Merge pull request #536 from Luen/main

Add in PDF link in publication fill

Author: arunkannawadi

scholarly/publication_parser.py

@@ -202,6 +202,10 @@ def _scholar_pub(self, __data, publication: Publication):
         if title.find('a'):
             publication['pub_url'] = title.find('a')['href']
 
+        pdf_div = __data.find('div', class_='gs_ggs gs_fl')
+        if pdf_div and pdf_div.find('a', href=True):
+            publication['eprint_url'] = pdf_div.find('a')['href']
+
         author_div_element = databox.find('div', class_='gs_a')
         authorinfo = author_div_element.text
         authorinfo = authorinfo.replace(u'\xa0', u' ')       # NBSP
@@ -286,6 +290,10 @@ def fill(self, publication: Publication)->Publication:
             if soup.find('a', class_='gsc_oci_title_link'):
                 publication['pub_url'] = soup.find(
                     'a', class_='gsc_oci_title_link')['href']
+            if soup.find('div', class_='gsc_oci_title_ggi'):
+                link = soup.find('a', attrs={'data-clk': True})
+                if link:
+                    publication['eprint_url'] = link['href']
             for item in soup.find_all('div', class_='gs_scl'):
                 key = item.find(class_='gsc_oci_field').text.strip().lower()
                 val = item.find(class_='gsc_oci_value')

test_module.py

@@ -724,7 +724,7 @@ def test_search_pubs(self):
         pubs = list(scholarly.search_pubs('"naive physics" stability "3d shape"'))
         # Check that the first entry in pubs is the same as pub.
         # Checking for quality holds for non-dict entries only.
-        for key in {'author_id', 'pub_url', 'num_citations'}:
+        for key in {'author_id', 'pub_url', 'eprint_url', 'num_citations'}:
             self.assertEqual(pub[key], pubs[0][key])
         for key in {'title', 'pub_year', 'venue'}:
             self.assertEqual(pub['bib'][key], pubs[0]['bib'][key])
@@ -784,6 +784,7 @@ def test_search_pubs_filling_publication_contents(self):
         self.assertTrue(f['bib']['publisher'] == u'The Association for Research in Vision and Ophthalmology')
         self.assertTrue(f['bib']['title'] == u'Creating correct blur and its effect on accommodation')
         self.assertTrue(f['pub_url'] == u'https://jov.arvojournals.org/article.aspx?articleid=2701817')
+        self.assertTrue(f['eprint_url'] == u'https://jov.arvojournals.org/arvo/content_public/journal/jov/937491/i1534-7362-18-9-1.pdf')
         self.assertTrue(f['bib']['volume'] == '18')
         self.assertTrue(f['bib']['pub_year'] == u'2018')
 
@@ -800,6 +801,7 @@ def test_related_articles_from_author(self):
         # Typically, the same publication is returned as the most related article
         same_article = next(related_articles)
         self.assertEqual(pub["pub_url"], same_article["pub_url"])
+        self.assertEqual(pub["eprint_url"], same_article["eprint_url"])
         for key in {'title', 'pub_year'}:
             self.assertEqual(str(pub['bib'][key]), (same_article['bib'][key]))
 
@@ -818,7 +820,7 @@ def test_related_articles_from_publication(self):
         related_articles = scholarly.get_related_articles(pub)
         # Typically, the same publication is returned as the most related article
         same_article = next(related_articles)
-        for key in {'author_id', 'pub_url', 'num_citations'}:
+        for key in {'author_id', 'pub_url', 'eprint_url', 'num_citations'}:
             self.assertEqual(pub[key], same_article[key])
         for key in {'title', 'pub_year'}:
             self.assertEqual(pub['bib'][key], same_article['bib'][key])