Fixing flickr parsing

[redakcja.git] / apps / cover / forms.py

diff --git a/apps/cover/forms.py b/apps/cover/forms.py
index 754d697..0173134 100755 (executable)
--- a/apps/cover/forms.py
+++ b/apps/cover/forms.py
@@ -6,6 +6,7 @@
 import re
 from urllib2 import urlopen
 from django import forms
+from django.utils.translation import ugettext_lazy as _
 from cover.models import Image
 
 class ImageAddForm(forms.ModelForm):
@@ -33,18 +34,21 @@ class ReadonlyImageEditForm(ImageEditForm):
 
 
 class FlickrForm(forms.Form):
-    source_url = forms.URLField()
+    source_url = forms.URLField(label=_('Flickr URL'))
 
     def clean_source_url(self):
+        def normalize_html(html):
+            return re.sub('[\t\n]', '', html)
+    
         url = self.cleaned_data['source_url']
-        m = re.match(r'(https?://)?(www.)?flickr.com/photos/(?P<author>[^/]+)/(?P<img>\d+)', url)
+        m = re.match(r'(https?://)?(www\.|secure\.)?flickr\.com/photos/(?P<author>[^/]+)/(?P<img>\d+)/?', url)
         if not m:
             raise forms.ValidationError("It doesn't look like Flickr URL.")
         author_slug, img_id = m.group('author'), m.group('img')
         base_url = "https://www.flickr.com/photos/%s/%s/" % (author_slug, img_id)
 
         try:
-            html = urlopen(url).read().decode('utf-8')
+            html = normalize_html(urlopen(url).read().decode('utf-8'))
         except:
             raise forms.ValidationError('Error reading page.')
         match = re.search(r'<a href="([^"]*)" rel="license cc:license">Some rights reserved</a>', html)
@@ -59,7 +63,7 @@ class FlickrForm(forms.Form):
         except AssertionError:
             raise forms.ValidationError('Error reading license name.')
 
-        m = re.search(r'<strong class="username">By <a href="[^"]*">([^<]*)</a></strong>', html)
+        m = re.search(r'"ownername":"([^"]*)', html)
         if m:
             self.cleaned_data['author'] = "%s@Flickr" % m.group(1)
         else:
@@ -71,7 +75,7 @@ class FlickrForm(forms.Form):
         self.cleaned_data['title'] = m.group(1)
 
         url_size = base_url + "sizes/o/"
-        html = urlopen(url_size).read().decode('utf-8')
+        html = normalize_html(urlopen(url_size).read().decode('utf-8'))
         m = re.search(r'<div id="allsizes-photo">\s*<img src="([^"]*)"', html)
         if m:
             self.cleaned_data['download_url'] = m.group(1)