Fix bug: dbxref crashes when a pfam entry is returned as html

dbxref/retrieve/pfam.py

@@ -2,6 +2,7 @@
 import dbxref.resolver
 import requests
 import xml.etree.ElementTree as ET
+from xml.etree.ElementTree import ParseError
 import logging
 import json
 import argparse
@@ -35,11 +36,12 @@ def retrieve(dbxrefs, basic=True, annotation=True):
         logger.debug('URL: %s', xml_url)
         r = requests.get(xml_url)
         logger.debug('Content: %s', r.text)
-        root = ET.fromstring(r.text)
 
         output = {'id': entry['dbxref']}
 
         try:
+          root = ET.fromstring(r.text)
+
           tree = str(ET.tostring(root))
           if '<error>' in tree:
                output['message'] = tree[tree.find('<error>')+7:tree.rfind('</error>')]
@@ -52,7 +54,7 @@ def retrieve(dbxrefs, basic=True, annotation=True):
         except (KeyError, AttributeError) as e:
             logger.warn('Error in retrieving %s', str(entry))
             raise
-        except RuntimeError as e:
+        except (ParseError, RuntimeError) as e:
             output['message'] = 'an error occurred'
             try:
                 html = HTML.document_fromstring(r.text.replace('\n', ' '))

tests/test_pfam.py

@@ -8,3 +8,11 @@ class TestPfam(unittest.TestCase):
       documents = pfam.retrieve([{'db': 'PFAM', 'id': 'PF00083.23'}])
       # this test failed due to an error due to missing None handling, 
       # so no assertions here. Once fixed. this should suffice
+
+    def test_renamed_family(self):
+      '''regression test for missing comment in pfam entry'''
+      documents = pfam.retrieve([{'db': 'PFAM', 'id': 'Tiny_TM_bacill'}])
+      # this test failed due to a redirect when a family was renamed
+      # unfortunately the redirect was not encoded in http headers, but in 
+      # html markup (<meta http-equiv="Refresh" content="5; URL=/family/PF09680" />)
+      # so no assertions here. Once fixed. this should suffice