Fixes #4076: Recommendations from collections.
[wolnelektury.git] / src / search / index.py
index 7120893..2d84cb4 100644 (file)
@@ -142,7 +142,8 @@ class Index(SolrIndex):
                     uids.add(res['uid'])
                 st += rows
         if uids:
-            self.index.delete(uids)
+            # FIXME: With Solr API change, this doesn't work.
+            #self.index.delete(uids)
             return True
         else:
             return False
@@ -389,16 +390,16 @@ class Index(SolrIndex):
             return
 
         def fix_format(text):
-            # separator = [u" ", u"\t", u".", u";", u","]
+            # separator = [" ", "\t", ".", ";", ","]
             if isinstance(text, list):
                 # need to join it first
                 text = filter(lambda s: s is not None, content)
-                text = u' '.join(text)
+                text = ' '.join(text)
                 # for i in range(len(text)):
                 #     if i > 0:
                 #         if text[i][0] not in separator\
                 #             and text[i - 1][-1] not in separator:
-                #          text.insert(i, u" ")
+                #          text.insert(i, " ")
 
             return re.sub("(?m)/$", "", text)
 
@@ -462,7 +463,7 @@ class Index(SolrIndex):
                     elif end is not None and footnote is not [] and end.tag in self.footnote_tags:
                         handle_text.pop()
                         doc = add_part(snippets, header_index=position, header_type=header.tag,
-                                       text=u''.join(footnote),
+                                       text=''.join(footnote),
                                        is_footnote=True)
                         self.index.add(doc)
                         footnote = []
@@ -612,7 +613,7 @@ class SearchResult(object):
         return result
 
     def __str__(self):
-        return u"<SR id=%d %d(%d) hits score=%f %d snippets>" % \
+        return "<SR id=%d %d(%d) hits score=%f %d snippets>" % \
             (self.book_id, len(self._hits),
              len(self._processed_hits) if self._processed_hits else -1,
              self._score, len(self.snippets))
@@ -634,7 +635,10 @@ class SearchResult(object):
     def get_book(self):
         if self._book is not None:
             return self._book
-        self._book = catalogue.models.Book.objects.get(id=self.book_id)
+        try:
+            self._book = catalogue.models.Book.objects.get(id=self.book_id)
+        except catalogue.models.Book.DoesNotExist:
+            self._book = None
         return self._book
 
     book = property(get_book)
@@ -661,20 +665,18 @@ class SearchResult(object):
             lambda f: f[self.POSITION][self.POSITION_INDEX] <= s[self.POSITION][self.POSITION_INDEX] <
                       f[self.POSITION][self.POSITION_INDEX] + f[self.POSITION][self.POSITION_SPAN], frags))), sect)
 
-        def remove_duplicates(lst, keyfn, compare):
+        def remove_duplicates(lst, keyfn, larger):
             els = {}
             for e in lst:
                 eif = keyfn(e)
                 if eif in els:
-                    if compare(els[eif], e) >= 1:
+                    if larger(els[eif], e):
                         continue
                 els[eif] = e
             return els.values()
 
         # remove fragments with duplicated fid's and duplicated snippets
-        frags = remove_duplicates(frags, lambda f: f[self.FRAGMENT], lambda a, b: cmp(a[self.SCORE], b[self.SCORE]))
-        # frags = remove_duplicates(frags, lambda f: f[OTHER]['snippet_pos'] and f[OTHER]['snippet_pos'] or f[FRAGMENT],
-        #                           lambda a, b: cmp(a[SCORE], b[SCORE]))
+        frags = remove_duplicates(frags, lambda f: f[self.FRAGMENT], lambda a, b: a[self.SCORE] > b[self.SCORE])
 
         # remove duplicate sections
         sections = {}
@@ -746,13 +748,17 @@ class SearchResult(object):
                     books[r.book_id] = r
         return books.values()
 
+    def get_sort_key(self):
+        return (-self.score,
+                self.published_date,
+                self.book.sort_key_author if self.book else '',
+                self.book.sort_key if self.book else '')
+
     def __lt__(self, other):
-        return (-self.score, self.published_date, self.book.sort_key_author, self.book.sort_key) > \
-               (-other.score, other.published_date, other.book.sort_key_author, other.book.sort_key)
+        return self.get_sort_key() > other.get_sort_key()
 
     def __eq__(self, other):
-        return (self.score, self.published_date, self.book.sort_key_author, self.book.sort_key) == \
-               (other.score, other.published_date, other.book.sort_key_author, other.book.sort_key)
+        return self.get_sort_key() == other.get_sort_key()
 
     def __len__(self):
         return len(self.hits)
@@ -794,7 +800,7 @@ class PictureResult(object):
             self._hits.append(hit)
 
     def __str__(self):
-        return u"<PR id=%d score=%f >" % (self.picture_id, self._score)
+        return "<PR id=%d score=%f >" % (self.picture_id, self._score)
 
     def __repr__(self):
         return str(self)