Fix section counting/targetting from search snippet

[wolnelektury.git] / apps / opds / views.py

diff --git a/apps/opds/views.py b/apps/opds/views.py
index 7ee09be..dc094bb 100644 (file)
--- a/apps/opds/views.py
+++ b/apps/opds/views.py
@@ -4,6 +4,8 @@
 #
 from base64 import b64encode
 import os.path
 #
 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
 
 from django.contrib.syndication.views import Feed
 from django.core.urlresolvers import reverse


@@ -15,8 +17,13 @@ from django.contrib.sites.models import Site
 
 from basicauth import logged_in_or_basicauth, factory_decorator
 from catalogue.models import Book, Tag
 
 from basicauth import logged_in_or_basicauth, factory_decorator
 from catalogue.models import Book, Tag

-from catalogue.views import books_starting_with

 
 

+from search import Search, SearchResult, JVM
+from lucene import Term, QueryWrapperFilter, TermQuery
+
+import re
+
+from stats.utils import piwik_track

 
 _root_feeds = (
     {
 
 _root_feeds = (
     {


@@ -58,7 +65,7 @@ _root_feeds = (
 
 
 def full_url(url):
 
 
 def full_url(url):

-    return "http://%s%s" % (Site.objects.get_current().domain, url)
+    return urljoin("http://%s" % Site.objects.get_current().domain, url)

 
 
 class OPDSFeed(Atom1Feed):
 
 
 class OPDSFeed(Atom1Feed):


@@ -185,7 +192,7 @@ class AcquisitionFeed(Feed):
     def item_enclosure_length(self, book):
         return book.root_ancestor.epub_file.size
 
     def item_enclosure_length(self, book):
         return book.root_ancestor.epub_file.size
 

-
+@piwik_track

 class RootFeed(Feed):
     feed_type = OPDSFeed
     title = u'Wolne Lektury'
 class RootFeed(Feed):
     feed_type = OPDSFeed
     title = u'Wolne Lektury'


@@ -206,7 +213,7 @@ class RootFeed(Feed):
     def item_description(self, item):
         return item['description']
 
     def item_description(self, item):
         return item['description']
 

-
+@piwik_track

 class ByCategoryFeed(Feed):
     feed_type = OPDSFeed
     link = u'http://www.wolnelektury.pl/'
 class ByCategoryFeed(Feed):
     feed_type = OPDSFeed
     link = u'http://www.wolnelektury.pl/'


@@ -227,7 +234,7 @@ class ByCategoryFeed(Feed):
         return feed['title']
 
     def items(self, 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
 
     def item_title(self, item):
         return item.name


@@ -238,7 +245,7 @@ class ByCategoryFeed(Feed):
     def item_description(self):
         return u''
 
     def item_description(self):
         return u''
 

-
+@piwik_track

 class ByTagFeed(AcquisitionFeed):
     def link(self, tag):
         return tag.get_absolute_url()
 class ByTagFeed(AcquisitionFeed):
     def link(self, tag):
         return tag.get_absolute_url()


@@ -262,7 +269,8 @@ class ByTagFeed(AcquisitionFeed):
         return books
 
 
         return books
 
 

-#@factory_decorator(logged_in_or_basicauth())
+@factory_decorator(logged_in_or_basicauth())
+@piwik_track

 class UserFeed(Feed):
     feed_type = OPDSFeed
     link = u'http://www.wolnelektury.pl/'
 class UserFeed(Feed):
     feed_type = OPDSFeed
     link = u'http://www.wolnelektury.pl/'


@@ -277,7 +285,7 @@ class UserFeed(Feed):
         return u"Półki użytkownika %s" % user.username
 
     def items(self, user):
         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
 
     def item_title(self, item):
         return item.name


@@ -289,10 +297,11 @@ class UserFeed(Feed):
         return u''
 
 # no class decorators in python 2.5
         return u''
 
 # no class decorators in python 2.5

-UserFeed = factory_decorator(logged_in_or_basicauth())(UserFeed)
+#UserFeed = factory_decorator(logged_in_or_basicauth())(UserFeed)

 
 
 
 

-#@factory_decorator(logged_in_or_basicauth())
+@factory_decorator(logged_in_or_basicauth())
+@piwik_track

 class UserSetFeed(AcquisitionFeed):
     def link(self, tag):
         return tag.get_absolute_url()
 class UserSetFeed(AcquisitionFeed):
     def link(self, tag):
         return tag.get_absolute_url()


@@ -310,22 +319,126 @@ class UserSetFeed(AcquisitionFeed):
         return Book.tagged.with_any([tag])
 
 # no class decorators in python 2.5
         return Book.tagged.with_any([tag])
 
 # no class decorators in python 2.5

-UserSetFeed = factory_decorator(logged_in_or_basicauth())(UserSetFeed)
+#UserSetFeed = factory_decorator(logged_in_or_basicauth())(UserSetFeed)

 
 
 
 

+@piwik_track

 class SearchFeed(AcquisitionFeed):
     description = u"Wyniki wyszukiwania na stronie WolneLektury.pl"
     title = u"Wyniki wyszukiwania"
 class SearchFeed(AcquisitionFeed):
     description = u"Wyniki wyszukiwania na stronie WolneLektury.pl"
     title = u"Wyniki wyszukiwania"

+
+    INLINE_QUERY_RE = re.compile(r"(author:(?P<author>[^ ]+)|title:(?P<title>[^ ]+)|categories:(?P<categories>[^ ]+)|description:(?P<description>[^ ]+))")

     
     def get_object(self, request):
     
     def get_object(self, request):

-        return request.GET.get('q', '')
+        """
+        For OPDS 1.1 We should handle a query for search terms
+        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).
+
+        but if search terms are empty, we should do a different search
+        (perhaps for is_book=True)
+
+        """
+        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
+                if not e:
+                    return None
+                c = e[0][position]
+                print 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
+            categories = get_criteria(inline_criteria, 'categories', 3)
+            query = get_criteria(inline_criteria, 'description', 4)
+        else:
+            author = request.GET.get('author', '')
+            title = request.GET.get('title', '')
+            translator = request.GET.get('translator', '')
+            categories = None
+            fuzzy = False
+
+
+        srch = Search()
+        hint = srch.hint()
+
+        # Scenario 1: full search terms provided.
+        # Use auxiliarry information to narrow it and make it better.
+        if query:
+            filters = []
+
+            if author:
+                print "narrow to author %s" % author
+                hint.tags(srch.search_tags(author, filter=srch.term_filter(Term('tag_category', 'author'))))
+
+            if translator:
+                print "filter by translator %s" % translator
+                filters.append(QueryWrapperFilter(
+                    srch.make_phrase(srch.get_tokens(translator, field='translators'),
+                                     field='translators')))
+
+            if categories:
+                filters.append(QueryWrapperFilter(
+                    srch.make_phrase(srch.get_tokens(categories, field="tag_name_pl"),
+                                     field='tag_name_pl')))
+
+            flt = srch.chain_filters(filters)
+            if title:
+                print "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))
+
+            toks = srch.get_tokens(query)
+            print "tokens: %s" % toks
+            #            import pdb; pdb.set_trace()
+            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]
+        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.
+            filters = []
+
+            fields = {
+                'author': author,
+                'translators': translator,
+                'title': title
+                }
+
+            for fld, q in fields.items():
+                if q:
+                    filters.append(QueryWrapperFilter(
+                        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)
+            return books

 
     def get_link(self, query):
 
     def get_link(self, query):

-        return "%s?q=%s" % (reverse('search'), query) 
+        return "%s?q=%s" % (reverse('search'), query)

 
 

-    def items(self, query):
+    def items(self, books):

         try:
         try:

-            return books_starting_with(query)
+            return books

         except ValueError:
             # too short a query
             return []
         except ValueError:
             # too short a query
             return []