Merge branch 'production' of git://github.com/fnp/wolnelektury into production
[wolnelektury.git] / apps / catalogue / models.py
index f22c9e1..3704b16 100644 (file)
@@ -5,10 +5,9 @@
 from collections import namedtuple
 
 from django.db import models
-from django.db.models import permalink, Q
+from django.db.models import permalink
 import django.dispatch
-from django.core.cache import cache
-from django.core.files.storage import DefaultStorage
+from django.core.cache import get_cache
 from django.utils.translation import ugettext_lazy as _
 from django.contrib.auth.models import User
 from django.template.loader import render_to_string
@@ -16,21 +15,20 @@ from django.utils.datastructures import SortedDict
 from django.utils.safestring import mark_safe
 from django.utils.translation import get_language
 from django.core.urlresolvers import reverse
-from django.db.models.signals import post_save, m2m_changed, pre_delete
+from django.db.models.signals import post_save, pre_delete, post_delete
+import jsonfield
 
 from django.conf import settings
 
 from newtagging.models import TagBase, tags_updated
 from newtagging import managers
-from catalogue.fields import JSONField, OverwritingFileField
-from catalogue.utils import create_zip
-from catalogue.tasks import touch_tag
-from shutil import copy
-from glob import glob
+from catalogue.fields import OverwritingFileField
+from catalogue.utils import create_zip, split_tags, truncate_html_words
+from catalogue import tasks
 import re
-from os import path
 
 
+# Those are hard-coded here so that makemessages sees them.
 TAG_CATEGORIES = (
     ('author', _('author')),
     ('epoch', _('epoch')),
@@ -41,8 +39,8 @@ TAG_CATEGORIES = (
     ('book', _('book')),
 )
 
-# not quite, but Django wants you to set a timeout
-CACHE_FOREVER = 2419200  # 28 days
+
+permanent_cache = get_cache('permanent')
 
 
 class TagSubcategoryManager(models.Manager):
@@ -55,13 +53,16 @@ class TagSubcategoryManager(models.Manager):
 
 
 class Tag(TagBase):
+    """A tag attachable to books and fragments (and possibly anything).
+    
+    Used to represent searchable metadata (authors, epochs, genres, kinds),
+    fragment themes (motifs) and some book hierarchy related kludges."""
     name = models.CharField(_('name'), max_length=50, db_index=True)
     slug = models.SlugField(_('slug'), 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)
     book_count = models.IntegerField(_('book count'), blank=True, null=True)
@@ -117,8 +118,8 @@ class Tag(TagBase):
             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)]
+                l_tags = Tag.objects.filter(slug__in=[book.book_tag_slug() for book in objects.iterator()])
+                descendants_keys = [book.pk for book in Book.tagged.with_any(l_tags).iterator()]
                 if descendants_keys:
                     objects = objects.exclude(pk__in=descendants_keys)
         return objects.count()
@@ -175,7 +176,11 @@ class Tag(TagBase):
             try:
                 tag_names = getattr(info, field_name)
             except:
-                tag_names = [getattr(info, category)]
+                try:
+                    tag_names = [getattr(info, category)]
+                except:
+                    # For instance, Pictures do not have 'genre' field.
+                    continue
             for tag_name in tag_names:
                 tag_sort_key = tag_name
                 if category == 'author':
@@ -210,30 +215,8 @@ def book_upload_path(ext=None, maxlen=100):
     return lambda *args: get_dynamic_path(*args, ext=ext, maxlen=maxlen)
 
 
-def get_customized_pdf_path(book, customizations):
-    """
-    Returns a MEDIA_ROOT relative path for a customized pdf. The name will contain a hash of customization options.
-    """
-    customizations.sort()
-    h = hash(tuple(customizations))
-
-    pdf_name = '%s-custom-%s' % (book.fileid(), h)
-    pdf_file = get_dynamic_path(None, pdf_name, ext='pdf')
-
-    return pdf_file
-
-
-def get_existing_customized_pdf(book):
-    """
-    Returns a list of paths to generated customized pdf of a book
-    """
-    pdf_glob = '%s-custom-' % (book.fileid(),)
-    pdf_glob = get_dynamic_path(None, pdf_glob, ext='pdf')
-    pdf_glob = re.sub(r"[.]([a-z0-9]+)$", "*.\\1", pdf_glob)
-    return glob(path.join(settings.MEDIA_ROOT, pdf_glob))
-
-
 class BookMedia(models.Model):
+    """Represents media attached to a book."""
     FileFormat = namedtuple("FileFormat", "name ext")
     formats = SortedDict([
         ('mp3', FileFormat(name='MP3', ext='mp3')),
@@ -247,7 +230,7 @@ class BookMedia(models.Model):
     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)
+    extra_info  = jsonfield.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)
 
@@ -265,8 +248,8 @@ class BookMedia(models.Model):
 
         try:
             old = BookMedia.objects.get(pk=self.pk)
-        except BookMedia.DoesNotExist, e:
-            pass
+        except BookMedia.DoesNotExist:
+            old = None
         else:
             # if name changed, change the file name, too
             if slughifi(self.name) != slughifi(old.name):
@@ -275,11 +258,13 @@ class BookMedia(models.Model):
         super(BookMedia, self).save(*args, **kwargs)
 
         # remove the zip package for book with modified media
-        remove_zip(self.book.fileid())
+        if old:
+            remove_zip("%s_%s" % (old.book.slug, old.type))
+        remove_zip("%s_%s" % (self.book.slug, self.type))
 
-        extra_info = self.get_extra_info_value()
+        extra_info = self.extra_info
         extra_info.update(self.read_meta())
-        self.set_extra_info_value(extra_info)
+        self.extra_info = extra_info
         self.source_sha1 = self.read_source_sha1(self.file.path, self.type)
         return super(BookMedia, self).save(*args, **kwargs)
 
@@ -342,26 +327,32 @@ class BookMedia(models.Model):
 
 
 class Book(models.Model):
+    """Represents a book imported from WL-XML."""
     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, db_index=True)
+    slug = models.SlugField(_('slug'), max_length=120, db_index=True,
+            unique=True)
+    common_slug = models.SlugField(_('slug'), max_length=120, db_index=True)
     language = models.CharField(_('language code'), max_length=3, db_index=True,
                     default=settings.CATALOGUE_DEFAULT_LANGUAGE)
     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='{}')
+    extra_info    = jsonfield.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
 
     cover = models.FileField(_('cover'), upload_to=book_upload_path('png'),
                 null=True, blank=True)
-    ebook_formats = ['pdf', 'epub', 'mobi', 'txt']
+    ebook_formats = ['pdf', 'epub', 'mobi', 'fb2', 'txt']
     formats = ebook_formats + ['html', 'xml']
 
     parent        = models.ForeignKey('self', blank=True, null=True, related_name='children')
+
+    _related_info = jsonfield.JSONField(blank=True, null=True, editable=False)
+
     objects  = models.Manager()
     tagged   = managers.ModelTaggedItemManager(Tag)
     tags     = managers.TagDescriptor(Tag)
@@ -369,14 +360,10 @@ class Book(models.Model):
     html_built = django.dispatch.Signal()
     published = django.dispatch.Signal()
 
-    URLID_RE = r'[a-z0-9-]+(?:/[a-z]{3})?'
-    FILEID_RE = r'[a-z0-9-]+(?:_[a-z]{3})?'
-
     class AlreadyExists(Exception):
         pass
 
     class Meta:
-        unique_together = [['slug', 'language']]
         ordering = ('sort_key',)
         verbose_name = _('book')
         verbose_name_plural = _('books')
@@ -384,42 +371,6 @@ class Book(models.Model):
     def __unicode__(self):
         return self.title
 
-    def urlid(self, sep='/'):
-        stem = self.slug
-        if self.language != settings.CATALOGUE_DEFAULT_LANGUAGE:
-            stem += sep + self.language
-        return stem
-
-    def fileid(self):
-        return self.urlid('_')
-
-    @staticmethod
-    def split_urlid(urlid, sep='/', default_lang=settings.CATALOGUE_DEFAULT_LANGUAGE):
-        """Splits a URL book id into slug and language code.
-        
-        Returns a dictionary usable i.e. for object lookup, or None.
-
-        >>> Book.split_urlid("a-slug/pol", default_lang="eng")
-        {'slug': 'a-slug', 'language': 'pol'}
-        >>> Book.split_urlid("a-slug", default_lang="eng")
-        {'slug': 'a-slug', 'language': 'eng'}
-        >>> Book.split_urlid("a-slug_pol", "_", default_lang="eng")
-        {'slug': 'a-slug', 'language': 'pol'}
-        >>> Book.split_urlid("a-slug/eng", default_lang="eng")
-
-        """
-        parts = urlid.rsplit(sep, 1)
-        if len(parts) == 2:
-            if parts[1] == default_lang:
-                return None
-            return {'slug': parts[0], 'language': parts[1]}
-        else:
-            return {'slug': urlid, 'language': default_lang}
-
-    @classmethod
-    def split_fileid(cls, fileid):
-        return cls.split_urlid(fileid, '_')
-
     def save(self, force_insert=False, force_update=False, reset_short_html=True, **kwargs):
         from sortify import sortify
 
@@ -434,18 +385,14 @@ class Book(models.Model):
 
     @permalink
     def get_absolute_url(self):
-        return ('catalogue.views.book_detail', [self.urlid()])
+        return ('catalogue.views.book_detail', [self.slug])
 
     @property
     def name(self):
         return self.title
 
     def book_tag_slug(self):
-        stem = 'l-' + self.slug
-        if self.language != settings.CATALOGUE_DEFAULT_LANGUAGE:
-            return stem[:116] + ' ' + self.language
-        else:
-            return stem[:120]
+        return ('l-' + self.slug)[:120]
 
     def book_tag(self):
         slug = self.book_tag_slug()
@@ -456,18 +403,18 @@ class Book(models.Model):
             book_tag.save()
         return book_tag
 
-    def has_media(self, type):
-        if type in Book.formats:
-            return bool(getattr(self, "%s_file" % type))
+    def has_media(self, type_):
+        if type_ in Book.formats:
+            return bool(getattr(self, "%s_file" % type_))
         else:
-            return self.media.filter(type=type).exists()
+            return self.media.filter(type=type_).exists()
 
-    def get_media(self, type):
-        if self.has_media(type):
-            if type in Book.formats:
-                return getattr(self, "%s_file" % type)
+    def get_media(self, type_):
+        if self.has_media(type_):
+            if type_ in Book.formats:
+                return getattr(self, "%s_file" % type_)
             else:                                             
-                return self.media.filter(type=type)
+                return self.media.filter(type=type_)
         else:
             return None
 
@@ -484,66 +431,11 @@ class Book(models.Model):
         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))
+        type(self).objects.filter(pk=self.pk).update(_related_info=None)
         # Fragment.short_html relies on book's tags, so reset it here too
-        for fragm in self.fragments.all():
+        for fragm in self.fragments.all().iterator():
             fragm.reset_short_html()
 
-    def short_html(self):
-        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 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 = []
-            # files generated during publication
-            if self.has_media("html"):
-                formats.append(u'<a href="%s">%s</a>' % (reverse('book_text', args=[self.fileid()]), _('Read online')))
-            for ebook_format in self.ebook_formats:
-                if self.has_media(ebook_format):
-                    formats.append(u'<a href="%s">%s</a>' % (
-                        self.get_media(ebook_format).url,
-                        ebook_format.upper()
-                    ))
-            # 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]
-
-            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)
-
-    def mini_box(self):
-        if self.id:
-            cache_key = "Book.mini_boxs/%d" % (self.id, )
-            short_html = cache.get(cache_key)
-        else:
-            short_html = None
-
-        if short_html is None:
-            authors = self.tags.filter(category='author')
-
-            short_html = unicode(render_to_string('catalogue/book_mini_box.html',
-                {'book': self, 'authors': authors, 'STATIC_URL': settings.STATIC_URL}))
-
-            if self.id:
-                cache.set(cache_key, short_html, CACHE_FOREVER)
-        return mark_safe(short_html)
-
     def has_description(self):
         return len(self.description) > 0
     has_description.short_description = _('description')
@@ -566,7 +458,7 @@ class Book(models.Model):
     has_daisy_file.boolean = True
 
     def wldocument(self, parse_dublincore=True):
-        from catalogue.utils import ORMDocProvider
+        from catalogue.import_utils import ORMDocProvider
         from librarian.parser import WLDocument
 
         return WLDocument.from_file(self.xml_file.path,
@@ -587,73 +479,10 @@ class Book(models.Model):
         cover.save(imgstr, 'png')
         self.cover.save(None, ContentFile(imgstr.getvalue()))
 
-    def build_pdf(self, customizations=None, file_name=None):
-        """ (Re)builds the pdf file.
-        customizations - customizations which are passed to LaTeX class file.
-        file_name - save the pdf file under a different name and DO NOT save it in db.
-        """
-        from os import unlink
-        from django.core.files import File
-        from catalogue.utils import remove_zip
-
-        pdf = self.wldocument().as_pdf(customizations=customizations)
-
-        if file_name is None:
-            # we'd like to be sure not to overwrite changes happening while
-            # (timely) pdf generation is taking place (async celery scenario)
-            current_self = Book.objects.get(id=self.id)
-            current_self.pdf_file.save('%s.pdf' % self.fileid(),
-                    File(open(pdf.get_filename())))
-            self.pdf_file = current_self.pdf_file
-
-            # remove cached downloadables
-            remove_zip(settings.ALL_PDF_ZIP)
-
-            for customized_pdf in get_existing_customized_pdf(self):
-                unlink(customized_pdf)
-        else:
-            print "saving %s" % file_name
-            print "to: %s" % DefaultStorage().path(file_name)
-            DefaultStorage().save(file_name, File(open(pdf.get_filename())))
-
-    def build_mobi(self):
-        """ (Re)builds the MOBI file.
-
-        """
-        from django.core.files import File
-        from catalogue.utils import remove_zip
-
-        mobi = self.wldocument().as_mobi()
-
-        self.mobi_file.save('%s.mobi' % self.fileid(), File(open(mobi.get_filename())))
-
-        # remove zip with all mobi files
-        remove_zip(settings.ALL_MOBI_ZIP)
-
-    def build_epub(self):
-        """(Re)builds the epub file."""
-        from django.core.files import File
-        from catalogue.utils import remove_zip
-
-        epub = self.wldocument().as_epub()
-
-        self.epub_file.save('%s.epub' % self.fileid(),
-                File(open(epub.get_filename())))
-
-        # remove zip package with all epub files
-        remove_zip(settings.ALL_EPUB_ZIP)
-
-    def build_txt(self):
-        from django.core.files.base import ContentFile
-
-        text = self.wldocument().as_text()
-        self.txt_file.save('%s.txt' % self.fileid(), ContentFile(text.get_string()))
-
-
     def build_html(self):
-        from markupstring import MarkupString
         from django.core.files.base import ContentFile
         from slughifi import slughifi
+        from sortify import sortify
         from librarian import html
 
         meta_tags = list(self.tags.filter(
@@ -662,7 +491,7 @@ class Book(models.Model):
 
         html_output = self.wldocument(parse_dublincore=False).as_html()
         if html_output:
-            self.html_file.save('%s.html' % self.fileid(),
+            self.html_file.save('%s.html' % self.slug,
                     ContentFile(html_output.get_string()))
 
             # get ancestor l-tags for adding to new fragments
@@ -688,16 +517,16 @@ class Book(models.Model):
                     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.sort_key = sortify(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])
+                short_text = truncate_html_words(text, 15)
+                if text == short_text:
+                    short_text = ''
                 new_fragment = Fragment.objects.create(anchor=fragment.id, book=self,
                     text=text, short_text=short_text)
 
@@ -708,27 +537,57 @@ class Book(models.Model):
             return True
         return False
 
+    # Thin wrappers for builder tasks
+    def build_pdf(self, *args, **kwargs):
+        """(Re)builds PDF."""
+        return tasks.build_pdf.delay(self.pk, *args, **kwargs)
+    def build_epub(self, *args, **kwargs):
+        """(Re)builds EPUB."""
+        return tasks.build_epub.delay(self.pk, *args, **kwargs)
+    def build_mobi(self, *args, **kwargs):
+        """(Re)builds MOBI."""
+        return tasks.build_mobi.delay(self.pk, *args, **kwargs)
+    def build_fb2(self, *args, **kwargs):
+        """(Re)build FB2"""
+        return tasks.build_fb2.delay(self.pk, *args, **kwargs)
+    def build_txt(self, *args, **kwargs):
+        """(Re)builds TXT."""
+        return tasks.build_txt.delay(self.pk, *args, **kwargs)
+
     @staticmethod
     def zip_format(format_):
         def pretty_file_name(book):
             return "%s/%s.%s" % (
-                b.get_extra_info_value()['author'],
-                b.fileid(),
+                b.extra_info['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,
+                    for b in books.iterator()]
+        return create_zip(paths,
                     getattr(settings, "ALL_%s_ZIP" % format_.upper()))
-        return result.wait()
 
-    def zip_audiobooks(self):
-        bm = BookMedia.objects.filter(book=self, type='mp3')
+    def zip_audiobooks(self, format_):
+        bm = BookMedia.objects.filter(book=self, type=format_)
         paths = map(lambda bm: (None, bm.file.path), bm)
-        result = create_zip.delay(paths, self.fileid())
-        return result.wait()
+        return create_zip(paths, "%s_%s" % (self.slug, format_))
+
+    def search_index(self, book_info=None, reuse_index=False, index_tags=True):
+        import search
+        if reuse_index:
+            idx = search.ReusableIndex()
+        else:
+            idx = search.Index()
+            
+        idx.open()
+        try:
+            idx.index_book(self, book_info)
+            if index_tags:
+                idx.index_tags()
+        finally:
+            idx.close()
 
     @classmethod
     def from_xml_file(cls, xml_file, **kwargs):
@@ -748,40 +607,42 @@ class Book(models.Model):
 
     @classmethod
     def from_text_and_meta(cls, raw_file, book_info, overwrite=False,
-            build_epub=True, build_txt=True, build_pdf=True, build_mobi=True):
-        import re
-        from sortify import sortify
+            build_epub=True, build_txt=True, build_pdf=True, build_mobi=True, build_fb2=True,
+            search_index=True, search_index_tags=True, search_index_reuse=False):
 
         # check for parts before we do anything
         children = []
         if hasattr(book_info, 'parts'):
             for part_url in book_info.parts:
                 try:
-                    children.append(Book.objects.get(
-                        slug=part_url.slug, language=part_url.language))
-                except Book.DoesNotExist, e:
-                    raise Book.DoesNotExist(_('Book "%s/%s" does not exist.') %
-                            (part_url.slug, part_url.language))
+                    children.append(Book.objects.get(slug=part_url.slug))
+                except Book.DoesNotExist:
+                    raise Book.DoesNotExist(_('Book "%s" does not exist.') %
+                            part_url.slug)
 
 
         # Read book metadata
         book_slug = book_info.url.slug
-        language = book_info.language
-        if re.search(r'[^a-zA-Z0-9-]', book_slug):
+        if re.search(r'[^a-z0-9-]', book_slug):
             raise ValueError('Invalid characters in slug')
-        book, created = Book.objects.get_or_create(slug=book_slug, language=language)
+        book, created = Book.objects.get_or_create(slug=book_slug)
 
         if created:
             book_shelves = []
         else:
             if not overwrite:
-                raise Book.AlreadyExists(_('Book %s/%s already exists') % (
-                        book_slug, language))
+                raise Book.AlreadyExists(_('Book %s already exists') % (
+                        book_slug))
             # Save shelves for this book
             book_shelves = list(book.tags.filter(category='set'))
 
+        book.language = book_info.language
         book.title = book_info.title
-        book.set_extra_info_value(book_info.to_dict())
+        if book_info.variant_of:
+            book.common_slug = book_info.variant_of.slug
+        else:
+            book.common_slug = book.slug
+        book.extra_info = book_info.to_dict()
         book.save()
 
         meta_tags = Tag.tags_from_info(book_info)
@@ -816,6 +677,13 @@ class Book(models.Model):
         if not settings.NO_BUILD_MOBI and build_mobi:
             book.build_mobi()
 
+        if not settings.NO_BUILD_FB2 and build_fb2:
+            book.build_fb2()
+
+        if not settings.NO_SEARCH_INDEX and search_index:
+            book.search_index(index_tags=search_index_tags, reuse_index=search_index_reuse)
+            #index_book.delay(book.id, book_info)
+
         book_descendants = list(book.children.all())
         descendants_tags = set()
         # add l-tag to descendants and their fragments
@@ -824,12 +692,12 @@ class Book(models.Model):
             descendants_tags.update(child_book.tags)
             child_book.tags = list(child_book.tags) + [book_tag]
             child_book.save()
-            for fragment in child_book.fragments.all():
+            for fragment in child_book.fragments.all().iterator():
                 fragment.tags = set(list(fragment.tags) + [book_tag])
             book_descendants += list(child_book.children.all())
 
         for tag in descendants_tags:
-            touch_tag.delay(tag)
+            tasks.touch_tag(tag)
 
         book.save()
 
@@ -840,12 +708,49 @@ class Book(models.Model):
         cls.published.send(sender=book)
         return book
 
+    def related_info(self):
+        """Keeps info about related objects (tags, media) in cache field."""
+        if self._related_info is not None:
+            return self._related_info
+        else:
+            rel = {'tags': {}, 'media': {}}
+
+            tags = self.tags.filter(category__in=(
+                    'author', 'kind', 'genre', 'epoch'))
+            tags = split_tags(tags)
+            for category in tags:
+                rel['tags'][category] = [
+                        (t.name, t.slug) for t in tags[category]]
+
+            for media_format in BookMedia.formats:
+                rel['media'][media_format] = self.has_media(media_format)
+
+            book = self
+            parents = []
+            while book.parent:
+                parents.append((book.parent.title, book.parent.slug))
+                book = book.parent
+            parents = parents[::-1]
+            if parents:
+                rel['parents'] = parents
+
+            if self.pk:
+                type(self).objects.filter(pk=self.pk).update(_related_info=rel)
+            return rel
+
+    def related_themes(self):
+        theme_counter = self.theme_counter
+        book_themes = list(Tag.objects.filter(pk__in=theme_counter.keys()))
+        for tag in book_themes:
+            tag.count = theme_counter[tag.pk]
+        return book_themes
+
     def reset_tag_counter(self):
         if self.id is None:
             return
 
         cache_key = "Book.tag_counter/%d" % self.id
-        cache.delete(cache_key)
+        permanent_cache.delete(cache_key)
         if self.parent:
             self.parent.reset_tag_counter()
 
@@ -853,20 +758,20 @@ class Book(models.Model):
     def tag_counter(self):
         if self.id:
             cache_key = "Book.tag_counter/%d" % self.id
-            tags = cache.get(cache_key)
+            tags = permanent_cache.get(cache_key)
         else:
             tags = None
 
         if tags is None:
             tags = {}
-            for child in self.children.all().order_by():
+            for child in self.children.all().order_by().iterator():
                 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():
+            for tag in self.tags.exclude(category__in=('book', 'theme', 'set')).order_by().iterator():
                 tags[tag.pk] = 1
 
             if self.id:
-                cache.set(cache_key, tags, CACHE_FOREVER)
+                permanent_cache.set(cache_key, tags)
         return tags
 
     def reset_theme_counter(self):
@@ -874,7 +779,7 @@ class Book(models.Model):
             return
 
         cache_key = "Book.theme_counter/%d" % self.id
-        cache.delete(cache_key)
+        permanent_cache.delete(cache_key)
         if self.parent:
             self.parent.reset_theme_counter()
 
@@ -882,18 +787,18 @@ class Book(models.Model):
     def theme_counter(self):
         if self.id:
             cache_key = "Book.theme_counter/%d" % self.id
-            tags = cache.get(cache_key)
+            tags = permanent_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():
+            for fragment in Fragment.tagged.with_any([self.book_tag()]).order_by().iterator():
+                for tag in fragment.tags.filter(category='theme').order_by().iterator():
                     tags[tag.pk] = tags.get(tag.pk, 0) + 1
 
             if self.id:
-                cache.set(cache_key, tags, CACHE_FOREVER)
+                permanent_cache.set(cache_key, tags)
         return tags
 
     def pretty_title(self, html_links=False):
@@ -915,7 +820,7 @@ class Book(models.Model):
 
     @classmethod
     def tagged_top_level(cls, tags):
-        """ Returns top-level books tagged with `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.
@@ -924,8 +829,9 @@ class Book(models.Model):
         # 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)]
+        l_tags = Tag.objects.filter(category='book',
+            slug__in=[book.book_tag_slug() for book in objects.iterator()])
+        descendants_keys = [book.pk for book in cls.tagged.with_any(l_tags).iterator()]
         if descendants_keys:
             objects = objects.exclude(pk__in=descendants_keys)
 
@@ -941,22 +847,23 @@ class Book(models.Model):
 
         books_by_parent = {}
         books = cls.objects.all().order_by('parent_number', 'sort_key').only(
-                'title', 'parent', 'slug', 'language')
+                'title', 'parent', 'slug')
         if filter:
             books = books.filter(filter).distinct()
-            book_ids = set((book.pk for book in books))
-            for book in books:
+            
+            book_ids = set(b['pk'] for b in books.values("pk").iterator())
+            for book in books.iterator():
                 parent = book.parent_id
                 if parent not in book_ids:
                     parent = None
                 books_by_parent.setdefault(parent, []).append(book)
         else:
-            for book in books:
+            for book in books.iterator():
                 books_by_parent.setdefault(book.parent_id, []).append(book)
 
         orphans = []
         books_by_author = SortedDict()
-        for tag in Tag.objects.filter(category='author'):
+        for tag in Tag.objects.filter(category='author').iterator():
             books_by_author[tag] = []
 
         for book in books_by_parent.get(None,()):
@@ -978,14 +885,25 @@ class Book(models.Model):
         "LP": (3, u"liceum"),
     }
     def audiences_pl(self):
-        audiences = self.get_extra_info_value().get('audiences', [])
+        audiences = self.extra_info.get('audiences', [])
         audiences = sorted(set([self._audiences_pl[a] for a in audiences]))
         return [a[1] for a in audiences]
 
+    def choose_fragment(self):
+        tag = self.book_tag()
+        fragments = Fragment.tagged.with_any([tag])
+        if fragments.exists():
+            return fragments.order_by('?')[0]
+        elif self.parent:
+            return self.parent.choose_fragment()
+        else:
+            return None
+
 
 def _has_factory(ftype):
     has = lambda self: bool(getattr(self, "%s_file" % ftype))
-    has.short_description = t.upper()
+    has.short_description = ftype.upper()
+    has.__doc__ = None
     has.boolean = True
     has.__name__ = "has_%s_file" % ftype
     return has
@@ -1002,6 +920,7 @@ for t in Book.formats:
 
 
 class Fragment(models.Model):
+    """Represents a themed fragment of a book."""
     text = models.TextField()
     short_text = models.TextField(editable=False)
     anchor = models.CharField(max_length=120)
@@ -1017,7 +936,7 @@ class Fragment(models.Model):
         verbose_name_plural = _('fragments')
 
     def get_absolute_url(self):
-        return '%s#m%s' % (self.book.get_html_url(), self.anchor)
+        return '%s#m%s' % (reverse('book_text', args=[self.book.slug]), self.anchor)
 
     def reset_short_html(self):
         if self.id is None:
@@ -1025,12 +944,16 @@ class Fragment(models.Model):
 
         cache_key = "Fragment.short_html/%d/%s"
         for lang, langname in settings.LANGUAGES:
-            cache.delete(cache_key % (self.id, lang))
+            permanent_cache.delete(cache_key % (self.id, lang))
+
+    def get_short_text(self):
+        """Returns short version of the fragment."""
+        return self.short_text if self.short_text else self.text
 
     def short_html(self):
         if self.id:
             cache_key = "Fragment.short_html/%d/%s" % (self.id, get_language())
-            short_html = cache.get(cache_key)
+            short_html = permanent_cache.get(cache_key)
         else:
             short_html = None
 
@@ -1040,10 +963,28 @@ class Fragment(models.Model):
             short_html = unicode(render_to_string('catalogue/fragment_short.html',
                 {'fragment': self}))
             if self.id:
-                cache.set(cache_key, short_html, CACHE_FOREVER)
+                permanent_cache.set(cache_key, short_html)
             return mark_safe(short_html)
 
 
+class Collection(models.Model):
+    """A collection of books, which might be defined before publishing them."""
+    title = models.CharField(_('title'), max_length=120, db_index=True)
+    slug = models.SlugField(_('slug'), max_length=120, primary_key=True)
+    description = models.TextField(_('description'), null=True, blank=True)
+
+    models.SlugField(_('slug'), max_length=120, unique=True, db_index=True)
+    book_slugs = models.TextField(_('book slugs'))
+
+    class Meta:
+        ordering = ('title',)
+        verbose_name = _('collection')
+        verbose_name_plural = _('collections')
+
+    def __unicode__(self):
+        return self.title
+
+
 ###########
 #
 # SIGNALS
@@ -1055,7 +996,7 @@ 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
     for tag in affected_tags:
-        touch_tag.delay(tag)
+        tasks.touch_tag(tag)
 
     # if book tags changed, reset book tag counter
     if isinstance(sender, Book) and \
@@ -1076,8 +1017,24 @@ def _pre_delete_handler(sender, instance, **kwargs):
         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)
+
+
+if not settings.NO_SEARCH_INDEX:
+    @django.dispatch.receiver(post_delete, sender=Book)
+    def _remove_book_from_index_handler(sender, instance, **kwargs):
+        """ remove the book from search index, when it is deleted."""
+        import search
+        search.JVM.attachCurrentThread()
+        idx = search.Index()
+        idx.open(timeout=10000)  # 10 seconds timeout.
+        try:
+            idx.remove_book(instance)
+            idx.index_tags()
+        finally:
+            idx.close()