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.
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, gallery_path
12 from celery.task import Task, task
13 from celery.utils.log import get_task_logger
14 from waiter.utils import clear_cache
16 task_logger = get_task_logger(__name__)
19 class EbookFieldFile(FieldFile):
20 """Represents contents of an ebook file field."""
23 """Build the ebook immediately."""
24 return self.field.builder.build(self)
26 def build_delay(self):
27 """Builds the ebook in a delayed task."""
28 return self.field.builder.delay(self.instance, self.field.attname)
31 class EbookField(models.FileField):
32 """Represents an ebook file field, attachable to a model."""
33 attr_class = EbookFieldFile
35 def __init__(self, format_name, *args, **kwargs):
36 super(EbookField, self).__init__(*args, **kwargs)
37 self.format_name = format_name
39 def deconstruct(self):
40 name, path, args, kwargs = super(EbookField, self).deconstruct()
41 args.insert(0, self.format_name)
42 return name, path, args, kwargs
46 """Finds a celery task suitable for the format of the field."""
47 return BuildEbook.for_format(self.format_name)
49 def contribute_to_class(self, cls, name):
50 super(EbookField, self).contribute_to_class(cls, name)
52 def has(model_instance):
53 return bool(getattr(model_instance, self.attname, None))
55 has.__name__ = str("has_%s" % self.attname)
56 has.short_description = self.name
58 setattr(cls, 'has_%s' % self.attname, has)
61 class BuildEbook(Task):
65 def register(cls, format_name):
66 """A decorator for registering subclasses for particular formats."""
68 cls.formats[format_name] = builder
73 def for_format(cls, format_name):
74 """Returns a celery task suitable for specified format."""
75 return cls.formats.get(format_name, BuildEbookTask)
78 def transform(wldoc, fieldfile):
79 """Transforms an librarian.WLDocument into an librarian.OutputFile.
81 By default, it just calls relevant wldoc.as_??? method.
84 return getattr(wldoc, "as_%s" % fieldfile.field.format_name)()
86 def run(self, obj, field_name):
87 """Just run `build` on FieldFile, can't pass it directly to Celery."""
88 task_logger.info("%s -> %s" % (obj.slug, field_name))
89 ret = self.build(getattr(obj, field_name))
93 def build(self, fieldfile):
94 book = fieldfile.instance
95 out = self.transform(book.wldocument(), fieldfile)
96 fieldfile.save(None, File(open(out.get_filename())), save=False)
97 if book.pk is not None:
98 type(book).objects.filter(pk=book.pk).update(**{
99 fieldfile.field.attname: fieldfile
101 if fieldfile.field.format_name in app_settings.FORMAT_ZIPS:
102 remove_zip(app_settings.FORMAT_ZIPS[fieldfile.field.format_name])
103 # Don't decorate BuildEbook, because we want to subclass it.
104 BuildEbookTask = task(BuildEbook, ignore_result=True)
107 @BuildEbook.register('txt')
108 @task(ignore_result=True)
109 class BuildTxt(BuildEbook):
111 def transform(wldoc, fieldfile):
112 return wldoc.as_text()
115 @BuildEbook.register('pdf')
116 @task(ignore_result=True)
117 class BuildPdf(BuildEbook):
119 def transform(wldoc, fieldfile):
120 return wldoc.as_pdf(morefloats=settings.LIBRARIAN_PDF_MOREFLOATS, cover=True,
121 ilustr_path=gallery_path(wldoc.book_info.url.slug))
123 def build(self, fieldfile):
124 BuildEbook.build(self, fieldfile)
125 clear_cache(fieldfile.instance.slug)
128 @BuildEbook.register('epub')
129 @task(ignore_result=True)
130 class BuildEpub(BuildEbook):
132 def transform(wldoc, fieldfile):
133 return wldoc.as_epub(cover=True)
136 @BuildEbook.register('html')
137 @task(ignore_result=True)
138 class BuildHtml(BuildEbook):
139 def build(self, fieldfile):
140 from django.core.files.base import ContentFile
141 from fnpdjango.utils.text.slughifi import slughifi
142 from sortify import sortify
143 from librarian import html
144 from catalogue.models import Fragment, Tag
146 book = fieldfile.instance
148 html_output = self.transform(
149 book.wldocument(parse_dublincore=False),
152 # Delete old fragments, create from scratch if necessary.
153 book.fragments.all().delete()
156 meta_tags = list(book.tags.filter(
157 category__in=('author', 'epoch', 'genre', 'kind')))
160 lang = LANGUAGES_3TO2.get(lang, lang)
161 if lang not in [ln[0] for ln in settings.LANGUAGES]:
164 fieldfile.save(None, ContentFile(html_output.get_string()), save=False)
165 type(book).objects.filter(pk=book.pk).update(**{
166 fieldfile.field.attname: fieldfile
170 closed_fragments, open_fragments = html.extract_fragments(fieldfile.path)
171 for fragment in closed_fragments.values():
173 theme_names = [s.strip() for s in fragment.themes.split(',')]
174 except AttributeError:
177 for theme_name in theme_names:
180 if lang == settings.LANGUAGE_CODE:
181 # Allow creating themes if book in default language.
182 tag, created = Tag.objects.get_or_create(
183 slug=slughifi(theme_name),
186 tag.name = theme_name
187 setattr(tag, "name_%s" % lang, theme_name)
188 tag.sort_key = sortify(theme_name.lower())
191 elif lang is not None:
192 # Don't create unknown themes in non-default languages.
194 tag = Tag.objects.get(category='theme', **{"name_%s" % lang: theme_name})
195 except Tag.DoesNotExist:
202 text = fragment.to_string()
203 short_text = truncate_html_words(text, 15)
204 if text == short_text:
206 new_fragment = Fragment.objects.create(anchor=fragment.id, book=book, text=text, short_text=short_text)
209 new_fragment.tags = set(meta_tags + themes)
210 book.html_built.send(sender=type(self), instance=book)
215 @BuildEbook.register('cover_thumb')
216 @task(ignore_result=True)
217 class BuildCoverThumb(BuildEbook):
219 def transform(cls, wldoc, fieldfile):
220 from librarian.cover import WLCover
221 return WLCover(wldoc.book_info, height=193).output_file()
224 class OverwritingFieldFile(FieldFile):
226 Deletes the old file before saving the new one.
229 def save(self, name, content, *args, **kwargs):
230 leave = kwargs.pop('leave', None)
231 # delete if there's a file already and there's a new one coming
232 if not leave and self and (not hasattr(content, 'path') or content.path != self.path):
233 self.delete(save=False)
234 return super(OverwritingFieldFile, self).save(name, content, *args, **kwargs)
237 class OverwritingFileField(models.FileField):
238 attr_class = OverwritingFieldFile