X-Git-Url: https://git.mdrn.pl/wolnelektury.git/blobdiff_plain/60b06883b6d5a336ef47c01103ec1ce25aafae69..80d789aa009be1449b83b04b5838ebcc15368d68:/apps/opds/views.py diff --git a/apps/opds/views.py b/apps/opds/views.py index 52335709b..f7658c22b 100644 --- a/apps/opds/views.py +++ b/apps/opds/views.py @@ -2,7 +2,6 @@ # 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 @@ -16,9 +15,16 @@ from django.contrib.sites.models import Site from basicauth import logged_in_or_basicauth, factory_decorator from catalogue.models import Book, Tag -from catalogue.views import books_starting_with -from opds.utils import piwik_track +from search.views import Search, SearchResult +from lucene import Term, QueryWrapperFilter, TermQuery +import operator +import logging +import re + +log = logging.getLogger('opds') + +from stats.utils import piwik_track _root_feeds = ( { @@ -86,7 +92,7 @@ class OPDSFeed(Atom1Feed): {u"href": reverse("opds_authors"), u"rel": u"start", u"type": u"application/atom+xml"}) - handler.addQuickElement(u"link", None, + handler.addQuickElement(u"link", None, {u"href": full_url(os.path.join(settings.STATIC_URL, "opensearch.xml")), u"rel": u"search", u"type": u"application/opensearchdescription+xml"}) @@ -182,19 +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 @@ -211,10 +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] @@ -229,7 +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 @@ -256,7 +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) @@ -271,7 +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 @@ -280,7 +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 @@ -316,20 +322,124 @@ class UserSetFeed(AcquisitionFeed): # no class decorators in python 2.5 #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" - + + QUOTE_OR_NOT = r'(?:(?=["])"([^"]+)"|([^ ]+))' + INLINE_QUERY_RE = re.compile( + r"author:" + QUOTE_OR_NOT + + "|translator:" + QUOTE_OR_NOT + + "|title:" + QUOTE_OR_NOT + + "|categories:" + QUOTE_OR_NOT + + "|description:" + QUOTE_OR_NOT + + "|text:" + QUOTE_OR_NOT + ) + MATCHES = { + 'author': (0, 1), + 'translator': (2, 3), + 'title': (4, 5), + 'categories': (6, 7), + 'description': (8, 9), + 'text': (10, 11), + } + + PARAMS_TO_FIELDS = { + 'author': 'authors', + 'translator': 'translators', + # 'title': 'title', + 'categories': 'tag_name_pl', + 'description': 'text', + # 'text': 'text', + } + + ATOM_PLACEHOLDER = re.compile(r"^{(atom|opds):\w+}$") + 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) + + """ + + query = request.GET.get('q', '') + + inline_criteria = re.findall(self.INLINE_QUERY_RE, query) + if inline_criteria: + remains = re.sub(self.INLINE_QUERY_RE, '', query) + remains = re.sub(r'[ \t]+', ' ', remains) + + def get_criteria(criteria, name): + for c in criteria: + for p in self.MATCHES[name]: + if c[p]: + if p % 2 == 0: + return c[p].replace('+', ' ') + return c[p] + return None + + criteria = dict(map( + lambda cn: (cn, get_criteria(inline_criteria, cn)), + ['author', 'translator', 'title', 'categories', + 'description', 'text'])) + query = remains + # empty query and text set case? + log.debug("Inline query = [%s], criteria: %s" % (query, criteria)) + else: + def remove_dump_data(val): + """Some clients don't get opds placeholders and just send them.""" + if self.ATOM_PLACEHOLDER.match(val): + return '' + return val + + criteria = dict([(cn, remove_dump_data(request.GET.get(cn, ''))) + for cn in self.MATCHES.keys()]) + # query is set above. + log.debug("Inline query = [%s], criteria: %s" % (query, criteria)) + + srch = Search() + + book_hit_filter = srch.index.Q(book_id__any=True) + filters = [book_hit_filter] + [srch.index.Q( + **{self.PARAMS_TO_FIELDS.get(cn, cn): criteria[cn]} + ) for cn in self.MATCHES.keys() if cn in criteria + if criteria[cn]] + + if query: + q = srch.index.query( + reduce(operator.or_, + [srch.index.Q(**{self.PARAMS_TO_FIELDS.get(cn, cn): query}) + for cn in self.MATCHES.keys()], + srch.index.Q())) + else: + q = srch.index.query(srch.index.Q()) + + q = srch.apply_filters(q, filters).field_limit(score=True, fields=['book_id']) + results = q.execute() + + book_scores = dict([(r['book_id'], r['score']) for r in results]) + books = Book.objects.filter(id__in=set([r['book_id'] for r in results])) + books = list(books) + books.sort(reverse=True, key=lambda book: book_scores[book.id]) + return books 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: - return books_starting_with(query) + return books except ValueError: # too short a query return []