BE in footer
[wolnelektury.git] / apps / catalogue / models.py
index 7e3ccfe..568679b 100644 (file)
@@ -2,13 +2,14 @@
 # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
 #
+from datetime import datetime
+
 from django.db import models
 from django.db.models import permalink, Q
 from django.utils.translation import ugettext_lazy as _
 from django.contrib.auth.models import User
 from django.core.files import File
 from django.template.loader import render_to_string
-from django.template.defaultfilters import slugify
 from django.utils.safestring import mark_safe
 from django.utils.translation import get_language
 from django.core.urlresolvers import reverse
@@ -18,10 +19,14 @@ from django.conf import settings
 
 from newtagging.models import TagBase, tags_updated
 from newtagging import managers
-from catalogue.fields import JSONField
+from catalogue.fields import JSONField, OverwritingFileField
+from catalogue.utils import ExistingFile
 
 from librarian import dcparser, html, epub, NoDublinCore
+import mutagen
 from mutagen import id3
+from slughifi import slughifi
+from sortify import sortify
 
 
 TAG_CATEGORIES = (
@@ -64,6 +69,12 @@ class Tag(TagBase):
     gazeta_link = models.CharField(blank=True, max_length=240)
     wiki_link = models.CharField(blank=True, max_length=240)
 
+    created_at    = models.DateTimeField(_('creation date'), auto_now_add=True, db_index=True)
+    changed_at    = models.DateTimeField(_('creation date'), auto_now=True, db_index=True)
+
+    class UrlDeprecationWarning(DeprecationWarning):
+        pass
+
     categories_rev = {
         'autor': 'author',
         'epoka': 'epoch',
@@ -122,19 +133,20 @@ class Tag(TagBase):
             real_tags = []
             ambiguous_slugs = []
             category = None
+            deprecated = False
             tags_splitted = tags.split('/')
-            for index, name in enumerate(tags_splitted):
-                if name in Tag.categories_rev:
+            for name in tags_splitted:
+                if category:
+                    real_tags.append(Tag.objects.get(slug=name, category=category))
+                    category = None
+                elif name in Tag.categories_rev:
                     category = Tag.categories_rev[name]
                 else:
-                    if category:
-                        real_tags.append(Tag.objects.get(slug=name, category=category))
-                        category = None
-                    else:
-                        try:
-                            real_tags.append(Tag.objects.exclude(category='book').get(slug=name))
-                        except Tag.MultipleObjectsReturned, e:
-                            ambiguous_slugs.append(name)
+                    try:
+                        real_tags.append(Tag.objects.exclude(category='book').get(slug=name))
+                        deprecated = True 
+                    except Tag.MultipleObjectsReturned, e:
+                        ambiguous_slugs.append(name)
 
             if category:
                 # something strange left off
@@ -145,8 +157,11 @@ class Tag(TagBase):
                 e.tags = real_tags
                 e.ambiguous_slugs = ambiguous_slugs
                 raise e
-            else:
-                return real_tags
+            if deprecated:
+                e = Tag.UrlDeprecationWarning()
+                e.tags = real_tags
+                raise e
+            return real_tags
         else:
             return TagBase.get_tag_list(tags)
 
@@ -156,25 +171,30 @@ class Tag(TagBase):
 
 
 # TODO: why is this hard-coded ?
-def book_upload_path(ext=None):
+def book_upload_path(ext=None, maxlen=100):
     def get_dynamic_path(media, filename, ext=ext):
         # how to put related book's slug here?
         if not ext:
-            ext = media.type
+            if media.type == 'daisy':
+                ext = 'daisy.zip'
+            else:
+                ext = media.type
         if not media.name:
-            name = slugify(filename.split(".")[0])
+            name = slughifi(filename.split(".")[0])
         else:
-            name = slugify(media.name)
-        return 'lektura/%s.%s' % (name, ext)
+            name = slughifi(media.name)
+        return 'book/%s/%s.%s' % (ext, name[:maxlen-len('book/%s/.%s' % (ext, ext))-4], ext)
     return get_dynamic_path
 
 
 class BookMedia(models.Model):
     type        = models.CharField(_('type'), choices=MEDIA_FORMATS, max_length="100")
-    name        = models.CharField(_('name'), max_length="100", blank=True)
-    file        = models.FileField(_('file'), upload_to=book_upload_path(), blank=True)
+    name        = models.CharField(_('name'), max_length="100")
+    file        = OverwritingFileField(_('file'), upload_to=book_upload_path())
     uploaded_at = models.DateTimeField(_('creation date'), auto_now_add=True, editable=False)
-    extra_info  = JSONField(_('extra information'), default='{}')
+    extra_info  = JSONField(_('extra information'), default='{}', editable=False)
+    book = models.ForeignKey('Book', related_name='media')
+    source_sha1 = models.CharField(null=True, blank=True, max_length=40, editable=False)
 
     def __unicode__(self):
         return "%s (%s)" % (self.name, self.file.name.split("/")[-1])
@@ -184,35 +204,87 @@ class BookMedia(models.Model):
         verbose_name        = _('book media')
         verbose_name_plural = _('book media')
 
-    def save(self, force_insert=False, force_update=False):
-        media = super(BookMedia, self).save(force_insert, force_update)
-        if self.type == 'mp3':
-            file = self.file
-            extra_info = self.get_extra_info_value()
-            extra_info.update(self.get_mp3_info())
-            self.set_extra_info_value(extra_info)
-            media = super(BookMedia, self).save(force_insert, force_update)
-        return media
-
-    def get_mp3_info(self):
-        """Retrieves artist and director names from audio ID3 tags."""
+    def save(self, *args, **kwargs):
         try:
-            audio = id3.ID3(self.file.path)
-            artist_name = ', '.join(', '.join(tag.text) for tag in audio.getall('TPE1'))
-            director_name = ', '.join(', '.join(tag.text) for tag in audio.getall('TPE3'))
-        except:
-            artist_name = director_name = ''
-        return {'artist_name': artist_name, 'director_name': director_name}
+            old = BookMedia.objects.get(pk=self.pk)
+        except BookMedia.DoesNotExist, e:
+            pass
+        else:
+            # if name changed, change the file name, too
+            if slughifi(self.name) != slughifi(old.name):
+                self.file.save(None, ExistingFile(self.file.path), save=False, leave=True)
+
+        super(BookMedia, self).save(*args, **kwargs)
+        extra_info = self.get_extra_info_value()
+        extra_info.update(self.read_meta())
+        self.set_extra_info_value(extra_info)
+        self.source_sha1 = self.read_source_sha1(self.file.path, self.type)
+        return super(BookMedia, self).save(*args, **kwargs)
+
+    def read_meta(self):
+        """
+            Reads some metadata from the audiobook.
+        """
+
+        artist_name = director_name = project = funded_by = ''
+        if self.type == 'mp3':
+            try:
+                audio = id3.ID3(self.file.path)
+                artist_name = ', '.join(', '.join(tag.text) for tag in audio.getall('TPE1'))
+                director_name = ', '.join(', '.join(tag.text) for tag in audio.getall('TPE3'))
+                project = ", ".join([t.data for t in audio.getall('PRIV') 
+                        if t.owner=='wolnelektury.pl?project'])
+                funded_by = ", ".join([t.data for t in audio.getall('PRIV') 
+                        if t.owner=='wolnelektury.pl?funded_by'])
+            except:
+                pass
+        elif self.type == 'ogg':
+            try:
+                audio = mutagen.File(self.file.path)
+                artist_name = ', '.join(audio.get('artist', []))
+                director_name = ', '.join(audio.get('conductor', []))
+                project = ", ".join(audio.get('project', []))
+                funded_by = ", ".join(audio.get('funded_by', []))
+            except:
+                pass
+        else:
+            return {}
+        return {'artist_name': artist_name, 'director_name': director_name,
+                'project': project, 'funded_by': funded_by}
+
+    @staticmethod
+    def read_source_sha1(filepath, filetype):
+        """
+            Reads source file SHA1 from audiobok metadata.
+        """
+
+        if filetype == 'mp3':
+            try:
+                audio = id3.ID3(filepath)
+                return [t.data for t in audio.getall('PRIV') 
+                        if t.owner=='wolnelektury.pl?flac_sha1'][0]
+            except:
+                return None
+        elif filetype == 'ogg':
+            try:
+                audio = mutagen.File(filepath)
+                return audio.get('flac_sha1', [None])[0] 
+            except:
+                return None
+        else:
+            return None
 
 
 class Book(models.Model):
     title         = models.CharField(_('title'), max_length=120)
+    sort_key = models.CharField(_('sort_key'), max_length=120, db_index=True, editable=False)
     slug          = models.SlugField(_('slug'), max_length=120, unique=True, db_index=True)
     description   = models.TextField(_('description'), blank=True)
-    created_at    = models.DateTimeField(_('creation date'), auto_now_add=True)
+    created_at    = models.DateTimeField(_('creation date'), auto_now_add=True, db_index=True)
+    changed_at    = models.DateTimeField(_('creation date'), auto_now=True, db_index=True)
     _short_html   = models.TextField(_('short HTML'), editable=False)
     parent_number = models.IntegerField(_('parent number'), default=0)
-    extra_info    = JSONField(_('extra information'))
+    extra_info    = JSONField(_('extra information'), default='{}')
     gazeta_link   = models.CharField(blank=True, max_length=240)
     wiki_link     = models.CharField(blank=True, max_length=240)
     # files generated during publication
@@ -221,9 +293,7 @@ class Book(models.Model):
     pdf_file      = models.FileField(_('PDF file'), upload_to=book_upload_path('pdf'), blank=True)
     epub_file     = models.FileField(_('EPUB file'), upload_to=book_upload_path('epub'), blank=True)    
     txt_file      = models.FileField(_('TXT file'), upload_to=book_upload_path('txt'), blank=True)        
-    # other files
-    medias        = models.ManyToManyField(BookMedia, blank=True)
-    
+
     parent        = models.ForeignKey('self', blank=True, null=True, related_name='children')
     objects  = models.Manager()
     tagged   = managers.ModelTaggedItemManager(Tag)
@@ -236,7 +306,7 @@ class Book(models.Model):
         pass
 
     class Meta:
-        ordering = ('title',)
+        ordering = ('sort_key',)
         verbose_name = _('book')
         verbose_name_plural = _('books')
 
@@ -244,6 +314,8 @@ class Book(models.Model):
         return self.title
 
     def save(self, force_insert=False, force_update=False, reset_short_html=True, **kwargs):
+        self.sort_key = sortify(self.title)
+
         if reset_short_html:
             # Reset _short_html during save
             update = {}
@@ -302,7 +374,7 @@ class Book(models.Model):
             else:
                 return False                          
         else:
-            if self.medias.filter(book=self, type=type).count() > 0:
+            if self.media.filter(type=type).exists():
                 return True
             else:
                 return False
@@ -320,7 +392,7 @@ class Book(models.Model):
             elif type == "pdf":
                 return self.pdf_file
             else:                                             
-                return self.medias.filter(book=self, type=type)
+                return self.media.filter(type=type)
         else:
             return None
 
@@ -354,7 +426,7 @@ class Book(models.Model):
             if self.has_media("txt"):
                 formats.append(u'<a href="%s">TXT</a>' % self.get_media('txt').url)
             # other files
-            for m in self.medias.order_by('type'):
+            for m in self.media.order_by('type'):
                 formats.append(u'<a href="%s">%s</a>' % (m.file.url, m.type.upper()))
 
             formats = [mark_safe(format) for format in formats]
@@ -453,8 +525,7 @@ class Book(models.Model):
         epub_file = StringIO()
         try:
             epub.transform(BookImportDocProvider(self), self.slug, output_file=epub_file)
-            self.epub_file.save('%s.epub' % self.slug, ContentFile(epub_file.getvalue()), save=False)
-            self.save()
+            self.epub_file.save('%s.epub' % self.slug, ContentFile(epub_file.getvalue()))
             FileRecord(slug=self.slug, type='epub', sha1=sha1(epub_file.getvalue()).hexdigest()).save()
         except NoDublinCore:
             pass
@@ -468,9 +539,73 @@ class Book(models.Model):
             child_book.save()
             book_descendants += list(child_book.children.all())
 
+    def build_txt(self):
+        from StringIO import StringIO
+        from django.core.files.base import ContentFile
+        from librarian import text
+
+        out = StringIO()
+        text.transform(open(self.xml_file.path), out)
+        self.txt_file.save('%s.txt' % self.slug, ContentFile(out.getvalue()))
+
+
+    def build_html(self):
+        from tempfile import NamedTemporaryFile
+        from markupstring import MarkupString
+
+        meta_tags = list(self.tags.filter(
+            category__in=('author', 'epoch', 'genre', 'kind')))
+        book_tag = self.book_tag()
+
+        html_file = NamedTemporaryFile()
+        if html.transform(self.xml_file.path, html_file, parse_dublincore=False):
+            self.html_file.save('%s.html' % self.slug, File(html_file))
+
+            # get ancestor l-tags for adding to new fragments
+            ancestor_tags = []
+            p = self.parent
+            while p:
+                ancestor_tags.append(p.book_tag())
+                p = p.parent
+
+            # Delete old fragments and create them from scratch
+            self.fragments.all().delete()
+            # Extract fragments
+            closed_fragments, open_fragments = html.extract_fragments(self.html_file.path)
+            for fragment in closed_fragments.values():
+                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 = theme_name.lower()
+                        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 = Fragment.objects.create(anchor=fragment.id, book=self,
+                    text=text, short_text=short_text)
+
+                new_fragment.save()
+                new_fragment.tags = set(meta_tags + themes + [book_tag] + ancestor_tags)
+            self.save()
+            return True
+        return False
+
 
     @classmethod
-    def from_xml_file(cls, xml_file, overwrite=False, build_epub=True):
+    def from_xml_file(cls, xml_file, **kwargs):
         # use librarian to parse meta-data
         book_info = dcparser.parse(xml_file)
 
@@ -478,16 +613,13 @@ class Book(models.Model):
             xml_file = File(open(xml_file))
 
         try:
-            return cls.from_text_and_meta(xml_file, book_info, overwrite, build_epub=build_epub)
+            return cls.from_text_and_meta(xml_file, book_info, **kwargs)
         finally:
             xml_file.close()
 
     @classmethod
-    def from_text_and_meta(cls, raw_file, book_info, overwrite=False, build_epub=True):
-        from tempfile import NamedTemporaryFile
-        from slughifi import slughifi
-        from markupstring import MarkupString
-        from django.core.files.storage import default_storage
+    def from_text_and_meta(cls, raw_file, book_info, overwrite=False, build_epub=True, build_txt=True):
+        import re
 
         # check for parts before we do anything
         children = []
@@ -502,6 +634,8 @@ class Book(models.Model):
 
         # Read book metadata
         book_base, book_slug = book_info.url.rsplit('/', 1)
+        if re.search(r'[^a-zA-Z0-9-]', book_slug):
+            raise ValueError('Invalid characters in slug')
         book, created = Book.objects.get_or_create(slug=book_slug)
 
         if created:
@@ -517,7 +651,7 @@ class Book(models.Model):
         book._short_html = ''
         book.save()
 
-        book_tags = []
+        meta_tags = []
         categories = (('kinds', 'kind'), ('genres', 'genre'), ('authors', 'author'), ('epochs', 'epoch'))
         for field_name, category in categories:
             try:
@@ -532,11 +666,11 @@ class Book(models.Model):
                 tag, created = Tag.objects.get_or_create(slug=slughifi(tag_name), category=category)
                 if created:
                     tag.name = tag_name
-                    tag.sort_key = tag_sort_key.lower()
+                    tag.sort_key = sortify(tag_sort_key.lower())
                     tag.save()
-                book_tags.append(tag)
+                meta_tags.append(tag)
 
-        book.tags = book_tags + book_shelves
+        book.tags = set(meta_tags + book_shelves)
 
         book_tag = book.book_tag()
 
@@ -551,46 +685,9 @@ class Book(models.Model):
         # delete old fragments when overwriting
         book.fragments.all().delete()
 
-        html_file = NamedTemporaryFile()
-        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)
-
-            # get ancestor l-tags for adding to new fragments
-            ancestor_tags = []
-            p = book.parent
-            while p:
-                ancestor_tags.append(p.book_tag())
-                p = p.parent
-
-            # Extract fragments
-            closed_fragments, open_fragments = html.extract_fragments(book.html_file.path)
-            for fragment in closed_fragments.values():
-                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 = theme_name.lower()
-                        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] + ancestor_tags)
+        if book.build_html():
+            if not settings.NO_BUILD_TXT and build_txt:
+                book.build_txt()
 
         if not settings.NO_BUILD_EPUB and build_epub:
             book.root_ancestor.build_epub()
@@ -675,6 +772,24 @@ class Book(models.Model):
 
         return ', '.join(names)
 
+    @classmethod
+    def tagged_top_level(cls, tags):
+        """ Returns top-level books tagged with `tags'.
+
+        It only returns those books which don't have ancestors which are
+        also tagged with those tags.
+
+        """
+        # get relevant books and their tags
+        objects = cls.tagged.with_all(tags)
+        # eliminate descendants
+        l_tags = Tag.objects.filter(category='book', slug__in=[book.book_tag_slug() for book in objects])
+        descendants_keys = [book.pk for book in cls.tagged.with_any(l_tags)]
+        if descendants_keys:
+            objects = objects.exclude(pk__in=descendants_keys)
+
+        return objects
+
 
 class Fragment(models.Model):
     text = models.TextField()
@@ -730,7 +845,8 @@ class FileRecord(models.Model):
 
 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)
+    # we want Tag.changed_at updated for API to know the tag was touched
+    Tag.objects.filter(pk__in=[tag.pk for tag in affected_tags]).update(book_count=None, changed_at=datetime.now())
 
     # if book tags changed, reset book tag counter
     if isinstance(sender, Book) and \
@@ -745,23 +861,14 @@ def _tags_updated_handler(sender, affected_tags, **kwargs):
 tags_updated.connect(_tags_updated_handler)
 
 
-def _m2m_changed_handler(sender, instance, action, reverse, pk_set, **kwargs):
-    """ refresh all the short_html stuff on BookMedia delete """
-    if sender == Book.medias.through and reverse and action == 'pre_clear':
-        for book in instance.book_set.all():
-            book.save()
-m2m_changed.connect(_m2m_changed_handler)
-
 def _pre_delete_handler(sender, instance, **kwargs):
-    """ explicitly clear m2m, so that Books can be refreshed """
+    """ refresh Book on BookMedia delete """
     if sender == BookMedia:
-        instance.book_set.clear()
+        instance.book.save()
 pre_delete.connect(_pre_delete_handler)
 
 def _post_save_handler(sender, instance, **kwargs):
     """ refresh all the short_html stuff on BookMedia update """
     if sender == BookMedia:
-        for book in instance.book_set.all():
-            book.save()
+        instance.book.save()
 post_save.connect(_post_save_handler)
-