[ckan-changes] commit/ckanext-inspire: amercader: Create a licence_url extra with the first URL found in the licence extra

Wed May 11 15:22:36 UTC 2011

1 new changeset in ckanext-inspire:

http://bitbucket.org/okfn/ckanext-inspire/changeset/106372eb0003/
changeset:   r28:106372eb0003
user:        amercader
date:        2011-05-11 17:22:24
summary:     Create a licence_url extra with the first URL found in the licence extra
affected #:  1 file (462 bytes)

--- a/ckanext/inspire/harvesters.py	Wed May 11 14:45:43 2011 +0100
+++ b/ckanext/inspire/harvesters.py	Wed May 11 16:22:24 2011 +0100
@@ -10,6 +10,7 @@
 '''
 from lxml import etree
 import urllib2
+from urlparse import urlparse
 from datetime import datetime
 
 import logging
@@ -249,6 +250,11 @@
             extras[name] = gemini_values[name]
 
         extras['licence'] = gemini_values.get('use-constraints', '')
+        if len(extras['licence']):
+            license_url_extracted = self._extract_first_license_url(extras['licence'])
+            if license_url_extracted:
+                extras['licence_url'] = license_url_extracted
+ 
         extras['access_constraints'] = gemini_values.get('limitations-on-public-access','')
         if gemini_values.has_key('temporal-extent-begin'):
             #gemini_values['temporal-extent-begin'].sort()
@@ -338,6 +344,13 @@
                 counter = counter + 1
             return None
 
+    def _extract_first_license_url(self,licences):
+        for licence in licences:
+            o = urlparse(licence)
+            if o.scheme and o.netloc:
+                return licence
+        return None
+
     def _create_package_from_data(self, package_data, package = None):
         '''
         {'extras': {'INSPIRE': 'True',

Repository URL: https://bitbucket.org/okfn/ckanext-inspire/

--

This is a commit notification from bitbucket.org. You are receiving
this because you have the service enabled, addressing the recipient of
this email.


