Bug fix when rejecting record with invalid OAI.

modules/harvest_tools/automaton.py

@@ -450,7 +450,7 @@ class Automaton(object):
         # the search is perform on a range of creation date
         # if not defined all element are return
         #
-        # The method use here minimise the memory usage
+        # The method use here minimize the memory usage
         # on the server as well as on the client side
 
         for collection in collections:
@@ -511,8 +511,9 @@ class Automaton(object):
         if self.dbg:
             print "process xml record"
 
-        # NOTE: BaseException and inherited class
-        # are catched by the previous stage
+        # NOTE
+        # BaseException and inherited class
+        # are caught by the previous stage
         records = self.marc12(xml)
 
         # process individual record
@@ -529,13 +530,15 @@ class Automaton(object):
 
             # reject record with undefined OAI field
             oai = record.oai()
-            if not oai:
+            if oai is None:
                 self.logs[-1].reject(MSG_NO_OAI, record.year())
+                continue
 
-            # reject record is not well form OAI
+            # reject the record when the OAI is not well
             match = REG_OAI.match(oai)
             if not match:
                 self.logs[-1].reject(MSG_WELL_FORM_OAI, record.year())
+                continue
 
             # check that the record is well formed
             # repair non-conformity as far as possible

modules/invenio_tools/checkandfix.py

@@ -409,26 +409,26 @@ class CheckAndFix(object):
                 raise CheckException(MSG_WELL_FORMED_CONF_DATES)
 
     def is_bad_oai_used(self, record):
-        """The id in the OAI field might be different from the record id.
-        It happens when an old record is redirected to new one for obscure
-        reasons.
-
-        Return C{TRue} if the "bad OAI" is found in the database.
+        """The id in the OAI field is different from the record id.
+        This happens when an old record is redirected to new one
+        for obscure reasons. The method return C{True} when a record
+        with the bad OAI is found in the database.
 
         @type record: L{Record}
         @param record:
 
         @rtype: bool
+        @return: C{True} when a record is found in the database with
+        the bad OAI.
 
         """
         value = record.oai()
         match = REG_OAI.match(value)
-        myid = record.id()
 
-        if match.group(2) != myid:
+        if match.group(2) != record.id():
             db = self.db
 
-            # The bad OAI is already used in the database
+            # a record with the bad OAI exists in the database
             bad_oai_url = OAI_URL % (match.group(1), match.group(2))
             if get_id(db.publications, origin=bad_oai_url):
                 return True