Fixes.

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

diff --git a/src/search/index.py b/src/search/index.py
index f943a4d..bd31a2a 100644 (file)
--- a/src/search/index.py
+++ b/src/search/index.py
@@ -1,29 +1,29 @@
-# -*- coding: utf-8 -*-
 # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
 #
-from django.conf import settings
-
+from functools import reduce, total_ordering
+from itertools import chain
+import logging
+import operator
 import os
 import re
+from django.conf import settings
 from librarian import dcparser
 from librarian.parser import WLDocument
 from lxml import etree
+import scorched
 import catalogue.models
 import picture.models
 from pdcounter.models import Author as PDCounterAuthor, BookStub as PDCounterBook
-from itertools import chain
-import scorched
-from . import custom
-import operator
-import logging
 from wolnelektury.utils import makedirs
+from . import custom
 
 log = logging.getLogger('search')
 
+
 if os.path.isfile(settings.SOLR_STOPWORDS):
     stopwords = set(
-        line.decode('utf-8').strip()
+        line.strip()
         for line in open(settings.SOLR_STOPWORDS) if not line.startswith('#'))
 else:
     stopwords = set()
@@ -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
@@ -557,6 +558,7 @@ class Index(SolrIndex):
         self.index.add(doc)
 
 
+@total_ordering
 class SearchResult(object):
     def __init__(self, doc, how_found=None, query_terms=None):
         self.boost = 1.0
@@ -633,7 +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)
@@ -653,12 +658,12 @@ class SearchResult(object):
         # to sections and fragments
         frags = filter(lambda r: r[self.FRAGMENT] is not None, self._hits)
 
-        sect = filter(lambda r: r[self.FRAGMENT] is None, self._hits)
+        sect = [hit for hit in self._hits if hit[self.FRAGMENT] is None]
 
         # sections not covered by fragments
-        sect = filter(lambda s: 0 == len(filter(
+        sect = filter(lambda s: 0 == len(list(filter(
             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)
+                      f[self.POSITION][self.POSITION_INDEX] + f[self.POSITION][self.POSITION_SPAN], frags))), sect)
 
         def remove_duplicates(lst, keyfn, compare):
             els = {}
@@ -691,7 +696,7 @@ class SearchResult(object):
             m.update(s[self.OTHER])
             sections[si] = m
 
-        hits = sections.values()
+        hits = list(sections.values())
 
         for f in frags:
             try:
@@ -712,12 +717,12 @@ class SearchResult(object):
                             break
 
             def theme_by_name(n):
-                th = filter(lambda t: t.name == n, themes)
+                th = list(filter(lambda t: t.name == n, themes))
                 if th:
                     return th[0]
                 else:
                     return None
-            themes_hit = filter(lambda a: a is not None, map(theme_by_name, themes_hit))
+            themes_hit = list(filter(lambda a: a is not None, map(theme_by_name, themes_hit)))
 
             m = {'score': f[self.SCORE],
                  'fragment': frag,
@@ -745,13 +750,17 @@ class SearchResult(object):
                     books[r.book_id] = r
         return books.values()
 
-    def __cmp__(self, other):
-        c = cmp(self.score, other.score)
-        if c == 0:
-            # this is inverted, because earlier date is better
-            return cmp(other.published_date, self.published_date)
-        else:
-            return c
+    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.get_sort_key() > other.get_sort_key()
+
+    def __eq__(self, other):
+        return self.get_sort_key() == other.get_sort_key()
 
     def __len__(self):
         return len(self.hits)
@@ -766,6 +775,7 @@ class SearchResult(object):
             return None
 
 
+@total_ordering
 class PictureResult(object):
     def __init__(self, doc, how_found=None, query_terms=None):
         self.boost = 1.0
@@ -866,8 +876,11 @@ class PictureResult(object):
                     books[r.picture_id] = r
         return books.values()
 
-    def __cmp__(self, other):
-        return cmp(self.score, other.score)
+    def __lt__(self, other):
+        return self.score < other.score
+
+    def __eq__(self, other):
+        return self.score == other.score
 
 
 class Search(SolrIndex):
@@ -975,8 +988,8 @@ class Search(SolrIndex):
         finally:
             snippets.close()
 
-            # remove verse end markers..
-        snips = map(lambda s: s and s.replace("/\n", "\n"), snips)
+        # remove verse end markers..
+        snips = [s.replace("/\n", "\n") if s else s for s in snips]
 
         searchresult.snippets = snips