OPDS advanced search

[wolnelektury.git] / apps / search / index.py

diff --git a/apps/search/index.py b/apps/search/index.py
index 9b7efa2..8b0cfb7 100644 (file)
--- a/apps/search/index.py
+++ b/apps/search/index.py
@@ -55,6 +55,7 @@ class WLAnalyzer(PerFieldAnalyzerWrapper):
         self.addAnalyzer("publisher", simple)
         self.addAnalyzer("author", simple)
         self.addAnalyzer("is_book", keyword)
+        # shouldn't the title have two forms? _pl and simple?
 
         self.addAnalyzer("themes", simple)
         self.addAnalyzer("themes_pl", polish)
@@ -62,6 +63,8 @@ class WLAnalyzer(PerFieldAnalyzerWrapper):
         self.addAnalyzer("tag_name", simple)
         self.addAnalyzer("tag_name_pl", polish)
 
+        self.addAnalyzer("translators", simple)
+
         self.addAnalyzer("KEYWORD", keyword)
         self.addAnalyzer("SIMPLE", simple)
         self.addAnalyzer("POLISH", polish)
@@ -117,13 +120,11 @@ class Snippets(object):
         self.file.write(txt)
         pos = (self.position, l)
         self.position += l
-        print "Snip<%s>%s</s>" %(pos, txt)
         return pos
 
     def get(self, pos):
         self.file.seek(pos[0], 0)
         txt = self.file.read(pos[1]).decode('utf-8')
-        print "got from snippets %d bytes from %s:" % (len(txt), pos)
         return txt
 
     def close(self):
@@ -214,7 +215,7 @@ class Index(IndexStore):
 
     def extract_metadata(self, book):
         fields = {}
-        book_info = dcparser.parse(book.xml_file)
+        book_info = dcparser.parse(open(book.xml_file.path))
 
         print("extract metadata for book %s id=%d, thread%d" % (book.slug, book.id, current_thread().ident))
 
@@ -317,6 +318,13 @@ class Index(IndexStore):
 
             return doc
 
+        def give_me_utf8(s):
+            if isinstance(s, unicode):
+                return s.encode('utf-8')
+            else:
+                return s
+
+
         fragments = {}
         snippets = Snippets(book.id).open('w')
         try:
@@ -339,13 +347,16 @@ class Index(IndexStore):
                         fragments[fid]['content'].append(start.tail)
                     elif start is not None and start.tag == 'motyw':
                         fid = start.attrib['id'][1:]
-                        fragments[fid]['themes'].append(start.text)
+                        if start.text is not None:
+                            fragments[fid]['themes'] += map(str.strip, map(give_me_utf8, start.text.split(',')))
                         fragments[fid]['content'].append(start.tail)
                     elif start is not None and start.tag == 'end':
                         fid = start.attrib['id'][1:]
                         if fid not in fragments:
                             continue  # a broken <end> node, skip it
                         frag = fragments[fid]
+                        if frag['themes'] == []:
+                            continue  # empty themes list.
                         del fragments[fid]
 
                         def jstr(l):
@@ -486,6 +497,7 @@ class Search(IndexStore):
             bks.append(catalogue.models.Book.objects.get(id=doc.get("book_id")))
         return (bks, tops.totalHits)
 
+
     def search(self, query, max_results=50):
         query = self.query(query)
         query = self.wrapjoins(query, ["content", "themes"])
@@ -615,10 +627,10 @@ class Hint(object):
         self.search = search
         self.book_tags = {}
         self.part_tags = []
-        self._book = None
+        self._books = []
 
-    def book(self, book):
-        self._book = book
+    def books(self, *books):
+        self._books = books
 
     def tags(self, tags):
         for t in tags:
@@ -652,12 +664,18 @@ class Hint(object):
         fs = []
         if self.part_tags:
             fs.append(self.tag_filter(self.part_tags, field='themes'))
-        if self._book is not None:
-            fs.append(NumericRangeFilter.newIntRange('book_id', self._book.id, self._book.id, True, True))
+            
+        if self._books != []:
+            bf = BooleanFilter()
+            for b in self._books:
+                id_filter = NumericRangeFilter.newIntRange('book_id', b.id, b.id, True, True)
+                bf.add(FilterClause(id_filter, BooleanClause.Occur.SHOULD))
+            fs.append(bf)
+            
         return MultiSearch.chain_filters(fs)
             
     def should_search_for_book(self):
-        return self._book is None
+        return self._books == []
 
     def just_search_in(self, all):
         """Holds logic to figure out which indexes should be search, when we have some hinst already"""
@@ -665,7 +683,7 @@ class Hint(object):
         for field in all:
             if field == 'author' and 'author' in self.book_tags:
                 continue
-            if field == 'title' and self._book is not None:
+            if field == 'title' and self._books != []:
                 continue
             if (field == 'themes' or field == 'themes_pl') and self.part_tags:
                 continue
@@ -738,9 +756,9 @@ class MultiSearch(Search):
             q.add(BooleanClause(term, modal))
         return q
 
-    def content_query(self, query):
-        return BlockJoinQuery(query, self.parent_filter,
-                              BlockJoinQuery.ScoreMode.Total)
+    # def content_query(self, query):
+    #     return BlockJoinQuery(query, self.parent_filter,
+    #                           BlockJoinQuery.ScoreMode.Total)
 
     def search_perfect_book(self, searched, max_results=20, fuzzy=False, hint=None):
         fields_to_search = ['author', 'title']
@@ -890,7 +908,6 @@ class MultiSearch(Search):
         #  highlighter.getBestTextFragments(tokenStream, text, False, 10)
         #        import pdb; pdb.set_trace()
         snip = highlighter.getBestFragments(tokenStream, text, 3, "...")
-        print('snips: %s' % snip)
 
         return [snip]
 
@@ -923,6 +940,14 @@ class MultiSearch(Search):
 
         return tags
 
+    def search_books(self, query, filter=None, max_results=10):
+        bks = []
+        tops = self.searcher.search(query, filter, max_results)
+        for found in tops.scoreDocs:
+            doc = self.searcher.doc(found.doc)
+            bks.append(catalogue.models.Book.objects.get(id=doc.get("book_id")))
+        return bks
+
     def create_prefix_phrase(self, toks, field):
         q = MultiPhraseQuery()
         for i in range(len(toks)):