basic integrity check,
[wolnelektury.git] / apps / opds / views.py
index cd91a74..f793d7d 100644 (file)
@@ -2,10 +2,8 @@
 # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
 #
-from base64 import b64encode
 import os.path
 from urlparse import urljoin
-from urllib2 import unquote
 
 from django.contrib.syndication.views import Feed
 from django.core.urlresolvers import reverse
@@ -18,11 +16,14 @@ from django.contrib.sites.models import Site
 from basicauth import logged_in_or_basicauth, factory_decorator
 from catalogue.models import Book, Tag
 
-from search import MultiSearch, SearchResult, JVM
+from search.views import get_search, SearchResult, JVM
 from lucene import Term, QueryWrapperFilter, TermQuery
 
+import logging
 import re
 
+log = logging.getLogger('opds')
+
 from stats.utils import piwik_track
 
 _root_feeds = (
@@ -187,19 +188,19 @@ class AcquisitionFeed(Feed):
             return u''
 
     def item_enclosure_url(self, book):
-        return full_url(book.root_ancestor.epub_file.url)
+        return full_url(book.epub_file.url) if book.epub_file else None
 
     def item_enclosure_length(self, book):
-        return book.root_ancestor.epub_file.size
+        return book.epub_file.size if book.epub_file else None
 
 @piwik_track
 class RootFeed(Feed):
     feed_type = OPDSFeed
     title = u'Wolne Lektury'
-    link = u'http://www.wolnelektury.pl/'
+    link = u'http://wolnelektury.pl/'
     description = u"Spis utworów na stronie http://WolneLektury.pl"
     author_name = u"Wolne Lektury"
-    author_link = u"http://www.wolnelektury.pl/"
+    author_link = u"http://wolnelektury.pl/"
 
     def items(self):
         return _root_feeds
@@ -216,10 +217,10 @@ class RootFeed(Feed):
 @piwik_track
 class ByCategoryFeed(Feed):
     feed_type = OPDSFeed
-    link = u'http://www.wolnelektury.pl/'
+    link = u'http://wolnelektury.pl/'
     description = u"Spis utworów na stronie http://WolneLektury.pl"
     author_name = u"Wolne Lektury"
-    author_link = u"http://www.wolnelektury.pl/"
+    author_link = u"http://wolnelektury.pl/"
 
     def get_object(self, request, category):
         feed = [feed for feed in _root_feeds if feed['category']==category]
@@ -234,7 +235,7 @@ class ByCategoryFeed(Feed):
         return feed['title']
 
     def items(self, feed):
-        return (tag for tag in Tag.objects.filter(category=feed['category']) if tag.get_count() > 0)
+        return Tag.objects.filter(category=feed['category']).exclude(book_count=0)
 
     def item_title(self, item):
         return item.name
@@ -261,7 +262,7 @@ class ByTagFeed(AcquisitionFeed):
 
     def items(self, tag):
         books = Book.tagged.with_any([tag])
-        l_tags = Tag.objects.filter(category='book', slug__in=[book.book_tag_slug() for book in books])
+        l_tags = Tag.objects.filter(category='book', slug__in=[book.book_tag_slug() for book in books.iterator()])
         descendants_keys = [book.pk for book in Book.tagged.with_any(l_tags)]
         if descendants_keys:
             books = books.exclude(pk__in=descendants_keys)
@@ -276,7 +277,7 @@ class UserFeed(Feed):
     link = u'http://www.wolnelektury.pl/'
     description = u"Półki użytkownika na stronie http://WolneLektury.pl"
     author_name = u"Wolne Lektury"
-    author_link = u"http://www.wolnelektury.pl/"
+    author_link = u"http://wolnelektury.pl/"
 
     def get_object(self, request):
         return request.user
@@ -285,7 +286,7 @@ class UserFeed(Feed):
         return u"Półki użytkownika %s" % user.username
 
     def items(self, user):
-        return (tag for tag in Tag.objects.filter(category='set', user=user) if tag.get_count() > 0)
+        return Tag.objects.filter(category='set', user=user).exclude(book_count=0)
 
     def item_title(self, item):
         return item.name
@@ -332,7 +333,11 @@ class SearchFeed(AcquisitionFeed):
     def get_object(self, request):
         """
         For OPDS 1.1 We should handle a query for search terms
-        and atom:author, atom:contributor, atom:title
+        and criteria provided either as opensearch or 'inline' query.
+        OpenSearch defines fields: atom:author, atom:contributor (treated as translator),
+        atom:title. Inline query provides author, title, categories (treated as book tags),
+        description (treated as content search terms).
+        
         if search terms are provided, we shall search for books
         according to Hint information (from author & contributror & title).
 
@@ -343,22 +348,21 @@ class SearchFeed(AcquisitionFeed):
         JVM.attachCurrentThread()
 
         query = request.GET.get('q', '')
-
+        
         inline_criteria = re.findall(self.INLINE_QUERY_RE, query)
         if inline_criteria:
             def get_criteria(criteria, name, position):
                 e = filter(lambda el: el[0][0:len(name)] == name, criteria)
-                print e
+                log.info("get_criteria: %s" % e)
                 if not e:
                     return None
                 c = e[0][position]
-                print c
+                log.info("get_criteria: %s" % c)
                 if c[0] == '"' and c[-1] == '"':
                     c = c[1:-1]
                     c = c.replace('+', ' ')
                 return c
 
-            #import pdb; pdb.set_trace()
             author = get_criteria(inline_criteria, 'author', 1)
             title = get_criteria(inline_criteria, 'title', 2)
             translator = None
@@ -368,11 +372,15 @@ class SearchFeed(AcquisitionFeed):
             author = request.GET.get('author', '')
             title = request.GET.get('title', '')
             translator = request.GET.get('translator', '')
+
+            # Our client didn't handle the opds placeholders
+            if author == '{atom:author}': author = ''       
+            if title == '{atom:title}': title = ''
+            if translator == '{atom:contributor}': translator = ''
             categories = None
             fuzzy = False
 
-
-        srch = MultiSearch()
+        srch = get_search()
         hint = srch.hint()
 
         # Scenario 1: full search terms provided.
@@ -381,11 +389,11 @@ class SearchFeed(AcquisitionFeed):
             filters = []
 
             if author:
-                print "narrow to author %s" % author
-                hint.tags(srch.search_tags(author, filter=srch.term_filter(Term('tag_category', 'author'))))
+                log.info( "narrow to author %s" % author)
+                hint.tags(srch.search_tags(author, filt=srch.term_filter(Term('tag_category', 'author'))))
 
             if translator:
-                print "filter by translator %s" % translator
+                log.info( "filter by translator %s" % translator)
                 filters.append(QueryWrapperFilter(
                     srch.make_phrase(srch.get_tokens(translator, field='translators'),
                                      field='translators')))
@@ -397,18 +405,25 @@ class SearchFeed(AcquisitionFeed):
 
             flt = srch.chain_filters(filters)
             if title:
-                print "hint by book title %s" % title
+                log.info( "hint by book title %s" % title)
                 q = srch.make_phrase(srch.get_tokens(title, field='title'), field='title')
-                hint.books(*srch.search_books(q, filter=flt))
+                hint.books(*srch.search_books(q, filt=flt))
 
             toks = srch.get_tokens(query)
-            print "tokens: %s" % toks
-            #            import pdb; pdb.set_trace()
+            log.info("tokens for query: %s" % toks)
+            
             results = SearchResult.aggregate(srch.search_perfect_book(toks, fuzzy=fuzzy, hint=hint),
                 srch.search_perfect_parts(toks, fuzzy=fuzzy, hint=hint),
                 srch.search_everywhere(toks, fuzzy=fuzzy, hint=hint))
             results.sort(reverse=True)
-            return [r.book for r in results]
+            books = []
+            for r in results:
+                try:
+                    books.append(r.book)
+                except Book.DoesNotExist:
+                    pass
+            log.info("books: %s" % books)
+            return books
         else:
             # Scenario 2: since we no longer have to figure out what the query term means to the user,
             # we can just use filters and not the Hint class.
@@ -426,7 +441,7 @@ class SearchFeed(AcquisitionFeed):
                         srch.make_phrase(srch.get_tokens(q, field=fld), field=fld)))
 
             flt = srch.chain_filters(filters)
-            books = srch.search_books(TermQuery(Term('is_book', 'true')), filter=flt)
+            books = srch.search_books(TermQuery(Term('is_book', 'true')), filt=flt)
             return books
 
     def get_link(self, query):