resolver voor DPO toegevoegd

Author: xiffy

kb/nl/api/oai.py

@@ -7,7 +7,7 @@
 OAI_BASEURL = 'http://services.kb.nl/mdo/oai'
 
 
-class oai():
+class oai:
 
     """
         OAI interface to the National Library of the Netherlands.

kb/nl/api/sru.py

@@ -1,7 +1,7 @@
 import sys
 import requests
 import urllib
-import pprint 
+import pprint
 
 try:
     from urllib import quote  # Python 2.X
@@ -118,7 +118,7 @@ def __next__(self):
             return response(record_data, self.sru)
         else:
             raise StopIteration
-    
+
     def next(self):
         return self.__next__()
 

kb/nl/collections/__init__.py

@@ -13,7 +13,7 @@
                 'recordschema': 'ddd',
                 'setname': 'DPO',
                 'time_period': [1781, 1800],
-                'resolver' : ''},
+                'resolver' : 'http://resolver.kb.nl/resolve?urn='},
         'BYVANCK': {'collection': 'BYVANCK',
                     'description_en': 'Medieval Illuminated Manuscripts',
                     'description_nl': 'Middeleeuwse Verluchte Handschriften',

kb/nl/helpers/__init__.py

@@ -18,11 +18,19 @@
                     import elementtree.ElementTree as etree
                 except ImportError:
                     raise("Failed to import ElementTree from any known place")
-
+import codecs
 
 def alto_to_text(alto_data):
     ''' Grab the selected text blocks and write them to disk '''
-    alto_data = etree.fromstring(alto_data.encode('utf-8'))
+    try:
+        # some files have a BOM available which does not get stripped on windows
+        if not alto_data[0] == '<':
+            alto_data = alto_data[3:]
+        alto_data = etree.fromstring(alto_data.encode('utf-8'))
+
+    except etree.XMLSyntaxError as e:
+        print ("error! %s" % alto_str[:40])
+        return None
 
     alto_text = u""
     prev_was_hyp = False