src/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.core.files.storage import FileSystemStorage
   8 from django.db import models
   9 from django.db.models.fields.files import FieldFile
  10 from catalogue import app_settings
  11 from catalogue.constants import LANGUAGES_3TO2
  12 from catalogue.utils import remove_zip, truncate_html_words, gallery_path, gallery_url
  13 from celery.task import Task, task
  14 from celery.utils.log import get_task_logger
  15 from waiter.utils import clear_cache
  16
  17 task_logger = get_task_logger(__name__)
  18
  19
  20 class EbookFieldFile(FieldFile):
  21     """Represents contents of an ebook file field."""
  22
  23     def build(self):
  24         """Build the ebook immediately."""
  25         return self.field.builder.build(self)
  26
  27     def build_delay(self):
  28         """Builds the ebook in a delayed task."""
  29         return self.field.builder.delay(self.instance, self.field.attname)
  30
  31
  32 class EbookField(models.FileField):
  33     """Represents an ebook file field, attachable to a model."""
  34     attr_class = EbookFieldFile
  35
  36     def __init__(self, format_name, *args, **kwargs):
  37         super(EbookField, self).__init__(*args, **kwargs)
  38         self.format_name = format_name
  39
  40     def deconstruct(self):
  41         name, path, args, kwargs = super(EbookField, self).deconstruct()
  42         args.insert(0, self.format_name)
  43         return name, path, args, kwargs
  44
  45     @property
  46     def builder(self):
  47         """Finds a celery task suitable for the format of the field."""
  48         return BuildEbook.for_format(self.format_name)
  49
  50     def contribute_to_class(self, cls, name):
  51         super(EbookField, self).contribute_to_class(cls, name)
  52
  53         def has(model_instance):
  54             return bool(getattr(model_instance, self.attname, None))
  55         has.__doc__ = None
  56         has.__name__ = str("has_%s" % self.attname)
  57         has.short_description = self.name
  58         has.boolean = True
  59         setattr(cls, 'has_%s' % self.attname, has)
  60
  61
  62 class BuildEbook(Task):
  63     formats = {}
  64
  65     @classmethod
  66     def register(cls, format_name):
  67         """A decorator for registering subclasses for particular formats."""
  68         def wrapper(builder):
  69             cls.formats[format_name] = builder
  70             return builder
  71         return wrapper
  72
  73     @classmethod
  74     def for_format(cls, format_name):
  75         """Returns a celery task suitable for specified format."""
  76         return cls.formats.get(format_name, BuildEbookTask)
  77
  78     @staticmethod
  79     def transform(wldoc, fieldfile):
  80         """Transforms an librarian.WLDocument into an librarian.OutputFile.
  81
  82         By default, it just calls relevant wldoc.as_??? method.
  83
  84         """
  85         return getattr(wldoc, "as_%s" % fieldfile.field.format_name)()
  86
  87     def run(self, obj, field_name):
  88         """Just run `build` on FieldFile, can't pass it directly to Celery."""
  89         task_logger.info("%s -> %s" % (obj.slug, field_name))
  90         ret = self.build(getattr(obj, field_name))
  91         obj.flush_includes()
  92         return ret
  93
  94     def build(self, fieldfile):
  95         book = fieldfile.instance
  96         out = self.transform(book.wldocument(), fieldfile)
  97         fieldfile.save(None, File(open(out.get_filename())), save=False)
  98         if book.pk is not None:
  99             type(book).objects.filter(pk=book.pk).update(**{
 100                 fieldfile.field.attname: fieldfile
 101             })
 102         if fieldfile.field.format_name in app_settings.FORMAT_ZIPS:
 103             remove_zip(app_settings.FORMAT_ZIPS[fieldfile.field.format_name])
 104 # Don't decorate BuildEbook, because we want to subclass it.
 105 BuildEbookTask = task(BuildEbook, ignore_result=True)
 106
 107
 108 @BuildEbook.register('txt')
 109 @task(ignore_result=True)
 110 class BuildTxt(BuildEbook):
 111     @staticmethod
 112     def transform(wldoc, fieldfile):
 113         return wldoc.as_text()
 114
 115
 116 @BuildEbook.register('pdf')
 117 @task(ignore_result=True)
 118 class BuildPdf(BuildEbook):
 119     @staticmethod
 120     def transform(wldoc, fieldfile):
 121         return wldoc.as_pdf(morefloats=settings.LIBRARIAN_PDF_MOREFLOATS, cover=True,
 122                             ilustr_path=gallery_path(wldoc.book_info.url.slug))
 123
 124     def build(self, fieldfile):
 125         BuildEbook.build(self, fieldfile)
 126         clear_cache(fieldfile.instance.slug)
 127
 128
 129 @BuildEbook.register('epub')
 130 @task(ignore_result=True)
 131 class BuildEpub(BuildEbook):
 132     @staticmethod
 133     def transform(wldoc, fieldfile):
 134         return wldoc.as_epub(cover=True, ilustr_path=gallery_path(wldoc.book_info.url.slug))
 135
 136
 137 @BuildEbook.register('mobi')
 138 @task(ignore_result=True)
 139 class BuildMobi(BuildEbook):
 140     @staticmethod
 141     def transform(wldoc, fieldfile):
 142         return wldoc.as_mobi(cover=True, ilustr_path=gallery_path(wldoc.book_info.url.slug))
 143
 144
 145 @BuildEbook.register('html')
 146 @task(ignore_result=True)
 147 class BuildHtml(BuildEbook):
 148     def build(self, fieldfile):
 149         from django.core.files.base import ContentFile
 150         from fnpdjango.utils.text.slughifi import slughifi
 151         from sortify import sortify
 152         from librarian import html
 153         from catalogue.models import Fragment, Tag
 154
 155         book = fieldfile.instance
 156
 157         html_output = self.transform(book.wldocument(parse_dublincore=False), fieldfile)
 158
 159         # Delete old fragments, create from scratch if necessary.
 160         book.fragments.all().delete()
 161
 162         if html_output:
 163             meta_tags = list(book.tags.filter(
 164                 category__in=('author', 'epoch', 'genre', 'kind')))
 165
 166             lang = book.language
 167             lang = LANGUAGES_3TO2.get(lang, lang)
 168             if lang not in [ln[0] for ln in settings.LANGUAGES]:
 169                 lang = None
 170
 171             fieldfile.save(None, ContentFile(html_output.get_string()), save=False)
 172             type(book).objects.filter(pk=book.pk).update(**{
 173                 fieldfile.field.attname: fieldfile
 174             })
 175
 176             # Extract fragments
 177             closed_fragments, open_fragments = html.extract_fragments(fieldfile.path)
 178             for fragment in closed_fragments.values():
 179                 try:
 180                     theme_names = [s.strip() for s in fragment.themes.split(',')]
 181                 except AttributeError:
 182                     continue
 183                 themes = []
 184                 for theme_name in theme_names:
 185                     if not theme_name:
 186                         continue
 187                     if lang == settings.LANGUAGE_CODE:
 188                         # Allow creating themes if book in default language.
 189                         tag, created = Tag.objects.get_or_create(
 190                                             slug=slughifi(theme_name),
 191                                             category='theme')
 192                         if created:
 193                             tag.name = theme_name
 194                             setattr(tag, "name_%s" % lang, theme_name)
 195                             tag.sort_key = sortify(theme_name.lower())
 196                             tag.save()
 197                         themes.append(tag)
 198                     elif lang is not None:
 199                         # Don't create unknown themes in non-default languages.
 200                         try:
 201                             tag = Tag.objects.get(category='theme', **{"name_%s" % lang: theme_name})
 202                         except Tag.DoesNotExist:
 203                             pass
 204                         else:
 205                             themes.append(tag)
 206                 if not themes:
 207                     continue
 208
 209                 text = fragment.to_string()
 210                 short_text = truncate_html_words(text, 15)
 211                 if text == short_text:
 212                     short_text = ''
 213                 new_fragment = Fragment.objects.create(anchor=fragment.id, book=book, text=text, short_text=short_text)
 214
 215                 new_fragment.save()
 216                 new_fragment.tags = set(meta_tags + themes)
 217             book.html_built.send(sender=type(self), instance=book)
 218             return True
 219         return False
 220
 221     @staticmethod
 222     def transform(wldoc, fieldfile):
 223         # ugly, but we can't use wldoc.book_info here
 224         from librarian import DCNS
 225         url_elem = wldoc.edoc.getroot().find('.//' + DCNS('identifier.url'))
 226         if url_elem is None:
 227             gallery = ''
 228         else:
 229             gallery = gallery_url(slug=url_elem.text.rsplit('/', 1)[1])
 230         return wldoc.as_html(options={'gallery': "'%s'" % gallery})
 231
 232
 233 @BuildEbook.register('cover_thumb')
 234 @task(ignore_result=True)
 235 class BuildCoverThumb(BuildEbook):
 236     @classmethod
 237     def transform(cls, wldoc, fieldfile):
 238         from librarian.cover import WLCover
 239         return WLCover(wldoc.book_info, height=193).output_file()
 240
 241
 242 # not used, but needed for migrations
 243 class OverwritingFieldFile(FieldFile):
 244     """
 245         Deletes the old file before saving the new one.
 246     """
 247
 248     def save(self, name, content, *args, **kwargs):
 249         leave = kwargs.pop('leave', None)
 250         # delete if there's a file already and there's a new one coming
 251         if not leave and self and (not hasattr(content, 'path') or content.path != self.path):
 252             self.delete(save=False)
 253         return super(OverwritingFieldFile, self).save(name, content, *args, **kwargs)
 254
 255
 256 class OverwritingFileField(models.FileField):
 257     attr_class = OverwritingFieldFile
 258
 259
 260 class OverwriteStorage(FileSystemStorage):
 261
 262     def get_available_name(self, name, max_length=None):
 263         self.delete(name)
 264         return name