
@@ -546,9 +394,13 @@ class ExerciseTranslationSerializer(serializers.ModelSerializer):
raise serializers.ValidationError(
f'There is already a translation for this exercise in {value["language"]}'
)
+
# Creating a new object
- # -> Check if the language already exists
- else:
+ # -> Check if the language already exists. When ``exercise`` isn't
+ # in the payload (submission flow creates a fresh exercise), the
+ # check can't run and isn't relevant and there are no duplicates
+ # by definition.
+ elif value.get('exercise'):
if Translation.objects.filter(
exercise=value['exercise'],
language=value['language'],
@@ -557,9 +409,84 @@ class ExerciseTranslationSerializer(serializers.ModelSerializer):
f'There is already a translation for this exercise in {value["language"]}'
)
+ # Verify the detected language of the description matches the declared
+ # language. On PATCH either field may be absent, so we fall back to the
+ # persisted value where available.
+ description = value.get('description_source')
+ language = value.get('language') or (self.instance and self.instance.language)
+ if description and language:
+ validate_language_matches(description, language, 'description')
+
return super().validate(value)
+class ExerciseTranslationSubmissionSerializer(ExerciseTranslationSerializer):
+ """
+ Translation serializer used as a nested child of ``ExerciseSubmissionSerializer``.
+
+ Differs from the regular serializer only because:
+ - the ``exercise`` FK isn't known until the parent creates it (passed via ``create()`` kwargs);
+ - the payload also accepts nested ``aliases`` and ``comments`` lists,
+ which the regular CRUD endpoint doesn't.
+ """
+
+ aliases = ExerciseAliasSerializer(many=True, required=False)
+ comments = ExerciseCommentSerializer(many=True, required=False)
+
+ class Meta(ExerciseTranslationSerializer.Meta):
+ fields = (
+ 'name',
+ 'description_source',
+ 'language',
+ 'aliases',
+ 'comments',
+ 'license_author',
+ )
+
+ def __init__(self, *args, **kwargs):
+ super().__init__(*args, **kwargs)
+
+ # ``translation`` is assigned by the parent submission in create()
+ for nested in ('aliases', 'comments'):
+ self.fields[nested].child.fields.pop('translation', None)
+
+ def validate(self, data):
+ data = super().validate(data)
+
+ # Language of each child comment must also match the one in the declared translation
+ language = data.get('language')
+ if language:
+ for comment_data in data.get('comments', []):
+ comment_text = comment_data.get('comment')
+ if comment_text:
+ validate_language_matches(comment_text, language, 'comment')
+
+ return data
+
+ def create(self, validated_data, **kwargs):
+ aliases_data = validated_data.pop('aliases', [])
+ comments_data = validated_data.pop('comments', [])
+ exercise = kwargs.get('exercise')
+ if not exercise:
+ raise serializers.ValidationError(
+ 'Exercise is required to create a translation via submission.'
+ )
+ validated_data['exercise'] = exercise
+
+ translation = Translation.objects.create(**validated_data)
+ _log_action_creation(self, translation)
+
+ for alias_data in aliases_data:
+ alias = Alias.objects.create(translation=translation, **alias_data)
+ _log_action_creation(self, alias)
+
+ for comment_data in comments_data:
+ comment = ExerciseComment.objects.create(translation=translation, **comment_data)
+ _log_action_creation(self, comment)
+
+ return translation
+
+
class ExerciseInfoSerializer(serializers.ModelSerializer):
"""
Exercise info serializer
diff --git a/wger/exercises/api/validators.py b/wger/exercises/api/validators.py
new file mode 100644
index 000000000..9a97f9c2c
--- /dev/null
+++ b/wger/exercises/api/validators.py
@@ -0,0 +1,51 @@
+# This file is part of wger Workout Manager.
+#
+# wger Workout Manager is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# wger Workout Manager is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with Workout Manager. If not, see <http://www.gnu.org/licenses/>.
+
+# Third Party
+from lingua import LanguageDetectorBuilder
+from rest_framework import serializers
+
+# wger
+from wger.core.models import Language
+
+
+def validate_language_matches(text: str, language: Language, source_label: str) -> None:
+ """
+ Raise ``ValidationError`` if the detected language of ``text`` doesn't
+ match the ``short_name`` of the given ``Language``.
+
+ ``source_label`` (e.g. ``'description'``, ``'comment'``) is used in the
+ error message to point the caller at the offending field.
+ """
+
+ detector = (
+ LanguageDetectorBuilder.from_all_languages()
+ .with_preloaded_language_models()
+ .build()
+ )
+
+ detected_language = detector.detect_language_of(text)
+ detected_language_code = detected_language.iso_code_639_1.name.lower()
+ if detected_language_code != language.short_name.lower():
+ raise serializers.ValidationError(
+ {
+ 'language': f'The detected language of the {source_label} is '
+ f'"{detected_language.name.capitalize()}" ({detected_language_code}), '
+ f'which does not match your selected language: "{language.full_name.capitalize()}" '
+ f'({language.short_name}). If you believe this is incorrect, try adding '
+ f'more content or rephrasing your text, as language detection works'
+ f'better with longer or more complete sentences.'
+ }
+ )
diff --git a/wger/exercises/tests/api/test_exercise_submission_api.py b/wger/exercises/tests/api/test_exercise_submission_api.py
index 3e6339863..0fe1ff8ac 100644
--- a/wger/exercises/tests/api/test_exercise_submission_api.py
+++ b/wger/exercises/tests/api/test_exercise_submission_api.py
@@ -154,6 +154,49 @@ class SearchSubmissionApiTestCase(BaseTestCase, ApiBaseTestCase):
response_data.get('translations'), ['You must provide at least one translation.']
)
+ def test_unsuccessful_submission_language_mismatch_description(self):
+ """
+ A translation whose description language doesn't match the declared language
+ field is rejected.
+ """
+ self.authenticate('admin')
+
+ payload = self.get_payload()
+ # Swap the EN description for a clearly German one.
+ payload['translations'][0]['description_source'] = (
+ 'Das ist eine deutsche Beschreibung der Übung, mit ausreichend Text '
+ 'damit die Spracherkennung sie zuverlässig erkennen kann.'
+ )
+
+ counts_before = self.get_counts()
+ response = self.client.post(self.url, data=payload)
+ counts_after = self.get_counts()
+
+ self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+ self.assertEqual(counts_before, counts_after)
+ self.assertIn('language', response.json().get('translations')[0])
+
+ def test_unsuccessful_submission_language_mismatch_comment(self):
+ """
+ A comment whose detected language doesn't match the parent translation's
+ language is rejected.
+ """
+ self.authenticate('admin')
+
+ payload = self.get_payload()
+ # Replace one of the English comments with a clearly French one.
+ payload['translations'][0]['comments'][0]['comment'] = (
+ 'Ceci est un long commentaire en français qui ne correspond pas du '
+ 'tout à la traduction anglaise et devrait être rejeté.'
+ )
+
+ counts_before = self.get_counts()
+ response = self.client.post(self.url, data=payload)
+ counts_after = self.get_counts()
+
+ self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+ self.assertEqual(counts_before, counts_after)
+
def test_unsuccessful_submission_no_english_translations(self):
"""
If any part of the exercise submission fails, no exercise is created.
diff --git a/wger/exercises/tests/test_exercise_comments.py b/wger/exercises/tests/test_exercise_comments.py
index 66fdfabc4..61c5846b2 100644
--- a/wger/exercises/tests/test_exercise_comments.py
+++ b/wger/exercises/tests/test_exercise_comments.py
@@ -12,6 +12,9 @@
#
# You should have received a copy of the GNU Affero General Public License
+# Third Party
+from rest_framework import status
+
# wger
from wger.core.tests.api_base_test import ExerciseCrudApiTestCase
from wger.core.tests.base_testcase import WgerTestCase
@@ -39,7 +42,29 @@ class ExerciseCommentApiTestCase(ActstreamApiMixin, ExerciseCrudApiTestCase):
pk = 1
resource = ExerciseComment
data = {
- 'comment': 'a cool comment',
+ 'comment': 'This is a clearly English comment used for testing purposes.',
'translation': '1',
'id': 1,
}
+
+ def test_post_rejects_language_mismatch(self):
+ """
+ Adding a comment whose language doesn't match the parent translation's
+ language is rejected.
+ """
+
+ # Translation pk=1 has language=2 (en); posting a German comment must
+ # be rejected.
+ self.authenticate('trainer1')
+ response = self.client.post(
+ self.url,
+ data={
+ 'translation': 1,
+ 'comment': (
+ 'Ein ausreichend langer deutscher Kommentar, damit die '
+ 'Spracherkennung greifen kann und ihn ablehnt.'
+ ),
+ },
+ )
+ self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+ self.assertIn('language', response.json())
diff --git a/wger/exercises/tests/test_exercise_translation.py b/wger/exercises/tests/test_exercise_translation.py
index bc92b5695..d42b962e2 100644
--- a/wger/exercises/tests/test_exercise_translation.py
+++ b/wger/exercises/tests/test_exercise_translation.py
@@ -195,7 +195,10 @@ class ExerciseTranslationCustomApiTestCase(ActstreamApiMixin, ExerciseCrudApiTes
"""
payload = {
'name': 'A new translation',
- 'description_source': 'Bend your knees and squat down.',
+ 'description_source': (
+ 'Beuge die Knie und gehe in die tiefe Hocke, halte dabei den '
+ 'Rücken gerade und die Brust aufrecht.'
+ ),
'language': 1,
'exercise': 1,
}
@@ -206,9 +209,9 @@ class ExerciseTranslationCustomApiTestCase(ActstreamApiMixin, ExerciseCrudApiTes
self.assertEqual(response.status_code, status.HTTP_201_CREATED)
translation = Translation.objects.get(pk=response.json()['id'])
- self.assertEqual(translation.description_source, 'Bend your knees and squat down.')
+ self.assertEqual(translation.description_source, payload['description_source'])
# Backend rendered the markdown into description on save()
- self.assertIn('Bend your knees', translation.description)
+ self.assertIn('Beuge die Knie', translation.description)
def test_cant_set_description_directly(self):
"""
@@ -218,7 +221,10 @@ class ExerciseTranslationCustomApiTestCase(ActstreamApiMixin, ExerciseCrudApiTes
payload = {
'name': 'A new translation',
'description': '<script>alert(1)</script><p>raw html</p>',
- 'description_source': 'safe markdown',
+ 'description_source': (
+ 'Sicherer Markdown-Text der nur in deutscher Sprache geschrieben '
+ 'ist um die Spracherkennung zuverlässig zu bestehen.'
+ ),
'language': 1,
'exercise': 1,
}
@@ -232,7 +238,7 @@ class ExerciseTranslationCustomApiTestCase(ActstreamApiMixin, ExerciseCrudApiTes
self.assertNotIn('<script>', translation.description)
self.assertNotIn('raw html', translation.description)
- self.assertIn('safe markdown', translation.description)
+ self.assertIn('Sicherer Markdown-Text', translation.description)
def test_cant_patch_description_directly(self):
"""
@@ -260,6 +266,71 @@ class ExerciseTranslationCustomApiTestCase(ActstreamApiMixin, ExerciseCrudApiTes
translation = Translation.objects.get(pk=self.pk)
self.assertEqual(translation.description, ' The wild boar is a suid native...')
+ def test_post_accepts_matching_language(self):
+ """
+ POSTing a translation whose description is in the declared language
+ succeeds (positive counterpart to ``test_post_rejects_language_mismatch``).
+ """
+ Translation.objects.filter(exercise_id=1, language_id=1).delete()
+
+ payload = {
+ 'name': 'Eine neue Übersetzung',
+ 'description_source': (
+ 'Halte die Hantel mit beiden Händen und führe die Bewegung '
+ 'kontrolliert aus, dabei den Rücken stets gerade.'
+ ),
+ 'language': 1,
+ 'exercise': 1,
+ }
+ self.authenticate('trainer1')
+ response = self.client.post(self.url, data=payload)
+ self.assertEqual(response.status_code, status.HTTP_201_CREATED)
+
+ def test_post_rejects_language_mismatch(self):
+ """
+ POSTing a translation whose description language doesn't match the
+ declared language field is rejected.
+ """
+
+ # Free up exercise=1/language=2 so the duplicate-translation check
+ # doesn't fire before the language-mismatch check.
+ Translation.objects.filter(exercise_id=1, language_id=2).delete()
+
+ payload = {
+ 'name': 'A new translation',
+ 'description_source': (
+ 'Das ist eine deutsche Beschreibung der Übung, mit ausreichend '
+ 'Text damit die Spracherkennung sie zuverlässig erkennen kann.'
+ ),
+ 'language': 2,
+ 'exercise': 1,
+ }
+ self.authenticate('trainer1')
+ response = self.client.post(self.url, data=payload)
+ self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+ self.assertIn('language', response.json())
+
+ def test_patch_rejects_language_mismatch(self):
+ """
+ PATCHing ``description_source`` to text in a different language than
+ the (existing) translation's language is rejected.
+ """
+
+ # Translation pk=1 has language=2 (en); push a clearly German
+ # description and expect the validator to reject it.
+ self.authenticate('trainer1')
+ response = self.client.patch(
+ self.url_detail,
+ data={
+ 'description_source': (
+ 'Eine ausreichend lange deutsche Beschreibung, damit die '
+ 'Spracherkennung sicher greifen kann.'
+ ),
+ },
+ )
+ self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+ self.assertIn('language', response.json())
+
def test_post_only_one_language_per_base(self):
"""
Test that it's not possible to add a second translation for the same