chore: refactor antispam API

Allow future extension to services supporting checking multiple strings
at once.

weblate/trans/forms.py

@@ -9,7 +9,7 @@
 import re
 from datetime import datetime
 from secrets import token_hex
-from typing import TYPE_CHECKING, Literal
+from typing import TYPE_CHECKING, ClassVar, Literal
 
 from crispy_forms.bootstrap import InlineCheckboxes, InlineRadios, Tab, TabHolder
 from crispy_forms.helper import FormHelper
@@ -1430,28 +1430,25 @@ def get_field_doc(self, field: forms.Field) -> tuple[str, str] | None:
 
 
 class SpamCheckMixin(forms.Form):
-    def spam_check(self, value) -> None:
-        if is_spam(value, self.request):
-            raise ValidationError(gettext("This field has been identified as spam!"))
+    spam_fields: ClassVar[tuple[str, ...]]
+
+    def clean(self) -> None:
+        data = self.cleaned_data
+        check_values: list[str] = [
+            data[field] for field in self.spam_fields if field in data
+        ]
+        if is_spam(self.request, check_values):
+            raise ValidationError(
+                gettext("This submission has been identified as spam!")
+            )
 
 
 class ComponentAntispamMixin(SpamCheckMixin):
-    def clean_agreement(self):
-        value = self.cleaned_data["agreement"]
-        self.spam_check(value)
-        return value
+    spam_fields = ("agreement",)
 
 
 class ProjectAntispamMixin(SpamCheckMixin):
-    def clean_web(self):
-        value = self.cleaned_data["web"]
-        self.spam_check(value)
-        return value
-
-    def clean_instructions(self):
-        value = self.cleaned_data["instructions"]
-        self.spam_check(value)
-        return value
+    spam_fields = ("web", "instructions")
 
 
 class ComponentSettingsForm(

weblate/trans/views/edit.py

@@ -300,7 +300,7 @@ def perform_suggestion(unit, form, request: AuthenticatedHttpRequest):
         return False
     # Spam check for unauthenticated users
     if not request.user.is_authenticated and is_spam(
-        "\n".join(form.cleaned_data["target"]), request
+        request, form.cleaned_data["target"]
     ):
         messages.error(request, gettext("Your suggestion has been identified as spam!"))
         return False

weblate/utils/antispam.py

@@ -29,30 +29,34 @@ def get_akismet():
     )
 
 
-def is_spam(text, request: AuthenticatedHttpRequest):
+def is_spam(request: AuthenticatedHttpRequest, texts: str | list[str]):
     """Check whether text is considered spam."""
-    if not text:
+    if not texts or not any(texts):
         return False
+    if isinstance(texts, str):
+        texts = [texts]
     akismet = get_akismet()
     if akismet is not None:
         from akismet import AkismetServerError, SpamStatus
 
         user_ip = get_ip_address(request)
         user_agent = get_user_agent_raw(request)
 
-        try:
-            result = akismet.check(
-                user_ip=user_ip,
-                user_agent=user_agent,
-                comment_content=text,
-                comment_type="comment",
-            )
-        except (OSError, AkismetServerError):
-            report_error("Akismet error")
-            return True
-        if result:
-            report_error("Akismet reported spam", level="info", message=True)
-        return result == SpamStatus.DefiniteSpam
+        for text in texts:
+            try:
+                result = akismet.check(
+                    user_ip=user_ip,
+                    user_agent=user_agent,
+                    comment_content=text,
+                    comment_type="comment",
+                )
+            except (OSError, AkismetServerError):
+                report_error("Akismet error")
+                return True
+            if result:
+                report_error("Akismet reported spam", level="info", message=True)
+            if result == SpamStatus.DefiniteSpam:
+                return True
     return False
 
 

weblate/utils/tests/test_antispam.py

@@ -21,7 +21,7 @@
 class SpamTest(TestCase):
     @override_settings(AKISMET_API_KEY=None)
     def test_disabled(self) -> None:
-        self.assertFalse(is_spam("text", HttpRequest()))
+        self.assertFalse(is_spam(HttpRequest(), "text"))
 
     def mock_akismet(self, body, **kwargs) -> None:
         responses.add(
@@ -40,26 +40,34 @@ def mock_akismet(self, body, **kwargs) -> None:
             responses.POST, "https://rest.akismet.com/1.1/verify-key", body="valid"
         )
 
+    @responses.activate
+    def test_akismet_spam_blank(self) -> None:
+        self.assertFalse(is_spam(HttpRequest(), ""))
+        self.assertFalse(is_spam(HttpRequest(), [""]))
+
     @skipIf(not HAS_AKISMET, "akismet module not installed")
     @responses.activate
     @override_settings(AKISMET_API_KEY="key")
     def test_akismet_spam(self) -> None:
         self.mock_akismet("true")
-        self.assertFalse(is_spam("text", HttpRequest()))
+        self.assertFalse(is_spam(HttpRequest(), "text"))
+        self.assertFalse(is_spam(HttpRequest(), ["text"]))
 
     @skipIf(not HAS_AKISMET, "akismet module not installed")
     @responses.activate
     @override_settings(AKISMET_API_KEY="key")
     def test_akismet_definite_spam(self) -> None:
         self.mock_akismet("true", headers={"X-Akismet-Pro-Tip": "discard"})
-        self.assertTrue(is_spam("text", HttpRequest()))
+        self.assertTrue(is_spam(HttpRequest(), "text"))
+        self.assertTrue(is_spam(HttpRequest(), ["text"]))
 
     @skipIf(not HAS_AKISMET, "akismet module not installed")
     @responses.activate
     @override_settings(AKISMET_API_KEY="key")
     def test_akismet_nospam(self) -> None:
         self.mock_akismet("false")
-        self.assertFalse(is_spam("text", HttpRequest()))
+        self.assertFalse(is_spam(HttpRequest(), "text"))
+        self.assertFalse(is_spam(HttpRequest(), ["text"]))
 
     @skipIf(not HAS_AKISMET, "akismet module not installed")
     @responses.activate