apps/catalogue/fields.py

   1 # -*- coding: utf-8 -*-
   2 # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
   3 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
   4 #
   5 from django.conf import settings
   6 from django.core.files import File
   7 from django.db import models
   8 from django.db.models.fields.files import FieldFile
   9 from catalogue import app_settings
  10 from catalogue.constants import LANGUAGES_3TO2
  11 from catalogue.utils import remove_zip, truncate_html_words
  12 from celery.task import Task, task
  13 from waiter.utils import clear_cache
  14
  15
  16 class EbookFieldFile(FieldFile):
  17     """Represents contents of an ebook file field."""
  18
  19     def build(self):
  20         """Build the ebook immediately."""
  21         return self.field.builder.build(self)
  22
  23     def build_delay(self):
  24         """Builds the ebook in a delayed task."""
  25         return self.field.builder.delay(self.instance, self.field.attname)
  26
  27
  28 class EbookField(models.FileField):
  29     """Represents an ebook file field, attachable to a model."""
  30     attr_class = EbookFieldFile
  31
  32     def __init__(self, format_name, *args, **kwargs):
  33         super(EbookField, self).__init__(*args, **kwargs)
  34         self.format_name = format_name
  35
  36     def deconstruct(self):
  37         name, path, args, kwargs = super(EbookField, self).deconstruct()
  38         args.insert(0, self.format_name)
  39         return name, path, args, kwargs
  40
  41     @property
  42     def builder(self):
  43         """Finds a celery task suitable for the format of the field."""
  44         return BuildEbook.for_format(self.format_name)
  45
  46     def contribute_to_class(self, cls, name):
  47         super(EbookField, self).contribute_to_class(cls, name)
  48
  49         def has(model_instance):
  50             return bool(getattr(model_instance, self.attname, None))
  51         has.__doc__ = None
  52         has.__name__ = str("has_%s" % self.attname)
  53         has.short_description = self.name
  54         has.boolean = True
  55         setattr(cls, 'has_%s' % self.attname, has)
  56
  57
  58 class BuildEbook(Task):
  59     formats = {}
  60
  61     @classmethod
  62     def register(cls, format_name):
  63         """A decorator for registering subclasses for particular formats."""
  64         def wrapper(builder):
  65             cls.formats[format_name] = builder
  66             return builder
  67         return wrapper
  68
  69     @classmethod
  70     def for_format(cls, format_name):
  71         """Returns a celery task suitable for specified format."""
  72         return cls.formats.get(format_name, BuildEbookTask)
  73
  74     @staticmethod
  75     def transform(wldoc, fieldfile):
  76         """Transforms an librarian.WLDocument into an librarian.OutputFile.
  77
  78         By default, it just calls relevant wldoc.as_??? method.
  79
  80         """
  81         return getattr(wldoc, "as_%s" % fieldfile.field.format_name)()
  82
  83     def run(self, obj, field_name):
  84         """Just run `build` on FieldFile, can't pass it directly to Celery."""
  85         return self.build(getattr(obj, field_name))
  86
  87     def build(self, fieldfile):
  88         book = fieldfile.instance
  89         out = self.transform(book.wldocument(), fieldfile)
  90         fieldfile.save(None, File(open(out.get_filename())), save=False)
  91         if book.pk is not None:
  92             type(book).objects.filter(pk=book.pk).update(**{
  93                 fieldfile.field.attname: fieldfile
  94             })
  95         if fieldfile.field.format_name in app_settings.FORMAT_ZIPS:
  96             remove_zip(app_settings.FORMAT_ZIPS[fieldfile.field.format_name])
  97 # Don't decorate BuildEbook, because we want to subclass it.
  98 BuildEbookTask = task(BuildEbook, ignore_result=True)
  99
 100
 101 @BuildEbook.register('txt')
 102 @task(ignore_result=True)
 103 class BuildTxt(BuildEbook):
 104     @staticmethod
 105     def transform(wldoc, fieldfile):
 106         return wldoc.as_text()
 107
 108
 109 @BuildEbook.register('pdf')
 110 @task(ignore_result=True)
 111 class BuildPdf(BuildEbook):
 112     @staticmethod
 113     def transform(wldoc, fieldfile):
 114         return wldoc.as_pdf(morefloats=settings.LIBRARIAN_PDF_MOREFLOATS,
 115             cover=True)
 116
 117     def build(self, fieldfile):
 118         BuildEbook.build(self, fieldfile)
 119         clear_cache(fieldfile.instance.slug)
 120
 121
 122 @BuildEbook.register('epub')
 123 @task(ignore_result=True)
 124 class BuildEpub(BuildEbook):
 125     @staticmethod
 126     def transform(wldoc, fieldfile):
 127         return wldoc.as_epub(cover=True)
 128
 129
 130 @BuildEbook.register('html')
 131 @task(ignore_result=True)
 132 class BuildHtml(BuildEbook):
 133     def build(self, fieldfile):
 134         from django.core.files.base import ContentFile
 135         from fnpdjango.utils.text.slughifi import slughifi
 136         from sortify import sortify
 137         from librarian import html
 138         from catalogue.models import Fragment, Tag
 139
 140         book = fieldfile.instance
 141
 142         html_output = self.transform(
 143                         book.wldocument(parse_dublincore=False),
 144                         fieldfile)
 145
 146         # Delete old fragments, create from scratch if necessary.
 147         book.fragments.all().delete()
 148
 149         if html_output:
 150             meta_tags = list(book.tags.filter(
 151                 category__in=('author', 'epoch', 'genre', 'kind')))
 152
 153             lang = book.language
 154             lang = LANGUAGES_3TO2.get(lang, lang)
 155             if lang not in [ln[0] for ln in settings.LANGUAGES]:
 156                 lang = None
 157
 158             fieldfile.save(None, ContentFile(html_output.get_string()),
 159                     save=False)
 160             type(book).objects.filter(pk=book.pk).update(**{
 161                 fieldfile.field.attname: fieldfile
 162             })
 163
 164             # Extract fragments
 165             closed_fragments, open_fragments = html.extract_fragments(fieldfile.path)
 166             for fragment in closed_fragments.values():
 167                 try:
 168                     theme_names = [s.strip() for s in fragment.themes.split(',')]
 169                 except AttributeError:
 170                     continue
 171                 themes = []
 172                 for theme_name in theme_names:
 173                     if not theme_name:
 174                         continue
 175                     if lang == settings.LANGUAGE_CODE:
 176                         # Allow creating themes if book in default language.
 177                         tag, created = Tag.objects.get_or_create(
 178                                             slug=slughifi(theme_name),
 179                                             category='theme')
 180                         if created:
 181                             tag.name = theme_name
 182                             setattr(tag, "name_%s" % lang, theme_name)
 183                             tag.sort_key = sortify(theme_name.lower())
 184                             tag.save()
 185                         themes.append(tag)
 186                     elif lang is not None:
 187                         # Don't create unknown themes in non-default languages.
 188                         try:
 189                             tag = Tag.objects.get(category='theme',
 190                                     **{"name_%s" % lang: theme_name})
 191                         except Tag.DoesNotExist:
 192                             pass
 193                         else:
 194                             themes.append(tag)
 195                 if not themes:
 196                     continue
 197
 198                 text = fragment.to_string()
 199                 short_text = truncate_html_words(text, 15)
 200                 if text == short_text:
 201                     short_text = ''
 202                 new_fragment = Fragment.objects.create(anchor=fragment.id,
 203                         book=book, text=text, short_text=short_text)
 204
 205                 new_fragment.save()
 206                 new_fragment.tags = set(meta_tags + themes)
 207             book.html_built.send(sender=book)
 208             return True
 209         return False
 210
 211 @BuildEbook.register('cover_thumb')
 212 @task(ignore_result=True)
 213 class BuildCoverThumb(BuildEbook):
 214     @classmethod
 215     def transform(cls, wldoc, fieldfile):
 216         from librarian.cover import WLCover
 217         return WLCover(wldoc.book_info, height=193).output_file()
 218
 219
 220
 221 class OverwritingFieldFile(FieldFile):
 222     """
 223         Deletes the old file before saving the new one.
 224     """
 225
 226     def save(self, name, content, *args, **kwargs):
 227         leave = kwargs.pop('leave', None)
 228         # delete if there's a file already and there's a new one coming
 229         if not leave and self and (not hasattr(content, 'path') or
 230                                    content.path != self.path):
 231             self.delete(save=False)
 232         return super(OverwritingFieldFile, self).save(
 233                 name, content, *args, **kwargs)
 234
 235
 236 class OverwritingFileField(models.FileField):
 237     attr_class = OverwritingFieldFile
 238
 239
 240 try:
 241     # check for south
 242     from south.modelsinspector import add_introspection_rules
 243 except ImportError:
 244     pass
 245 else:
 246     add_introspection_rules([
 247         (
 248             [EbookField],
 249             [],
 250             {'format_name': ('format_name', {})}
 251         )
 252     ], ["^catalogue\.fields\.EbookField"])
 253     add_introspection_rules([], ["^catalogue\.fields\.OverwritingFileField"])