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         ret = self.build(getattr(obj, field_name))
  86         obj.flush_includes()
  87         return ret
  88
  89     def build(self, fieldfile):
  90         book = fieldfile.instance
  91         out = self.transform(book.wldocument(), fieldfile)
  92         fieldfile.save(None, File(open(out.get_filename())), save=False)
  93         if book.pk is not None:
  94             type(book).objects.filter(pk=book.pk).update(**{
  95                 fieldfile.field.attname: fieldfile
  96             })
  97         if fieldfile.field.format_name in app_settings.FORMAT_ZIPS:
  98             remove_zip(app_settings.FORMAT_ZIPS[fieldfile.field.format_name])
  99 # Don't decorate BuildEbook, because we want to subclass it.
 100 BuildEbookTask = task(BuildEbook, ignore_result=True)
 101
 102
 103 @BuildEbook.register('txt')
 104 @task(ignore_result=True)
 105 class BuildTxt(BuildEbook):
 106     @staticmethod
 107     def transform(wldoc, fieldfile):
 108         return wldoc.as_text()
 109
 110
 111 @BuildEbook.register('pdf')
 112 @task(ignore_result=True)
 113 class BuildPdf(BuildEbook):
 114     @staticmethod
 115     def transform(wldoc, fieldfile):
 116         return wldoc.as_pdf(morefloats=settings.LIBRARIAN_PDF_MOREFLOATS,
 117             cover=True)
 118
 119     def build(self, fieldfile):
 120         BuildEbook.build(self, fieldfile)
 121         clear_cache(fieldfile.instance.slug)
 122
 123
 124 @BuildEbook.register('epub')
 125 @task(ignore_result=True)
 126 class BuildEpub(BuildEbook):
 127     @staticmethod
 128     def transform(wldoc, fieldfile):
 129         return wldoc.as_epub(cover=True)
 130
 131
 132 @BuildEbook.register('html')
 133 @task(ignore_result=True)
 134 class BuildHtml(BuildEbook):
 135     def build(self, fieldfile):
 136         from django.core.files.base import ContentFile
 137         from fnpdjango.utils.text.slughifi import slughifi
 138         from sortify import sortify
 139         from librarian import html
 140         from catalogue.models import Fragment, Tag
 141
 142         book = fieldfile.instance
 143
 144         html_output = self.transform(
 145                         book.wldocument(parse_dublincore=False),
 146                         fieldfile)
 147
 148         # Delete old fragments, create from scratch if necessary.
 149         book.fragments.all().delete()
 150
 151         if html_output:
 152             meta_tags = list(book.tags.filter(
 153                 category__in=('author', 'epoch', 'genre', 'kind')))
 154
 155             lang = book.language
 156             lang = LANGUAGES_3TO2.get(lang, lang)
 157             if lang not in [ln[0] for ln in settings.LANGUAGES]:
 158                 lang = None
 159
 160             fieldfile.save(None, ContentFile(html_output.get_string()),
 161                     save=False)
 162             type(book).objects.filter(pk=book.pk).update(**{
 163                 fieldfile.field.attname: fieldfile
 164             })
 165
 166             # Extract fragments
 167             closed_fragments, open_fragments = html.extract_fragments(fieldfile.path)
 168             for fragment in closed_fragments.values():
 169                 try:
 170                     theme_names = [s.strip() for s in fragment.themes.split(',')]
 171                 except AttributeError:
 172                     continue
 173                 themes = []
 174                 for theme_name in theme_names:
 175                     if not theme_name:
 176                         continue
 177                     if lang == settings.LANGUAGE_CODE:
 178                         # Allow creating themes if book in default language.
 179                         tag, created = Tag.objects.get_or_create(
 180                                             slug=slughifi(theme_name),
 181                                             category='theme')
 182                         if created:
 183                             tag.name = theme_name
 184                             setattr(tag, "name_%s" % lang, theme_name)
 185                             tag.sort_key = sortify(theme_name.lower())
 186                             tag.save()
 187                         themes.append(tag)
 188                     elif lang is not None:
 189                         # Don't create unknown themes in non-default languages.
 190                         try:
 191                             tag = Tag.objects.get(category='theme',
 192                                     **{"name_%s" % lang: theme_name})
 193                         except Tag.DoesNotExist:
 194                             pass
 195                         else:
 196                             themes.append(tag)
 197                 if not themes:
 198                     continue
 199
 200                 text = fragment.to_string()
 201                 short_text = truncate_html_words(text, 15)
 202                 if text == short_text:
 203                     short_text = ''
 204                 new_fragment = Fragment.objects.create(anchor=fragment.id,
 205                         book=book, text=text, short_text=short_text)
 206
 207                 new_fragment.save()
 208                 new_fragment.tags = set(meta_tags + themes)
 209             book.html_built.send(sender=type(self), instance=book)
 210             return True
 211         return False
 212
 213 @BuildEbook.register('cover_thumb')
 214 @task(ignore_result=True)
 215 class BuildCoverThumb(BuildEbook):
 216     @classmethod
 217     def transform(cls, wldoc, fieldfile):
 218         from librarian.cover import WLCover
 219         return WLCover(wldoc.book_info, height=193).output_file()
 220
 221
 222
 223 class OverwritingFieldFile(FieldFile):
 224     """
 225         Deletes the old file before saving the new one.
 226     """
 227
 228     def save(self, name, content, *args, **kwargs):
 229         leave = kwargs.pop('leave', None)
 230         # delete if there's a file already and there's a new one coming
 231         if not leave and self and (not hasattr(content, 'path') or
 232                                    content.path != self.path):
 233             self.delete(save=False)
 234         return super(OverwritingFieldFile, self).save(
 235                 name, content, *args, **kwargs)
 236
 237
 238 class OverwritingFileField(models.FileField):
 239     attr_class = OverwritingFieldFile