move checks for existing commentary and bad request to ArxivQueryForm

commentaries/forms.py

@@ -6,7 +6,7 @@ from django.urls import reverse
 
 from .models import Commentary
 
-from scipost.services import DOICaller
+from scipost.services import DOICaller, ArxivCaller
 from scipost.models import Contributor
 
 
@@ -48,10 +48,21 @@ class ArxivQueryForm(forms.Form):
 
     identifier = forms.RegexField(regex=VALID_ARXIV_IDENTIFIER_REGEX, strip=True)
 
-#         identifierpattern_new = re.compile("^[0-9]{4,}.[0-9]{4,5}v[0-9]{1,2}$")
-#         identifierpattern_old = re.compile("^[-.a-z]+/[0-9]{7,}v[0-9]{1,2}$")
+    def clean_identifier(self):
+        identifier = self.cleaned_data['identifier']
+
+        if Commentary.objects.filter(arxiv_identifier=identifier).exists():
+            error_message = 'There already exists a Commentary Page on this arXiv preprint.'
+            raise forms.ValidationError(error_message)
 
+        caller = ArxivCaller(identifier)
+        if caller.is_valid:
+            self.arxiv_data = ArxivCaller(identifier).data
+        else:
+            error_message = 'Could not find a resource for that arXiv identifier.'
+            raise forms.ValidationError(error_message)
 
+        return identifier
 
 
 # class IdentifierToQueryForm(forms.Form):

commentaries/test_forms.py

@@ -12,6 +12,9 @@ from common.helpers.test import add_groups_and_permissions
 
 
 class TestArxivQueryForm(TestCase):
+    def setUp(self):
+        add_groups_and_permissions()
+
     def test_new_arxiv_identifier_is_valid(self):
         new_identifier_data = {'identifier': '1612.07611v1'}
         form = ArxivQueryForm(new_identifier_data)
@@ -37,6 +40,19 @@ class TestArxivQueryForm(TestCase):
         form = ArxivQueryForm(data)
         self.assertFalse(form.is_valid())
 
+    def test_arxiv_identifier_that_already_has_commentary_page_is_invalid(self):
+        unvetted_commentary = UnvettedCommentaryFactory()
+        invalid_data = {'identifier': unvetted_commentary.arxiv_identifier}
+        form = ArxivQueryForm(invalid_data)
+        self.assertFalse(form.is_valid())
+        error_message = form.errors['identifier'][0]
+        self.assertRegexpMatches(error_message, re.compile('already exist'))
+
+    def test_valid_but_non_existent_identifier_is_invalid(self):
+        invalid_data = {'identifier': '1613.07611v1'}
+        form = ArxivQueryForm(invalid_data)
+        self.assertFalse(form.is_valid())
+
 
 class TestDOIToQueryForm(TestCase):
     def setUp(self):

scipost/services.py

@@ -77,9 +77,10 @@ class ArxivCaller:
     def _call_arxiv(self):
         url = self.query_base_url % self.identifier
         request = requests.get(url)
-        if request.ok:
+        arxiv_data = feedparser.parse(request.content)['entries'][0]
+        if self._search_result_present(arxiv_data):
             self.is_valid = True
-            self._arxiv_data = feedparser.parse(request.content)['entries'][0]
+            self._arxiv_data = arxiv_data
         else:
             self.is_valid = False
 
@@ -97,6 +98,9 @@ class ArxivCaller:
             'pub_abstract': abstract,
         }
 
+    def _search_result_present(self, data):
+        return 'title' in data
+
 
 # I'm going to revamp this whole thing...
 class BaseCaller(object):

scipost/test_services.py

@@ -24,7 +24,7 @@ class ArxivCallerTest(TestCase):
 
     def valid_but_nonexistent_identifier(self):
         caller = ArxivCaller('1613.07611v1')
-        self.assertFalse(caller.is_valid)
+        self.assertEqual(caller.is_valid, False)
 
 
 class DOICallerTest(TestCase):