Search engine now should be db change change-proof
[wolnelektury.git] / apps / catalogue / models.py
index e4cd8c4..ea09c04 100644 (file)
@@ -7,7 +7,7 @@ from collections import namedtuple
 from django.db import models
 from django.db.models import permalink, Q
 import django.dispatch
-from django.core.cache import cache
+from django.core.cache import get_cache
 from django.core.files.storage import DefaultStorage
 from django.utils.translation import ugettext_lazy as _
 from django.contrib.auth.models import User
@@ -16,23 +16,24 @@ 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, m2m_changed, 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, split_tags
+from catalogue.utils import create_zip, split_tags, truncate_html_words
 from catalogue.tasks import touch_tag, index_book
 from shutil import copy
 from glob import glob
 import re
 from os import path
 
-
 import search
 
+# Those are hard-coded here so that makemessages sees them.
 TAG_CATEGORIES = (
     ('author', _('author')),
     ('epoch', _('epoch')),
@@ -43,8 +44,8 @@ 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):
@@ -215,14 +216,17 @@ def book_upload_path(ext=None, maxlen=100):
     return lambda *args: get_dynamic_path(*args, ext=ext, maxlen=maxlen)
 
 
+def customizations_hash(customizations):
+    customizations.sort()
+    return hash(tuple(customizations))
+
+
 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)
+    h = customizations_hash(customizations)
+    pdf_name = '%s-custom-%s' % (book.slug, h)
     pdf_file = get_dynamic_path(None, pdf_name, ext='pdf')
 
     return pdf_file
@@ -232,7 +236,7 @@ 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 = '%s-custom-' % (book.slug,)
     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))
@@ -271,7 +275,7 @@ class BookMedia(models.Model):
         try:
             old = BookMedia.objects.get(pk=self.pk)
         except BookMedia.DoesNotExist, e:
-            pass
+            old = None
         else:
             # if name changed, change the file name, too
             if slughifi(self.name) != slughifi(old.name):
@@ -280,7 +284,9 @@ 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.update(self.read_meta())
@@ -349,7 +355,9 @@ class BookMedia(models.Model):
 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, 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)
@@ -367,6 +375,9 @@ class Book(models.Model):
     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)
@@ -374,14 +385,10 @@ 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')
@@ -389,42 +396,6 @@ 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
 
@@ -439,18 +410,14 @@ 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()
@@ -489,58 +456,11 @@ 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))
-        cache.delete("Book.mini_box/%d" % (self.id, ))
+        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():
             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(category__in=('author', 'kind', 'genre', 'epoch'))
-            tags = split_tags(tags)
-
-            formats = {}
-            # files generated during publication
-            for ebook_format in self.ebook_formats:
-                if self.has_media(ebook_format):
-                    formats[ebook_format] = self.get_media(ebook_format)
-
-
-            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_box/%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')
@@ -563,7 +483,7 @@ 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,
@@ -599,7 +519,7 @@ class Book(models.Model):
             # 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(),
+            current_self.pdf_file.save('%s.pdf' % self.slug,
                     File(open(pdf.get_filename())))
             self.pdf_file = current_self.pdf_file
 
@@ -622,7 +542,7 @@ class Book(models.Model):
 
         mobi = self.wldocument().as_mobi()
 
-        self.mobi_file.save('%s.mobi' % self.fileid(), File(open(mobi.get_filename())))
+        self.mobi_file.save('%s.mobi' % self.slug, File(open(mobi.get_filename())))
 
         # remove zip with all mobi files
         remove_zip(settings.ALL_MOBI_ZIP)
@@ -634,7 +554,7 @@ class Book(models.Model):
 
         epub = self.wldocument().as_epub()
 
-        self.epub_file.save('%s.epub' % self.fileid(),
+        self.epub_file.save('%s.epub' % self.slug,
                 File(open(epub.get_filename())))
 
         # remove zip package with all epub files
@@ -644,11 +564,10 @@ class Book(models.Model):
         from django.core.files.base import ContentFile
 
         text = self.wldocument().as_text()
-        self.txt_file.save('%s.txt' % self.fileid(), ContentFile(text.get_string()))
+        self.txt_file.save('%s.txt' % self.slug, ContentFile(text.get_string()))
 
 
     def build_html(self):
-        from markupstring import MarkupString
         from django.core.files.base import ContentFile
         from slughifi import slughifi
         from librarian import html
@@ -659,7 +578,7 @@ 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
@@ -692,9 +611,9 @@ class Book(models.Model):
                     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)
 
@@ -710,7 +629,7 @@ class Book(models.Model):
         def pretty_file_name(book):
             return "%s/%s.%s" % (
                 b.get_extra_info_value()['author'],
-                b.fileid(),
+                b.slug,
                 format_)
 
         field_name = "%s_file" % format_
@@ -721,14 +640,14 @@ class Book(models.Model):
                     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())
+        result = create_zip.delay(paths, "%s_%s" % (self.slug, format_))
         return result.wait()
 
-    def search_index(self, book_info=None):
-        if settings.CELERY_ALWAYS_EAGER:
+    def search_index(self, book_info=None, reuse_index=False, index_tags=True):
+        if reuse_index:
             idx = search.ReusableIndex()
         else:
             idx = search.Index()
@@ -736,7 +655,8 @@ class Book(models.Model):
         idx.open()
         try:
             idx.index_book(self, book_info)
-            idx.index_tags()
+            if index_tags:
+                idx.index_tags()
         finally:
             idx.close()
 
@@ -759,7 +679,7 @@ 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,
-            search_index=True):
+            search_index=True, search_index_tags=True, search_index_reuse=False):
         import re
         from sortify import sortify
 
@@ -768,30 +688,33 @@ class Book(models.Model):
         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))
+                    children.append(Book.objects.get(slug=part_url.slug))
                 except Book.DoesNotExist, e:
-                    raise Book.DoesNotExist(_('Book "%s/%s" does not exist.') %
-                            (part_url.slug, part_url.language))
+                    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
+        if book_info.variant_of:
+            book.common_slug = book_info.variant_of.slug
+        else:
+            book.common_slug = book.slug
         book.set_extra_info_value(book_info.to_dict())
         book.save()
 
@@ -828,7 +751,8 @@ class Book(models.Model):
             book.build_mobi()
 
         if not settings.NO_SEARCH_INDEX and search_index:
-            index_book.delay(book.id, book_info)
+            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()
@@ -843,7 +767,7 @@ class Book(models.Model):
             book_descendants += list(child_book.children.all())
 
         for tag in descendants_tags:
-            touch_tag.delay(tag)
+            touch_tag(tag)
 
         book.save()
 
@@ -854,12 +778,42 @@ 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 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()
 
@@ -867,7 +821,7 @@ 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
 
@@ -880,7 +834,7 @@ class Book(models.Model):
                 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):
@@ -888,7 +842,7 @@ 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()
 
@@ -896,7 +850,7 @@ 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
 
@@ -907,7 +861,7 @@ class Book(models.Model):
                     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):
@@ -955,7 +909,7 @@ 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))
@@ -996,6 +950,16 @@ class Book(models.Model):
         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))
@@ -1039,12 +1003,16 @@ 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
 
@@ -1054,10 +1022,28 @@ 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
@@ -1069,7 +1055,7 @@ 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)
+        touch_tag(tag)
 
     # if book tags changed, reset book tag counter
     if isinstance(sender, Book) and \
@@ -1090,8 +1076,22 @@ 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)
+
+
+@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."""
+    search.JVM.attachCurrentThread()
+    idx = search.Index()
+    idx.open(timeout=10000)  # 10 seconds timeout.
+    try:
+        idx.remove_book(instance)
+        idx.index_tags()
+    finally:
+        idx.close()