upgrade

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

diff --git a/src/search/index.py b/src/search/index.py
index e471cd6..2be60fd 100644 (file)
--- a/src/search/index.py
+++ b/src/search/index.py
@@ -95,7 +95,10 @@ class Snippets(object):
         of the snippet stored there.
         """
         self.file.seek(pos[0], 0)
-        txt = self.file.read(pos[1]).decode('utf-8')
+        try:
+            txt = self.file.read(pos[1]).decode('utf-8')
+        except:
+            return ''
         return txt
 
     def close(self):
@@ -636,7 +639,7 @@ class SearchResult(object):
         if self._book is not None:
             return self._book
         try:
-            self._book = catalogue.models.Book.objects.get(id=self.book_id)
+            self._book = catalogue.models.Book.objects.get(id=self.book_id, findable=True)
         except catalogue.models.Book.DoesNotExist:
             self._book = None
         return self._book
@@ -665,20 +668,18 @@ 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 = {}
@@ -905,7 +906,7 @@ class Search(SolrIndex):
 
     def search_by_author(self, words):
         from catalogue.models import Book
-        books = Book.objects.filter(parent=None).order_by('-popularity__count')
+        books = Book.objects.filter(parent=None, findable=True).order_by('-popularity__count')
         for word in words:
             books = books.filter(cached_author__iregex='\m%s\M' % word).select_related('popularity__count')
         return [SearchResult.from_book(book, how_found='search_by_author', query_terms=words) for book in books[:30]]
@@ -979,7 +980,7 @@ class Search(SolrIndex):
                 idx += 1
 
         except IOError as e:
-            book = catalogue.models.Book.objects.filter(id=book_id)
+            book = catalogue.models.Book.objects.filter(id=book_id, findable=True)
             if not book:
                 log.error("Book does not exist for book id = %d" % book_id)
             elif not book.get().children.exists():