Merge branch 'hotfix/23.16.6' into develop

osf/external/spam/tasks.py

@@ -46,15 +46,20 @@ def _check_resource_for_domains(guid, content):
     resource = guid.referent
     spammy_domains = []
     referrer_content_type = ContentType.objects.get_for_model(resource)
-    for domain in _extract_domains(content):
-        notable_domain, _ = NotableDomain.objects.get_or_create(domain=domain)
+    for domain, note in _extract_domains(content):
+        notable_domain, _ = NotableDomain.objects.get_or_create(
+            domain=domain,
+            defaults={'note': note}
+        )
         if notable_domain.note == NotableDomain.Note.EXCLUDE_FROM_ACCOUNT_CREATION_AND_CONTENT:
             spammy_domains.append(notable_domain.domain)
         DomainReference.objects.get_or_create(
             domain=notable_domain,
             referrer_object_id=resource.id,
             referrer_content_type=referrer_content_type,
-            defaults={'is_triaged': notable_domain.note != NotableDomain.Note.UNKNOWN}
+            defaults={
+                'is_triaged': notable_domain.note not in (NotableDomain.Note.UNKNOWN, NotableDomain.Note.UNVERIFIED)
+            }
         )
     if spammy_domains:
         resource.confirm_spam(save=True, domains=list(spammy_domains))
@@ -72,8 +77,11 @@ def check_resource_for_domains_async(guid, content):
 
 
 def _extract_domains(content):
+    from osf.models import NotableDomain
+
     extracted_domains = set()
     for match in DOMAIN_REGEX.finditer(content):
+        note = NotableDomain.Note.UNKNOWN
         domain = match.group('domain')
         if not domain or domain in extracted_domains:
             continue
@@ -85,10 +93,11 @@ def _extract_domains(content):
 
         try:
             response = requests.head(constructed_url, timeout=settings.DOMAIN_EXTRACTION_TIMEOUT)
-        except (requests.exceptions.ConnectionError, requests.exceptions.InvalidURL):
+        except requests.exceptions.InvalidURL:
+            # Likely false-positive from a filename.ext
             continue
         except requests.exceptions.RequestException:
-            pass
+            note = NotableDomain.Note.UNVERIFIED
         else:
             # Store the redirect location (to help catch link shorteners)
             if response.status_code in REDIRECT_CODES and 'location' in response.headers:
@@ -99,7 +108,7 @@ def _extract_domains(content):
         # Avoid returning a duplicate domain discovered via redirect
         if domain not in extracted_domains:
             extracted_domains.add(domain)
-            yield domain
+            yield domain, note
 
 
 @run_postcommit(once_per_request=False, celery=True)

osf/migrations/0017_alter_notabledomain_note.py

@@ -0,0 +1,19 @@
+# Generated by Django 3.2.17 on 2023-12-12 19:02
+
+from django.db import migrations, models
+import osf.models.notable_domain
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('osf', '0016_auto_20230828_1810'),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name='notabledomain',
+            name='note',
+            field=models.IntegerField(choices=[(0, 'EXCLUDE_FROM_ACCOUNT_CREATION_AND_CONTENT'), (1, 'ASSUME_HAM_UNTIL_REPORTED'), (2, 'UNKNOWN'), (3, 'IGNORED'), (4, 'UNVERIFIED')], default=osf.models.notable_domain.NotableDomain.Note['UNKNOWN']),
+        ),
+    ]

osf/models/notable_domain.py

@@ -14,6 +14,7 @@ class Note(IntEnum):
         ASSUME_HAM_UNTIL_REPORTED = 1
         UNKNOWN = 2
         IGNORED = 3
+        UNVERIFIED = 4  # Timedout couldn't determine
 
         @classmethod
         def choices(cls):

osf_tests/test_notable_domains.py

@@ -30,31 +30,34 @@ def test_extract_domains__optional_components(self, protocol_component, www_comp
         sample_text = f'This is a link: {test_url}'
         with mock.patch.object(spam_tasks.requests, 'head'):
             domains = list(spam_tasks._extract_domains(sample_text))
-        assert domains == ['osf.io']
+        assert domains == [('osf.io', NotableDomain.Note.UNKNOWN)]
 
     def test_extract_domains__url_in_quotes(self):
         sample_text = '"osf.io"'
         with mock.patch.object(spam_tasks.requests, 'head'):
             domains = list(spam_tasks._extract_domains(sample_text))
-        assert domains == ['osf.io']
+        assert domains == [('osf.io', NotableDomain.Note.UNKNOWN)]
 
     def test_extract_domains__url_in_parens(self):
         sample_text = '(osf.io)'
         with mock.patch.object(spam_tasks.requests, 'head'):
             domains = list(spam_tasks._extract_domains(sample_text))
-        assert domains == ['osf.io']
+        assert domains == [('osf.io', NotableDomain.Note.UNKNOWN)]
 
     def test_extract_domains__captures_domain_with_multiple_subdomains(self):
         sample_text = 'This is a link: https://api.test.osf.io'
         with mock.patch.object(spam_tasks.requests, 'head'):
             domains = list(spam_tasks._extract_domains(sample_text))
-        assert domains == ['api.test.osf.io']
+        assert domains == [('api.test.osf.io', NotableDomain.Note.UNKNOWN)]
 
     def test_extract_domains__captures_multiple_domains(self):
         sample_text = 'This is a domain: http://osf.io. This is another domain: www.cos.io'
         with mock.patch.object(spam_tasks.requests, 'head'):
             domains = set(spam_tasks._extract_domains(sample_text))
-        assert domains == {'osf.io', 'cos.io'}
+        assert domains == {
+            ('osf.io', NotableDomain.Note.UNKNOWN),
+            ('cos.io', NotableDomain.Note.UNKNOWN),
+        }
 
     def test_extract_domains__no_domains(self):
         sample_text = 'http://fakeout!'
@@ -63,19 +66,19 @@ def test_extract_domains__no_domains(self):
         assert not domains
         mock_head.assert_not_called()
 
-    def test_extract_domains__ignored_if_does_not_resolve(self):
+    def test_extract_domains__unverfied_if_does_not_resolve(self):
         sample_text = 'This.will.not.connect'
         with mock.patch.object(spam_tasks.requests, 'head') as mock_head:
             mock_head.side_effect = spam_tasks.requests.exceptions.ConnectionError
             domains = set(spam_tasks._extract_domains(sample_text))
-        assert not domains
+        assert domains == {('This.will.not.connect', NotableDomain.Note.UNVERIFIED)}
 
     def test_actract_domains__returned_on_error(self):
         sample_text = 'This.will.timeout'
         with mock.patch.object(spam_tasks.requests, 'head') as mock_head:
             mock_head.side_effect = spam_tasks.requests.exceptions.Timeout
             domains = set(spam_tasks._extract_domains(sample_text))
-        assert domains == {sample_text}
+        assert domains == {(sample_text, NotableDomain.Note.UNVERIFIED)}
 
     @pytest.mark.parametrize('status_code', [301, 302, 303, 307, 308])
     def test_extract_domains__follows_redirect(self, status_code):
@@ -85,7 +88,7 @@ def test_extract_domains__follows_redirect(self, status_code):
         sample_text = 'redirect.me'
         with mock.patch.object(spam_tasks.requests, 'head', return_value=mock_response):
             domains = list(spam_tasks._extract_domains(sample_text))
-        assert domains == ['redirected.com']
+        assert domains == [('redirected.com', NotableDomain.Note.UNKNOWN)]
 
     def test_extract_domains__redirect_code_no_location(self):
         mock_response = SimpleNamespace()
@@ -94,7 +97,7 @@ def test_extract_domains__redirect_code_no_location(self):
         sample_text = 'redirect.me'
         with mock.patch.object(spam_tasks.requests, 'head', return_value=mock_response):
             domains = list(spam_tasks._extract_domains(sample_text))
-        assert domains == ['redirect.me']
+        assert domains == [('redirect.me', NotableDomain.Note.UNKNOWN)]
 
     def test_extract_domains__redirect_code_bad_location(self):
         mock_response = SimpleNamespace()
@@ -103,7 +106,7 @@ def test_extract_domains__redirect_code_bad_location(self):
         sample_text = 'redirect.me'
         with mock.patch.object(spam_tasks.requests, 'head', return_value=mock_response):
             domains = list(spam_tasks._extract_domains(sample_text))
-        assert domains == ['redirect.me']
+        assert domains == [('redirect.me', NotableDomain.Note.UNKNOWN)]
 
     def test_extract_domains__redirect_with_full_url_no_protocol(self):
         mock_response = SimpleNamespace()
@@ -114,7 +117,7 @@ def test_extract_domains__redirect_with_full_url_no_protocol(self):
         with mock.patch.object(spam_tasks.requests, 'head', return_value=mock_response) as mock_object:
             domains = list(spam_tasks._extract_domains(sample_text))
             mock_object.assert_called_once_with(f'https://{target_url}', timeout=60)
-        assert domains == ['osf.io']
+        assert domains == [('osf.io', NotableDomain.Note.UNKNOWN)]
 
     def test_extract_domains__redirect_with_full_url_and_protocol(self):
         mock_response = SimpleNamespace()
@@ -125,13 +128,13 @@ def test_extract_domains__redirect_with_full_url_and_protocol(self):
         with mock.patch.object(spam_tasks.requests, 'head', return_value=mock_response) as mock_object:
             domains = list(spam_tasks._extract_domains(sample_text))
             mock_object.assert_called_once_with(target_url, timeout=60)
-        assert domains == ['osf.io']
+        assert domains == [('osf.io', NotableDomain.Note.UNKNOWN)]
 
     def test_extract_domains__deduplicates(self):
         sample_text = 'osf.io osf.io osf.io and, oh, yeah, osf.io'
         with mock.patch.object(spam_tasks.requests, 'head'):
             domains = list(spam_tasks._extract_domains(sample_text))
-        assert domains == ['osf.io']
+        assert domains == [('osf.io', NotableDomain.Note.UNKNOWN)]
 
     def test_extract_domains__ignores_floats(self):
         sample_text = 'this is a number 3.1415 not a domain'