Merge pull request #575 from mouse-reeve/openlibrary-editions

Ignore openlibrary editions with little to no metadata

bookwyrm/connectors/openlibrary.py

@@ -27,9 +27,9 @@ class Connector(AbstractConnector):
             Mapping('series', formatter=get_first),
             Mapping('seriesNumber', remote_field='series_number'),
             Mapping('subjects'),
-            Mapping('subjectPlaces'),
+            Mapping('subjectPlaces', remote_field='subject_places'),
-            Mapping('isbn13', formatter=get_first),
+            Mapping('isbn13', remote_field='isbn_13', formatter=get_first),
-            Mapping('isbn10', formatter=get_first),
+            Mapping('isbn10', remote_field='isbn_10', formatter=get_first),
             Mapping('lccn', formatter=get_first),
             Mapping(
                 'oclcNumber', remote_field='oclc_numbers',
@@ -144,9 +144,34 @@ class Connector(AbstractConnector):
         # we can mass download edition data from OL to avoid repeatedly querying
         edition_options = self.load_edition_data(work.openlibrary_key)
         for edition_data in edition_options.get('entries'):
+            # does this edition have ANY interesting data?
+            if ignore_edition(edition_data):
+                continue
             self.create_edition_from_data(work, edition_data)
 
 
+def ignore_edition(edition_data):
+    ''' don't load a million editions that have no metadata '''
+    # an isbn, we love to see it
+    if edition_data.get('isbn_13') or edition_data.get('isbn_10'):
+        print(edition_data.get('isbn_10'))
+        return False
+    # grudgingly, oclc can stay
+    if edition_data.get('oclc_numbers'):
+        print(edition_data.get('oclc_numbers'))
+        return False
+    # if it has a cover it can stay
+    if edition_data.get('covers'):
+        print(edition_data.get('covers'))
+        return False
+    # keep non-english editions
+    if edition_data.get('languages') and \
+            'languages/eng' not in str(edition_data.get('languages')):
+        print(edition_data.get('languages'))
+        return False
+    return True
+
+
 def get_description(description_blob):
     ''' descriptions can be a string or a dict '''
     if isinstance(description_blob, dict):

bookwyrm/tests/connectors/test_openlibrary_connector.py

@@ -190,3 +190,19 @@ class Openlibrary(TestCase):
         ''' detect if the loaded json is an edition '''
         edition = pick_default_edition(self.edition_list_data['entries'])
         self.assertEqual(edition['key'], '/books/OL9788823M')
+
+
+    def test_create_edition_from_data(self):
+        ''' okay but can it actually create an edition with proper metadata '''
+        work = models.Work.objects.create(title='Hello')
+        result = self.connector.create_edition_from_data(
+            work, self.edition_data)
+        self.assertEqual(result.parent_work, work)
+        self.assertEqual(result.title, 'Sabriel')
+        self.assertEqual(result.isbn_10, '0060273224')
+        self.assertIsNotNone(result.description)
+        self.assertEqual(result.languages[0], 'English')
+        self.assertEqual(result.publishers[0], 'Harper Trophy')
+        self.assertEqual(result.pages, 491)
+        self.assertEqual(result.subjects[0], 'Fantasy.')
+        self.assertEqual(result.physical_format, 'Hardcover')

bookwyrm/tests/data/ol_edition.json

@@ -9,6 +9,7 @@
             "518848"
         ]
     },
+    "physical_format": "Hardcover",
     "lc_classifications": [
         "PZ7.N647 Sab 1995"
     ],