better treatment to lack of diacritics in search
[wolnelektury.git] / src / picture / models.py
index dc7142c..455ed10 100644 (file)
@@ -12,6 +12,9 @@ from django.core.files.storage import FileSystemStorage
 from django.utils.datastructures import SortedDict
 from fnpdjango.utils.text.slughifi import slughifi
 from ssify import flush_ssi_includes
+
+from catalogue.models.tag import prefetched_relations
+from catalogue.utils import split_tags
 from picture import tasks
 from StringIO import StringIO
 import jsonfield
@@ -112,8 +115,8 @@ class Picture(models.Model):
         self.sort_key = sortify(self.title)[:120]
 
         try:
-            author = self.tags.filter(category='author')[0].sort_key
-        except IndexError:
+            author = self.authors().first().sort_key
+        except AttributeError:
             author = u''
         self.sort_key_author = author
 
@@ -124,15 +127,22 @@ class Picture(models.Model):
     def __unicode__(self):
         return self.title
 
-    def author_str(self):
-        return ", ".join(str(t) for t in self.tags.filter(category='author'))
+    def authors(self):
+        return self.tags.filter(category='author')
 
     def tag_unicode(self, category):
-        return ", ".join(unicode(t) for t in self.tags.filter(category=category))
+        relations = prefetched_relations(self, category)
+        if relations:
+            return ', '.join(rel.tag.name for rel in relations)
+        else:
+            return ', '.join(self.tags.filter(category=category).values_list('name', flat=True))
 
     def author_unicode(self):
         return self.tag_unicode('author')
 
+    def tags_by_category(self):
+        return split_tags(self.tags)
+
     @permalink
     def get_absolute_url(self):
         return 'picture.views.picture_detail', [self.slug]
@@ -156,7 +166,7 @@ class Picture(models.Model):
             return None
 
     @classmethod
-    def from_xml_file(cls, xml_file, image_file=None, image_store=None, overwrite=False):
+    def from_xml_file(cls, xml_file, image_file=None, image_store=None, overwrite=False, search_index=True):
         """
         Import xml and it's accompanying image file.
         If image file is missing, it will be fetched by librarian.picture.ImageStore
@@ -192,8 +202,10 @@ class Picture(models.Model):
             picture.extra_info = picture_xml.picture_info.to_dict()
 
             picture_tags = set(catalogue.models.Tag.tags_from_info(picture_xml.picture_info))
-            motif_tags = set()
-            thing_tags = set()
+            for tag in picture_tags:
+                if not tag.for_pictures:
+                    tag.for_pictures = True
+                    tag.save()
 
             area_data = {'themes': {}, 'things': {}}
 
@@ -207,23 +219,30 @@ class Picture(models.Model):
                 if part.get('object', None) is not None:
                     _tags = set()
                     for objname in part['object'].split(','):
-                        objname = objname.strip().capitalize()
+                        objname = objname.strip()
+                        assert objname, 'Empty object name'
+                        # str.capitalize() is wrong, because it also lowers letters
+                        objname = objname[0].upper() + objname[1:]
                         tag, created = catalogue.models.Tag.objects.get_or_create(
                             slug=slughifi(objname), category='thing')
                         if created:
                             tag.name = objname
                             setattr(tag, 'name_%s' % lang, tag.name)
                             tag.sort_key = sortify(tag.name)
+                            tag.for_pictures = True
                             tag.save()
-                        # thing_tags.add(tag)
                         area_data['things'][tag.slug] = {
                             'object': objname,
                             'coords': part['coords'],
                             }
 
                         _tags.add(tag)
+                        if not tag.for_pictures:
+                            tag.for_pictures = True
+                            tag.save()
                     area = PictureArea.rectangle(picture, 'thing', part['coords'])
                     area.save()
+                    # WTF thing area does not inherit tags from picture and theme area does, is it intentional?
                     area.tags = _tags
                 else:
                     _tags = set()
@@ -234,9 +253,13 @@ class Picture(models.Model):
                             if created:
                                 tag.name = motif
                                 tag.sort_key = sortify(tag.name)
+                                tag.for_pictures = True
                                 tag.save()
                             # motif_tags.add(tag)
                             _tags.add(tag)
+                            if not tag.for_pictures:
+                                tag.for_pictures = True
+                                tag.save()
                             area_data['themes'][tag.slug] = {
                                 'theme': motif,
                                 'coords': part['coords']
@@ -247,7 +270,7 @@ class Picture(models.Model):
                     area.save()
                     area.tags = _tags.union(picture_tags)
 
-            picture.tags = picture_tags.union(motif_tags).union(thing_tags)
+            picture.tags = picture_tags
             picture.areas_json = area_data
 
             if image_file is not None:
@@ -268,6 +291,8 @@ class Picture(models.Model):
             picture.xml_file.save("%s.xml" % picture.slug, File(xml_file))
             picture.save()
             tasks.generate_picture_html(picture.id)
+            if not settings.NO_SEARCH_INDEX and search_index:
+                tasks.index_picture.delay(picture.id, picture_info=picture_xml.picture_info)
 
         if close_xml_file:
             xml_file.close()
@@ -319,7 +344,7 @@ class Picture(models.Model):
             pics_by_author[tag] = []
 
         for pic in pics.iterator():
-            authors = list(pic.tags.filter(category='author'))
+            authors = list(pic.authors().only('pk'))
             if authors:
                 for author in authors:
                     pics_by_author[author].append(pic)
@@ -338,9 +363,7 @@ class Picture(models.Model):
         return self._info
 
     def pretty_title(self, html_links=False):
-        picture = self
-        names = [(tag.name, tag.get_absolute_url())
-                 for tag in self.tags.filter(category='author')]
+        names = [(tag.name, tag.get_absolute_url()) for tag in self.authors().only('name', 'category', 'slug')]
         names.append((self.title, self.get_absolute_url()))
 
         if html_links:
@@ -366,3 +389,17 @@ class Picture(models.Model):
                 ]
             for lang in languages
             ])
+
+    def search_index(self, picture_info=None, index=None, index_tags=True, commit=True):
+        if index is None:
+            from search.index import Index
+            index = Index()
+        try:
+            index.index_picture(self, picture_info)
+            if index_tags:
+                index.index_tags()
+            if commit:
+                index.index.commit()
+        except Exception, e:
+            index.index.rollback()
+            raise e