new librarian sub-module: custom pdfs
[wolnelektury.git] / apps / catalogue / models.py
index 2008f09..1a2e8f8 100644 (file)
@@ -2,8 +2,12 @@
 # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
 #
 # 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.db import models
 from django.db.models import permalink, Q
+import django.dispatch
+from django.core.cache import cache
 from django.utils.translation import ugettext_lazy as _
 from django.contrib.auth.models import User
 from django.core.files import File
 from django.utils.translation import ugettext_lazy as _
 from django.contrib.auth.models import User
 from django.core.files import File
@@ -11,15 +15,23 @@ from django.template.loader import render_to_string
 from django.utils.safestring import mark_safe
 from django.utils.translation import get_language
 from django.core.urlresolvers import reverse
 from django.utils.safestring import mark_safe
 from django.utils.translation import get_language
 from django.core.urlresolvers import reverse
-from datetime import datetime
+from django.db.models.signals import post_save, m2m_changed, pre_delete
+
+from django.conf import settings
 
 
-from newtagging.models import TagBase
+from newtagging.models import TagBase, tags_updated
 from newtagging import managers
 from newtagging import managers
-from catalogue.fields import JSONField
+from catalogue.fields import JSONField, OverwritingFileField
+from catalogue.utils import ExistingFile, ORMDocProvider, create_zip, remove_zip
 
 
-from librarian import html, dcparser
+from librarian import dcparser, html, epub, NoDublinCore
+import mutagen
 from mutagen import id3
 from mutagen import id3
+from slughifi import slughifi
+from sortify import sortify
+from os import unlink
 
 
+import search
 
 TAG_CATEGORIES = (
     ('author', _('author')),
 
 TAG_CATEGORIES = (
     ('author', _('author')),
@@ -31,6 +43,16 @@ TAG_CATEGORIES = (
     ('book', _('book')),
 )
 
     ('book', _('book')),
 )
 
+MEDIA_FORMATS = (
+    ('odt', _('ODT file')),
+    ('mp3', _('MP3 file')),
+    ('ogg', _('OGG file')),
+    ('daisy', _('DAISY file')), 
+)
+
+# not quite, but Django wants you to set a timeout
+CACHE_FOREVER = 2419200  # 28 days
+
 
 class TagSubcategoryManager(models.Manager):
     def __init__(self, subcategory):
 
 class TagSubcategoryManager(models.Manager):
     def __init__(self, subcategory):
@@ -44,18 +66,23 @@ class TagSubcategoryManager(models.Manager):
 class Tag(TagBase):
     name = models.CharField(_('name'), max_length=50, db_index=True)
     slug = models.SlugField(_('slug'), max_length=120, db_index=True)
 class Tag(TagBase):
     name = models.CharField(_('name'), max_length=50, db_index=True)
     slug = models.SlugField(_('slug'), max_length=120, db_index=True)
-    sort_key = models.SlugField(_('sort key'), max_length=120, db_index=True)
+    sort_key = models.CharField(_('sort key'), max_length=120, db_index=True)
     category = models.CharField(_('category'), max_length=50, blank=False, null=False,
         db_index=True, choices=TAG_CATEGORIES)
     description = models.TextField(_('description'), blank=True)
     main_page = models.BooleanField(_('main page'), default=False, db_index=True, help_text=_('Show tag on main page'))
 
     user = models.ForeignKey(User, blank=True, null=True)
     category = models.CharField(_('category'), max_length=50, blank=False, null=False,
         db_index=True, choices=TAG_CATEGORIES)
     description = models.TextField(_('description'), blank=True)
     main_page = models.BooleanField(_('main page'), default=False, db_index=True, help_text=_('Show tag on main page'))
 
     user = models.ForeignKey(User, blank=True, null=True)
-    book_count = models.IntegerField(_('book count'), default=0, blank=False, null=False)
-    death = models.IntegerField(_(u'year of death'), blank=True, null=True)
+    book_count = models.IntegerField(_('book count'), blank=True, null=True)
     gazeta_link = models.CharField(blank=True, max_length=240)
     wiki_link = models.CharField(blank=True, max_length=240)
     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',
     categories_rev = {
         'autor': 'author',
         'epoka': 'epoch',
@@ -87,16 +114,26 @@ class Tag(TagBase):
     has_description.short_description = _('description')
     has_description.boolean = True
 
     has_description.short_description = _('description')
     has_description.boolean = True
 
-    def alive(self):
-        return self.death is None
-
-    def in_pd(self):
-        """ tests whether an author is in public domain """
-        return self.death is not None and self.goes_to_pd() <= datetime.now().year
+    def get_count(self):
+        """ returns global book count for book tags, fragment count for themes """
 
 
-    def goes_to_pd(self):
-        """ calculates the year of public domain entry for an author """
-        return self.death + 71 if self.death is not None else None
+        if self.book_count is None:
+            if self.category == 'book':
+                # never used
+                objects = Book.objects.none()
+            elif self.category == 'theme':
+                objects = Fragment.tagged.with_all((self,))
+            else:
+                objects = Book.tagged.with_all((self,)).order_by()
+                if self.category != 'set':
+                    # eliminate descendants
+                    l_tags = Tag.objects.filter(slug__in=[book.book_tag_slug() for book in objects])
+                    descendants_keys = [book.pk for book in Book.tagged.with_any(l_tags)]
+                    if descendants_keys:
+                        objects = objects.exclude(pk__in=descendants_keys)
+            self.book_count = objects.count()
+            self.save()
+        return self.book_count
 
     @staticmethod
     def get_tag_list(tags):
 
     @staticmethod
     def get_tag_list(tags):
@@ -104,20 +141,21 @@ class Tag(TagBase):
             real_tags = []
             ambiguous_slugs = []
             category = None
             real_tags = []
             ambiguous_slugs = []
             category = None
+            deprecated = False
             tags_splitted = tags.split('/')
             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:
                     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
                 raise Tag.DoesNotExist()
             if category:
                 # something strange left off
                 raise Tag.DoesNotExist()
@@ -127,81 +165,170 @@ class Tag(TagBase):
                 e.tags = real_tags
                 e.ambiguous_slugs = ambiguous_slugs
                 raise e
                 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)
         else:
             return TagBase.get_tag_list(tags)
-    
+
     @property
     def url_chunk(self):
         return '/'.join((Tag.categories_dict[self.category], self.slug))
 
 
     @property
     def url_chunk(self):
         return '/'.join((Tag.categories_dict[self.category], self.slug))
 
 
-# TODO: why is this hard-coded ? 
-def book_upload_path(ext):
-    def get_dynamic_path(book, filename):
-        return 'lektura/%s.%s' % (book.slug, ext)
+# TODO: why is this hard-coded ?
+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:
+            if media.type == 'daisy':
+                ext = 'daisy.zip'
+            else:
+                ext = media.type
+        if not media.name:
+            name = slughifi(filename.split(".")[0])
+        else:
+            name = slughifi(media.name)
+        return 'book/%s/%s.%s' % (ext, name[:maxlen-len('book/%s/.%s' % (ext, ext))-4], ext)
     return get_dynamic_path
 
 
     return get_dynamic_path
 
 
-class Book(models.Model):
-    title = models.CharField(_('title'), max_length=120)
-    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=True)
-    _short_html = models.TextField(_('short HTML'), editable=False)
-    parent_number = models.IntegerField(_('parent number'), default=0)
-    extra_info = JSONField(_('extra information'))
-    gazeta_link = models.CharField(blank=True, max_length=240)
-    wiki_link = models.CharField(blank=True, max_length=240)
+class BookMedia(models.Model):
+    type        = models.CharField(_('type'), choices=MEDIA_FORMATS, max_length="100")
+    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='{}', 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])
 
 
-    # Formats
-    xml_file = models.FileField(_('XML file'), upload_to=book_upload_path('xml'), blank=True)
-    html_file = models.FileField(_('HTML file'), upload_to=book_upload_path('html'), blank=True)
-    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)
-    odt_file = models.FileField(_('ODT file'), upload_to=book_upload_path('odt'), blank=True)
-    txt_file = models.FileField(_('TXT file'), upload_to=book_upload_path('txt'), blank=True)
-    mp3_file = models.FileField(_('MP3 file'), upload_to=book_upload_path('mp3'), blank=True)
-    ogg_file = models.FileField(_('OGG file'), upload_to=book_upload_path('ogg'), blank=True)
+    class Meta:
+        ordering            = ('type', 'name')
+        verbose_name        = _('book media')
+        verbose_name_plural = _('book media')
 
 
-    parent = models.ForeignKey('self', blank=True, null=True, related_name='children')
+    def save(self, *args, **kwargs):
+        try:
+            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)
+
+        # remove the zip package for book with modified media
+        remove_zip(self.book.slug)
+
+        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}
 
 
-    objects = models.Manager()
-    tagged = managers.ModelTaggedItemManager(Tag)
-    tags = managers.TagDescriptor(Tag)
+    @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, db_index=True)
+    changed_at    = models.DateTimeField(_('creation date'), auto_now=True, db_index=True)
+    parent_number = models.IntegerField(_('parent number'), default=0)
+    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
+
+    file_types = ['epub', 'html', 'mobi', 'pdf', 'txt', 'xml']
     
     
-    _tag_counter = JSONField(null=True, editable=False)
-    _theme_counter = JSONField(null=True, editable=False)
+    parent        = models.ForeignKey('self', blank=True, null=True, related_name='children')
+    objects  = models.Manager()
+    tagged   = managers.ModelTaggedItemManager(Tag)
+    tags     = managers.TagDescriptor(Tag)
+
+    html_built = django.dispatch.Signal()
 
     class AlreadyExists(Exception):
         pass
 
     class Meta:
 
     class AlreadyExists(Exception):
         pass
 
     class Meta:
-        ordering = ('title',)
+        ordering = ('sort_key',)
         verbose_name = _('book')
         verbose_name_plural = _('books')
 
     def __unicode__(self):
         return self.title
 
         verbose_name = _('book')
         verbose_name_plural = _('books')
 
     def __unicode__(self):
         return self.title
 
-    def save(self, force_insert=False, force_update=False, reset_short_html=True, refresh_mp3=True):
-        if reset_short_html:
-            # Reset _short_html during save
-            for key in filter(lambda x: x.startswith('_short_html'), self.__dict__):
-                self.__setattr__(key, '')
+    def save(self, force_insert=False, force_update=False, reset_short_html=True, **kwargs):
+        self.sort_key = sortify(self.title)
 
 
-        book = super(Book, self).save(force_insert, force_update)
+        ret = super(Book, self).save(force_insert, force_update)
 
 
-        if refresh_mp3 and self.mp3_file:
-            print self.mp3_file, self.mp3_file.path
-            extra_info = self.get_extra_info_value()
-            extra_info.update(self.get_mp3_info())
-            self.set_extra_info_value(extra_info)
-            book = super(Book, self).save(force_insert, force_update)
+        if reset_short_html:
+            self.reset_short_html()
 
 
-        return book
+        return ret
 
     @permalink
     def get_absolute_url(self):
 
     @permalink
     def get_absolute_url(self):
@@ -210,103 +337,343 @@ class Book(models.Model):
     @property
     def name(self):
         return self.title
     @property
     def name(self):
         return self.title
-    
+
+    def book_tag_slug(self):
+        return ('l-' + self.slug)[:120]
+
     def book_tag(self):
     def book_tag(self):
-        slug = ('l-' + self.slug)[:120]
+        slug = self.book_tag_slug()
         book_tag, created = Tag.objects.get_or_create(slug=slug, category='book')
         if created:
             book_tag.name = self.title[:50]
         book_tag, created = Tag.objects.get_or_create(slug=slug, category='book')
         if created:
             book_tag.name = self.title[:50]
-            book_tag.sort_key = slug
+            book_tag.sort_key = self.title.lower()
             book_tag.save()
         return book_tag
 
             book_tag.save()
         return book_tag
 
+    def has_media(self, type):
+        if type in Book.file_types:
+            return bool(getattr(self, "%s_file" % type))
+        else:
+            return self.media.filter(type=type).exists()
+
+    def get_media(self, type):
+        if self.has_media(type):
+            if type in Book.file_types:
+                return getattr(self, "%s_file" % type)
+            else:                                             
+                return self.media.filter(type=type)
+        else:
+            return None
+
+    def get_mp3(self):
+        return self.get_media("mp3")
+    def get_odt(self):
+        return self.get_media("odt")
+    def get_ogg(self):
+        return self.get_media("ogg")
+    def get_daisy(self):
+        return self.get_media("daisy")                       
+
+    def reset_short_html(self):
+        if self.id is None:
+            return
+
+        cache_key = "Book.short_html/%d/%s"
+        for lang, langname in settings.LANGUAGES:
+            cache.delete(cache_key % (self.id, lang))
+        # Fragment.short_html relies on book's tags, so reset it here too
+        for fragm in self.fragments.all():
+            fragm.reset_short_html()
+
     def short_html(self):
     def short_html(self):
-        key = '_short_html_%s' % get_language()
-        short_html = getattr(self, key)
+        if self.id:
+            cache_key = "Book.short_html/%d/%s" % (self.id, get_language())
+            short_html = cache.get(cache_key)
+        else:
+            short_html = None
 
 
-        if short_html and len(short_html):
+        if short_html is not None:
             return mark_safe(short_html)
         else:
             tags = self.tags.filter(~Q(category__in=('set', 'theme', 'book')))
             tags = [mark_safe(u'<a href="%s">%s</a>' % (tag.get_absolute_url(), tag.name)) for tag in tags]
 
             formats = []
             return mark_safe(short_html)
         else:
             tags = self.tags.filter(~Q(category__in=('set', 'theme', 'book')))
             tags = [mark_safe(u'<a href="%s">%s</a>' % (tag.get_absolute_url(), tag.name)) for tag in tags]
 
             formats = []
-            if self.html_file:
+            # files generated during publication
+            if self.has_media("html"):
                 formats.append(u'<a href="%s">%s</a>' % (reverse('book_text', kwargs={'slug': self.slug}), _('Read online')))
                 formats.append(u'<a href="%s">%s</a>' % (reverse('book_text', kwargs={'slug': self.slug}), _('Read online')))
-            if self.pdf_file:
-                formats.append(u'<a href="%s">PDF</a>' % self.pdf_file.url)
-            if self.epub_file:
-                formats.append(u'<a href="%s">EPUB</a>' % self.epub_file.url)
-            if self.odt_file:
-                formats.append(u'<a href="%s">ODT</a>' % self.odt_file.url)
-            if self.txt_file:
-                formats.append(u'<a href="%s">TXT</a>' % self.txt_file.url)
-            if self.mp3_file:
-                formats.append(u'<a href="%s">MP3</a>' % self.mp3_file.url)
-            if self.ogg_file:
-                formats.append(u'<a href="%s">OGG</a>' % self.ogg_file.url)
+            if self.has_media("pdf"):
+                formats.append(u'<a href="%s">PDF</a>' % self.get_media('pdf').url)
+            if self.has_media("mobi"):
+                formats.append(u'<a href="%s">MOBI</a>' % self.get_media('mobi').url)
+            if self.root_ancestor.has_media("epub"):
+                formats.append(u'<a href="%s">EPUB</a>' % self.root_ancestor.get_media('epub').url)
+            if self.has_media("txt"):
+                formats.append(u'<a href="%s">TXT</a>' % self.get_media('txt').url)
+            # other files
+            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]
 
 
             formats = [mark_safe(format) for format in formats]
 
-            setattr(self, key, unicode(render_to_string('catalogue/book_short.html',
-                {'book': self, 'tags': tags, 'formats': formats})))
-            self.save(reset_short_html=False)
-            return mark_safe(getattr(self, key))
+            short_html = unicode(render_to_string('catalogue/book_short.html',
+                {'book': self, 'tags': tags, 'formats': formats}))
+
+            if self.id:
+                cache.set(cache_key, short_html, CACHE_FOREVER)
+            return mark_safe(short_html)
+
+    @property
+    def root_ancestor(self):
+        """ returns the oldest ancestor """
 
 
+        if not hasattr(self, '_root_ancestor'):
+            book = self
+            while book.parent:
+                book = book.parent
+            self._root_ancestor = book
+        return self._root_ancestor
 
 
-    def get_mp3_info(self):
-        """Retrieves artist and director names from audio ID3 tags."""
-        audio = id3.ID3(self.mp3_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'))
-        return {'artist_name': artist_name, 'director_name': director_name}
 
     def has_description(self):
         return len(self.description) > 0
     has_description.short_description = _('description')
     has_description.boolean = True
 
 
     def has_description(self):
         return len(self.description) > 0
     has_description.short_description = _('description')
     has_description.boolean = True
 
-    def has_pdf_file(self):
-        return bool(self.pdf_file)
-    has_pdf_file.short_description = 'PDF'
-    has_pdf_file.boolean = True
-
-    def has_epub_file(self):
-        return bool(self.epub_file)
-    has_epub_file.short_description = 'EPUB'
-    has_epub_file.boolean = True
-
+    # ugly ugly ugly
     def has_odt_file(self):
     def has_odt_file(self):
-        return bool(self.odt_file)
+        return bool(self.has_media("odt"))
     has_odt_file.short_description = 'ODT'
     has_odt_file.boolean = True
 
     has_odt_file.short_description = 'ODT'
     has_odt_file.boolean = True
 
-    def has_html_file(self):
-        return bool(self.html_file)
-    has_html_file.short_description = 'HTML'
-    has_html_file.boolean = True
+    def has_mp3_file(self):
+        return bool(self.has_media("mp3"))
+    has_mp3_file.short_description = 'MP3'
+    has_mp3_file.boolean = True
+
+    def has_ogg_file(self):
+        return bool(self.has_media("ogg"))
+    has_ogg_file.short_description = 'OGG'
+    has_ogg_file.boolean = True
+
+    def has_daisy_file(self):
+        return bool(self.has_media("daisy"))
+    has_daisy_file.short_description = 'DAISY'
+    has_daisy_file.boolean = True
+
+    def build_pdf(self):
+        """ (Re)builds the pdf file.
+
+        """
+        from librarian import pdf
+        from tempfile import NamedTemporaryFile
+        import os
+
+        try:
+            pdf_file = NamedTemporaryFile(delete=False)
+            pdf.transform(ORMDocProvider(self),
+                      file_path=str(self.xml_file.path),
+                      output_file=pdf_file,
+                      )
+
+            self.pdf_file.save('%s.pdf' % self.slug, File(open(pdf_file.name)))
+        finally:
+            unlink(pdf_file.name)
+
+        # remove zip with all pdf files
+        remove_zip(settings.ALL_PDF_ZIP)
+
+    def build_mobi(self):
+        """ (Re)builds the MOBI file.
+
+        """
+        from librarian import mobi
+        from tempfile import NamedTemporaryFile
+        import os
+
+        try:
+            mobi_file = NamedTemporaryFile(suffix='.mobi', delete=False)
+            mobi.transform(ORMDocProvider(self), verbose=1,
+                      file_path=str(self.xml_file.path),
+                      output_file=mobi_file.name,
+                      )
+
+            self.mobi_file.save('%s.mobi' % self.slug, File(open(mobi_file.name)))
+        finally:
+            unlink(mobi_file.name)
+
+        # remove zip with all mobi files
+        remove_zip(settings.ALL_MOBI_ZIP)
+
+    def build_epub(self, remove_descendants=True):
+        """ (Re)builds the epub file.
+            If book has a parent, does nothing.
+            Unless remove_descendants is False, descendants' epubs are removed.
+        """
+        from StringIO import StringIO
+        from hashlib import sha1
+        from django.core.files.base import ContentFile
+
+        if self.parent:
+            # don't need an epub
+            return
+
+        epub_file = StringIO()
+        try:
+            epub.transform(ORMDocProvider(self), self.slug, output_file=epub_file)
+            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
+
+        book_descendants = list(self.children.all())
+        while len(book_descendants) > 0:
+            child_book = book_descendants.pop(0)
+            if remove_descendants and child_book.has_epub_file():
+                child_book.epub_file.delete()
+            # save anyway, to refresh short_html
+            child_book.save()
+            book_descendants += list(child_book.children.all())
+
+        # remove zip package with all epub files
+        remove_zip(settings.ALL_EPUB_ZIP)
+
+    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()
+            self.html_built.send(sender=self)
+            return True
+        return False
+
+    @staticmethod
+    def zip_format(format_):
+        def pretty_file_name(book):
+            return "%s/%s.%s" % (
+                b.get_extra_info_value()['author'],
+                b.slug,
+                format_)
+
+        field_name = "%s_file" % format_
+        books = Book.objects.filter(parent=None).exclude(**{field_name: ""})
+        paths = [(pretty_file_name(b), getattr(b, field_name).path)
+                    for b in books]
+        result = create_zip.delay(paths,
+                    getattr(settings, "ALL_%s_ZIP" % format_.upper()))
+        return result.wait()
+
+    def zip_audiobooks(self):
+        bm = BookMedia.objects.filter(book=self, type='mp3')
+        paths = map(lambda bm: (None, bm.file.path), bm)
+        result = create_zip.delay(paths, self.slug)
+        return result.wait()
+
+    def search_index(self):
+        if settings.SEARCH_INDEX_PARALLEL:
+            if instance(settings.SEARCH_INDEX_PARALLEL, int):
+                idx = search.ReusableIndex(threads=4)
+            else:
+                idx = search.ReusableIndex()
+        else:
+            idx = search.Index()
+            
+        idx.open()
+        try:
+            idx.index_book(self)
+        finally:
+            idx.close()
 
     @classmethod
 
     @classmethod
-    def from_xml_file(cls, xml_file, overwrite=False):
+    def from_xml_file(cls, xml_file, **kwargs):
         # use librarian to parse meta-data
         book_info = dcparser.parse(xml_file)
 
         if not isinstance(xml_file, File):
         # use librarian to parse meta-data
         book_info = dcparser.parse(xml_file)
 
         if not isinstance(xml_file, File):
-            xml_file = File(xml_file)
+            xml_file = File(open(xml_file))
 
         try:
 
         try:
-            return cls.from_text_and_meta(xml_file, book_info, overwrite)
+            return cls.from_text_and_meta(xml_file, book_info, **kwargs)
         finally:
             xml_file.close()
 
     @classmethod
         finally:
             xml_file.close()
 
     @classmethod
-    def from_text_and_meta(cls, raw_file, book_info, overwrite=False):
-        from tempfile import NamedTemporaryFile
-        from slughifi import slughifi
-        from markupstring import MarkupString
+    def from_text_and_meta(cls, raw_file, book_info, overwrite=False,
+            build_epub=True, build_txt=True, build_pdf=True, build_mobi=True,
+            search_index=True):
+        import re
+
+        # check for parts before we do anything
+        children = []
+        if hasattr(book_info, 'parts'):
+            for part_url in book_info.parts:
+                base, slug = part_url.rsplit('/', 1)
+                try:
+                    children.append(Book.objects.get(slug=slug))
+                except Book.DoesNotExist, e:
+                    raise Book.DoesNotExist(_('Book with slug = "%s" does not exist.') % slug)
+
 
         # Read book metadata
         book_base, book_slug = book_info.url.rsplit('/', 1)
 
         # 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:
         book, created = Book.objects.get_or_create(slug=book_slug)
 
         if created:
@@ -319,39 +686,61 @@ class Book(models.Model):
 
         book.title = book_info.title
         book.set_extra_info_value(book_info.to_dict())
 
         book.title = book_info.title
         book.set_extra_info_value(book_info.to_dict())
-        book._short_html = ''
         book.save()
 
         book.save()
 
-        book_tags = []
-        for category in ('kind', 'genre', 'author', 'epoch'):
-            tag_name = getattr(book_info, category)
-            tag_sort_key = tag_name
-            if category == 'author':
-                tag_sort_key = tag_name.last_name
-                tag_name = ' '.join(tag_name.first_names) + ' ' + tag_name.last_name
-            tag, created = Tag.objects.get_or_create(slug=slughifi(tag_name), category=category)
-            if created:
-                tag.name = tag_name
-                tag.sort_key = slughifi(tag_sort_key)
-                tag.save()
-            book_tags.append(tag)
-
-        book.tags = book_tags
+        meta_tags = []
+        categories = (('kinds', 'kind'), ('genres', 'genre'), ('authors', 'author'), ('epochs', 'epoch'))
+        for field_name, category in categories:
+            try:
+                tag_names = getattr(book_info, field_name)
+            except:
+                tag_names = [getattr(book_info, category)]
+            for tag_name in tag_names:
+                tag_sort_key = tag_name
+                if category == 'author':
+                    tag_sort_key = tag_name.last_name
+                    tag_name = ' '.join(tag_name.first_names) + ' ' + tag_name.last_name
+                tag, created = Tag.objects.get_or_create(slug=slughifi(tag_name), category=category)
+                if created:
+                    tag.name = tag_name
+                    tag.sort_key = sortify(tag_sort_key.lower())
+                    tag.save()
+                meta_tags.append(tag)
+
+        book.tags = set(meta_tags + book_shelves)
 
         book_tag = book.book_tag()
 
 
         book_tag = book.book_tag()
 
-        if hasattr(book_info, 'parts'):
-            for n, part_url in enumerate(book_info.parts):
-                base, slug = part_url.rsplit('/', 1)
-                try:
-                    child_book = Book.objects.get(slug=slug)
-                    child_book.parent = book
-                    child_book.parent_number = n
-                    child_book.save()
-                except Book.DoesNotExist, e:
-                    raise Book.DoesNotExist(_('Book with slug = "%s" does not exist.') % slug)
+        for n, child_book in enumerate(children):
+            child_book.parent = book
+            child_book.parent_number = n
+            child_book.save()
+
+        # Save XML and HTML files
+        book.xml_file.save('%s.xml' % book.slug, raw_file, save=False)
+
+        # delete old fragments when overwriting
+        book.fragments.all().delete()
+
+        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()
+
+        if not settings.NO_BUILD_PDF and build_pdf:
+            book.root_ancestor.build_pdf()
+
+        if not settings.NO_BUILD_MOBI and build_mobi:
+            book.build_mobi()
+
+        if not settings.NO_SEARCH_INDEX and search_index:
+            book.search_index()
 
         book_descendants = list(book.children.all())
 
         book_descendants = list(book.children.all())
+        # add l-tag to descendants and their fragments
+        # delete unnecessary EPUB files
         while len(book_descendants) > 0:
             child_book = book_descendants.pop(0)
             child_book.tags = list(child_book.tags) + [book_tag]
         while len(book_descendants) > 0:
             child_book = book_descendants.pop(0)
             child_book.tags = list(child_book.tags) + [book_tag]
@@ -360,85 +749,127 @@ class Book(models.Model):
                 fragment.tags = set(list(fragment.tags) + [book_tag])
             book_descendants += list(child_book.children.all())
 
                 fragment.tags = set(list(fragment.tags) + [book_tag])
             book_descendants += list(child_book.children.all())
 
-        # Save XML and HTML files
-        book.xml_file.save('%s.xml' % book.slug, raw_file, save=False)
+        book.save()
 
 
-        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)
+        # refresh cache
+        book.reset_tag_counter()
+        book.reset_theme_counter()
 
 
-            # Extract fragments
-            closed_fragments, open_fragments = html.extract_fragments(book.html_file.path)
-            book_themes = []
-            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})
+        return book
 
 
-                try:
-                    theme_names = [s.strip() for s in fragment.themes.split(',')]
-                except AttributeError:
-                    continue
-                themes = []
-                for theme_name in theme_names:
-                    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)
-                new_fragment.save()
-                new_fragment.tags = set(list(book.tags) + themes + [book_tag])
-                book_themes += themes
+    def reset_tag_counter(self):
+        if self.id is None:
+            return
 
 
-            book_themes = set(book_themes)
-            book.tags = list(book.tags) + list(book_themes) + book_shelves
+        cache_key = "Book.tag_counter/%d" % self.id
+        cache.delete(cache_key)
+        if self.parent:
+            self.parent.reset_tag_counter()
 
 
-        book.save()
-        return book
-    
-    
-    def refresh_tag_counter(self):
-        tags = {}
-        for child in self.children.all().order_by():
-            for tag_pk, value in child.tag_counter.iteritems():
-                tags[tag_pk] = tags.get(tag_pk, 0) + value
-        for tag in self.tags.exclude(category__in=('book', 'theme', 'set')).order_by():
-            tags[tag.pk] = 1
-        self.set__tag_counter_value(tags)
-        self.save(reset_short_html=False, refresh_mp3=False)
-        return tags
-    
     @property
     def tag_counter(self):
     @property
     def tag_counter(self):
-        if self._tag_counter is None:
-            return self.refresh_tag_counter()
-        return dict((int(k), v) for k, v in self.get__tag_counter_value().iteritems())
-
-    def refresh_theme_counter(self):
-        tags = {}
-        for fragment in Fragment.tagged.with_any([self.book_tag()]).order_by():
-            for tag in fragment.tags.filter(category='theme').order_by():
-                tags[tag.pk] = tags.get(tag.pk, 0) + 1
-        self.set__theme_counter_value(tags)
-        self.save(reset_short_html=False, refresh_mp3=False)
+        if self.id:
+            cache_key = "Book.tag_counter/%d" % self.id
+            tags = cache.get(cache_key)
+        else:
+            tags = None
+
+        if tags is None:
+            tags = {}
+            for child in self.children.all().order_by():
+                for tag_pk, value in child.tag_counter.iteritems():
+                    tags[tag_pk] = tags.get(tag_pk, 0) + value
+            for tag in self.tags.exclude(category__in=('book', 'theme', 'set')).order_by():
+                tags[tag.pk] = 1
+
+            if self.id:
+                cache.set(cache_key, tags, CACHE_FOREVER)
         return tags
         return tags
-    
+
+    def reset_theme_counter(self):
+        if self.id is None:
+            return
+
+        cache_key = "Book.theme_counter/%d" % self.id
+        cache.delete(cache_key)
+        if self.parent:
+            self.parent.reset_theme_counter()
+
     @property
     def theme_counter(self):
     @property
     def theme_counter(self):
-        if self._theme_counter is None:
-            return self.refresh_theme_counter()
-        return dict((int(k), v) for k, v in self.get__theme_counter_value().iteritems())
+        if self.id:
+            cache_key = "Book.theme_counter/%d" % self.id
+            tags = cache.get(cache_key)
+        else:
+            tags = None
+
+        if tags is None:
+            tags = {}
+            for fragment in Fragment.tagged.with_any([self.book_tag()]).order_by():
+                for tag in fragment.tags.filter(category='theme').order_by():
+                    tags[tag.pk] = tags.get(tag.pk, 0) + 1
+
+            if self.id:
+                cache.set(cache_key, tags, CACHE_FOREVER)
+        return tags
+
+    def pretty_title(self, html_links=False):
+        book = self
+        names = list(book.tags.filter(category='author'))
+
+        books = []
+        while book:
+            books.append(book)
+            book = book.parent
+        names.extend(reversed(books))
+
+        if html_links:
+            names = ['<a href="%s">%s</a>' % (tag.get_absolute_url(), tag.name) for tag in names]
+        else:
+            names = [tag.name for tag in names]
+
+        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
+
+
+def _has_factory(ftype):
+    has = lambda self: bool(getattr(self, "%s_file" % ftype))
+    has.short_description = t.upper()
+    has.boolean = True
+    has.__name__ = "has_%s_file" % ftype
+    return has
+
     
     
+# add the file fields
+for t in Book.file_types:
+    field_name = "%s_file" % t
+    models.FileField(_("%s file" % t.upper()),
+            upload_to=book_upload_path(t),
+            blank=True).contribute_to_class(Book, field_name)
+
+    setattr(Book, "has_%s_file" % t, _has_factory(t))
 
 
 class Fragment(models.Model):
     text = models.TextField()
     short_text = models.TextField(editable=False)
 
 
 class Fragment(models.Model):
     text = models.TextField()
     short_text = models.TextField(editable=False)
-    _short_html = models.TextField(editable=False)
     anchor = models.CharField(max_length=120)
     book = models.ForeignKey(Book, related_name='fragments')
 
     anchor = models.CharField(max_length=120)
     book = models.ForeignKey(Book, related_name='fragments')
 
@@ -454,46 +885,78 @@ class Fragment(models.Model):
     def get_absolute_url(self):
         return '%s#m%s' % (reverse('book_text', kwargs={'slug': self.book.slug}), self.anchor)
 
     def get_absolute_url(self):
         return '%s#m%s' % (reverse('book_text', kwargs={'slug': self.book.slug}), self.anchor)
 
+    def reset_short_html(self):
+        if self.id is None:
+            return
+
+        cache_key = "Fragment.short_html/%d/%s"
+        for lang, langname in settings.LANGUAGES:
+            cache.delete(cache_key % (self.id, lang))
+
     def short_html(self):
     def short_html(self):
-        key = '_short_html_%s' % get_language()
-        short_html = getattr(self, key)
-        if short_html and len(short_html):
-            return mark_safe(short_html)
+        if self.id:
+            cache_key = "Fragment.short_html/%d/%s" % (self.id, get_language())
+            short_html = cache.get(cache_key)
         else:
         else:
-            book_authors = [mark_safe(u'<a href="%s">%s</a>' % (tag.get_absolute_url(), tag.name))
-                for tag in self.book.tags if tag.category == 'author']
+            short_html = None
 
 
-            setattr(self, key, unicode(render_to_string('catalogue/fragment_short.html',
-                {'fragment': self, 'book': self.book, 'book_authors': book_authors})))
-            self.save()
-            return mark_safe(getattr(self, key))
+        if short_html is not None:
+            return mark_safe(short_html)
+        else:
+            short_html = unicode(render_to_string('catalogue/fragment_short.html',
+                {'fragment': self}))
+            if self.id:
+                cache.set(cache_key, short_html, CACHE_FOREVER)
+            return mark_safe(short_html)
 
 
 
 
-class BookStub(models.Model):
-    title = models.CharField(_('title'), max_length=120)
-    author = models.CharField(_('author'), max_length=120)
-    pd = models.IntegerField(_('goes to public domain'), null=True, blank=True)
-    slug = models.SlugField(_('slug'), max_length=120, unique=True, db_index=True)
-    translator = models.TextField(_('translator'), blank=True)
-    translator_death = models.TextField(_('year of translator\'s death'), blank=True)
+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:
 
     class Meta:
-        ordering = ('title',)
-        verbose_name = _('book stub')
-        verbose_name_plural = _('book stubs')
+        ordering = ('-time','-slug', '-type')
+        verbose_name = _('file record')
+        verbose_name_plural = _('file records')
 
     def __unicode__(self):
 
     def __unicode__(self):
-        return self.title
-
-    @permalink
-    def get_absolute_url(self):
-        return ('catalogue.views.book_detail', [self.slug])
-
-    def in_pd(self):
-        return self.pd is not None and self.pd <= datetime.now().year
-
-    @property
-    def name(self):
-        return self.title
-
+        return "%s %s.%s" % (self.sha1,  self.slug, self.type)
 
 
+###########
+#
+# SIGNALS
+#
+###########
+
+
+def _tags_updated_handler(sender, affected_tags, **kwargs):
+    # reset tag global counter
+    # 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 \
+                Tag.objects.filter(pk__in=(tag.pk for tag in affected_tags)).\
+                    exclude(category__in=('book', 'theme', 'set')).count():
+        sender.reset_tag_counter()
+    # if fragment theme changed, reset book theme counter
+    elif isinstance(sender, Fragment) and \
+                Tag.objects.filter(pk__in=(tag.pk for tag in affected_tags)).\
+                    filter(category='theme').count():
+        sender.book.reset_theme_counter()
+tags_updated.connect(_tags_updated_handler)
+
+
+def _pre_delete_handler(sender, instance, **kwargs):
+    """ refresh Book on BookMedia delete """
+    if sender == BookMedia:
+        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:
+        instance.book.save()
+post_save.connect(_post_save_handler)