Fixes #754, differentiate neighboring theme blocks

[wolnelektury.git] / apps / catalogue / models.py

diff --git a/apps/catalogue/models.py b/apps/catalogue/models.py
index a7e04f1..dd251c1 100644 (file)
--- a/apps/catalogue/models.py
+++ b/apps/catalogue/models.py
@@ -17,7 +17,7 @@ from newtagging.models import TagBase, tags_updated
 from newtagging import managers
 from catalogue.fields import JSONField
 
-from librarian import html, dcparser
+from librarian import dcparser, html, epub, NoDublinCore
 from mutagen import id3
 
 
@@ -332,6 +332,10 @@ class Book(models.Model):
         from tempfile import NamedTemporaryFile
         from slughifi import slughifi
         from markupstring import MarkupString
+        from hashlib import sha1
+        from django.core.files.base import ContentFile
+        from django.core.files.storage import default_storage
+        from StringIO import StringIO
 
         # Read book metadata
         book_base, book_slug = book_info.url.rsplit('/', 1)
@@ -400,28 +404,42 @@ class Book(models.Model):
         if html.transform(book.xml_file.path, html_file, parse_dublincore=False):
             book.html_file.save('%s.html' % book.slug, File(html_file), save=False)
 
+            # Create EPUB
+            epub_file = StringIO()
+            try:
+                epub.transform(book.xml_file, epub_file)
+                book.epub_file.save('%s.epub' % book.slug, ContentFile(epub_file.getvalue()), save=False)
+                FileRecord(slug=book.slug, type='epub', sha1=sha1(epub_file.getvalue()).hexdigest()).save()
+            except NoDublinCore:
+                pass
+
             # Extract fragments
             closed_fragments, open_fragments = html.extract_fragments(book.html_file.path)
             for fragment in closed_fragments.values():
-                text = fragment.to_string()
-                short_text = ''
-                if (len(MarkupString(text)) > 240):
-                    short_text = unicode(MarkupString(text)[:160])
-                new_fragment, created = Fragment.objects.get_or_create(anchor=fragment.id, book=book,
-                    defaults={'text': text, 'short_text': short_text})
-
                 try:
                     theme_names = [s.strip() for s in fragment.themes.split(',')]
                 except AttributeError:
                     continue
                 themes = []
                 for theme_name in theme_names:
+                    if not theme_name:
+                        continue
                     tag, created = Tag.objects.get_or_create(slug=slughifi(theme_name), category='theme')
                     if created:
                         tag.name = theme_name
                         tag.sort_key = slughifi(theme_name)
                         tag.save()
                     themes.append(tag)
+                if not themes:
+                    continue
+
+                text = fragment.to_string()
+                short_text = ''
+                if (len(MarkupString(text)) > 240):
+                    short_text = unicode(MarkupString(text)[:160])
+                new_fragment, created = Fragment.objects.get_or_create(anchor=fragment.id, book=book,
+                    defaults={'text': text, 'short_text': short_text})
+
                 new_fragment.save()
                 new_fragment.tags = set(book_tags + themes + [book_tag])
 
@@ -538,6 +556,21 @@ class BookStub(models.Model):
         return self.title
 
 
+class FileRecord(models.Model):
+    slug = models.SlugField(_('slug'), max_length=120, db_index=True)
+    type = models.CharField(_('type'), max_length=20, db_index=True)
+    sha1 = models.CharField(_('sha-1 hash'), max_length=40)
+    time = models.DateTimeField(_('time'), auto_now_add=True)
+
+    class Meta:
+        ordering = ('-time','-slug', '-type')
+        verbose_name = _('file record')
+        verbose_name_plural = _('file records')
+
+    def __unicode__(self):
+        return "%s %s.%s" % (self.sha1,  self.slug, self.type)
+
+
 def _tags_updated_handler(sender, affected_tags, **kwargs):
     # reset tag global counter
     Tag.objects.filter(pk__in=[tag.pk for tag in affected_tags]).update(book_count=None)