Merge pull request #266 from scholarly-python-package/develop

Develop

docs/index.rst

@@ -7,7 +7,7 @@ Welcome to scholarly's documentation!
 ========================================
 
 .. toctree::
-   :maxdepth: 2
+   :maxdepth: 5
    :caption: Index
 
    quickstart

scholarly/_scholarly.py

@@ -407,11 +407,14 @@ def pprint(self, object: Author or Publication)->None:
         print(pprint.pformat(to_print))
 
     def search_org(self, name: str, fromauthor: bool = False) -> list:
-        """Search by organization name and return a list of possible disambiguations
+        """
+        Search by organization name and return a list of possible disambiguations
+
         :Example::
             .. testcode::
                 search_query = scholarly.search_org('ucla')
                 print(search_query)
+
         :Output::
             .. testoutput::
                 [{'Organization': 'University of California, Los Angeles',

scholarly/author_parser.py

@@ -85,7 +85,10 @@ def _fill_basics(self, soup, author):
             if "avatar_scholar" in picture:
                 picture = _HOST.format(picture)
             author['url_picture'] = picture
-
+        index = soup.find_all('td', class_='gsc_rsb_std')
+        if index:
+            author['citedby'] = int(index[0].text)
+
     def _fill_indices(self, soup, author):
         index = soup.find_all('td', class_='gsc_rsb_std')
         if index:

scholarly/publication_parser.py

@@ -51,6 +51,7 @@ def __init__(self, nav, url: str):
         self._url = url
         self._nav = nav
         self._load_url(url)
+        self.total_results = self._get_total_results()
         self.pub_parser = PublicationParser(self._nav)
 
     def _load_url(self, url: str):
@@ -59,6 +60,14 @@ def _load_url(self, url: str):
         self._pos = 0
         self._rows = self._soup.find_all('div', class_='gs_r gs_or gs_scl')
 
+    def _get_total_results(self):
+        for x in self._soup.find_all('div', class_='gs_ab_mdw'):
+            # Decimal separator is set by Google independent of language setting
+            match = re.match(pattern=r'(^|\s*About)\s*([0-9,\.]+)', string=x.text)
+            if match:
+                return int(re.sub(pattern=r'[,\.]',repl='', string=match.group(2)))
+        return None
+
     # Iterator protocol
 
     def __iter__(self):

setup.py

@@ -5,7 +5,7 @@
 
 setuptools.setup(
     name='scholarly',
-    version='1.0.7',
+    version='1.1.0',
     author='Steven A. Cholewiak, Panos Ipeirotis, Victor Silva',
     author_email='steven@cholewiak.com, panos@stern.nyu.edu, vsilva@ualberta.ca',
     description='Simple access to Google Scholar authors and citations',

test_module.py

@@ -180,6 +180,16 @@ def test_search_pubs(self):
 
         self.assertIn('Visual perception of the physical stability of asymmetric three-dimensional objects', pubs)
 
+    def test_search_pubs_total_results(self):
+        """
+        As of February 4, 2021 there are 32 pubs that fit the search term:
+        ["naive physics" stability "3d shape"].
+
+        Check that the total results for that search term equals 32.
+        """
+        pubs = scholarly.search_pubs('"naive physics" stability "3d shape"')
+        self.assertGreaterEqual(pubs.total_results, 32)
+
     def test_search_pubs_filling_publication_contents(self):
         '''
         This process  checks the process of filling a publication that is derived