[UK] Fix broken table parsing.

Author: dracos

pyscraper/gidmatching.py

@@ -46,8 +46,8 @@ def PrepareXMLForDiff(scrapeversion):
 		assert chk[0] == chk[3]  # chunk type (this can fail if due to the lack of two \n's between the two labels, and thus detects an empty speech, which should not be there.  
 		# new_chk = chk[2]
 		new_chk = re.sub(
-			'(?s)(<p\s[^>]*>)(.*?)(<\/p>)',
-			lambda m: (u''.join((m.group(1), re.sub('\n', ' ', m.group(2)), m.group(3)))),
+			r'(?s)(<(p|tr)\s[^>]*>)(.*?)(<\/\2>)',
+			lambda m: (u''.join((m.group(1), re.sub('\n', ' ', m.group(3)), m.group(4)))),
 			chk[2]
 		)
 		essxindx.append(len(essxlist))

pyscraper/new_hansard.py

@@ -763,20 +763,23 @@ def parse_votelist(self, votes, direction, vote_list, is_teller=False):
         return vote_list
 
     def parse_table(self, wrapper):
-        rows = wrapper.xpath('.//row')
+        rows = wrapper.xpath('.//ns:row', namespaces=self.ns_map)
         tag = etree.Element('table')
         body = etree.Element('tbody')
         url = None
         for row in rows:
             row_tag = etree.Element('tr')
             row_tag.set('pid', self.get_pid())
 
-            for entry in row.xpath('(.//hs_brev|.//hs_Para)'):
+            for entry in row.xpath('(.//ns:hs_brev|.//ns:hs_Para)', namespaces=self.ns_map):
                 if url is None:
                     url = entry.get('url')
-                row_tag.append(list(entry))
+                td_tag = etree.Element('td')
+                td_tag.text = self.get_single_line_text_from_element(entry)
+                row_tag.append(td_tag)
 
-            body.append(row_tag)
+            if len(row_tag):
+                body.append(row_tag)
 
         tag.append(body)