Don't break catalogue.pdf on unknown dc:audience.
[wolnelektury.git] / apps / catalogue / fields.py
index 654dbfb..bb9f26e 100644 (file)
 # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
 #
 # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
 #
-import datetime
-
 from django.conf import settings
 from django.conf import settings
+from django.core.files import File
 from django.db import models
 from django.db import models
-from django.db.models import signals
-from django import forms
-from django.forms.widgets import flatatt
-from django.forms.util import smart_unicode
-from django.utils import simplejson as json
-from django.utils.html import escape
-from django.utils.safestring import mark_safe
-
-
-class JSONEncoder(json.JSONEncoder):
-    def default(self, obj):
-        if isinstance(obj, datetime.datetime):
-            return obj.strftime('%Y-%m-%d %H:%M:%S')
-        elif isinstance(obj, datetime.date):
-            return obj.strftime('%Y-%m-%d')
-        elif isinstance(obj, datetime.time):
-            return obj.strftime('%H:%M:%S')
-        return json.JSONEncoder.default(self, obj)
+from django.db.models.fields.files import FieldFile
+from catalogue import app_settings
+from catalogue.utils import remove_zip, truncate_html_words
+from celery.task import Task, task
+from waiter.utils import clear_cache
 
 
 
 
-def dumps(data):
-    return JSONEncoder().encode(data)
+class EbookFieldFile(FieldFile):
+    """Represents contents of an ebook file field."""
 
 
+    def build(self):
+        """Build the ebook immediately."""
+        return self.field.builder.build(self)
 
 
-def loads(str):
-    return json.loads(str, encoding=settings.DEFAULT_CHARSET)
+    def build_delay(self):
+        """Builds the ebook in a delayed task."""
+        return self.field.builder.delay(self.instance, self.field.attname)
 
 
 
 
-class JSONFormField(forms.CharField):
-    widget = forms.Textarea
-    
-    def clean(self, value):
-        try:
-            loads(value)
-            return value
-        except ValueError, e:
-            raise forms.ValidationError(_('Enter a valid JSON value. Error: %s') % e)
+class EbookField(models.FileField):
+    """Represents an ebook file field, attachable to a model."""
+    attr_class = EbookFieldFile
 
 
+    def __init__(self, format_name, *args, **kwargs):
+        super(EbookField, self).__init__(*args, **kwargs)
+        self.format_name = format_name
 
 
-class JSONField(models.TextField):
-    def formfield(self, **kwargs):
-        defaults = {'form_class': JSONFormField}
-        defaults.update(kwargs)
-        return super(JSONField, self).formfield(**defaults)
-
-    def db_type(self):
-        return 'text'
-
-    def get_internal_type(self):
-        return 'TextField'
+    @property
+    def builder(self):
+        """Finds a celery task suitable for the format of the field."""
+        return BuildEbook.for_format(self.format_name)
 
     def contribute_to_class(self, cls, name):
 
     def contribute_to_class(self, cls, name):
-        super(JSONField, self).contribute_to_class(cls, name)
-        
-        def get_value(model_instance):
-            return loads(getattr(model_instance, self.attname, None))
-        setattr(cls, 'get_%s_value' % self.name, get_value)
-
-        def set_value(model_instance, json):
-            return setattr(model_instance, self.attname, dumps(json))
-        setattr(cls, 'set_%s_value' % self.name, set_value)
-
-
-class JQueryAutoCompleteWidget(forms.TextInput):
-    def __init__(self, source, options=None, *args, **kwargs):
-        self.source = source
-        self.options = None
-        if options:
-            self.options = dumps(options)
-        super(JQueryAutoCompleteWidget, self).__init__(*args, **kwargs)
-    
-    def render_js(self, field_id):
-        source = "'%s'" % escape(self.source)
-        options = ''
-        if self.options:
-            options += ', %s' % self.options
-        
-        return u'$(\'#%s\').autocomplete(%s%s).result(autocomplete_result_handler);' % (field_id, source, options)
-    
-    def render(self, name, value=None, attrs=None):
-        final_attrs = self.build_attrs(attrs, name=name)
-        if value:
-            final_attrs['value'] = smart_unicode(value)
-        
-        if not self.attrs.has_key('id'):
-            final_attrs['id'] = 'id_%s' % name
-        
-        html = u'''<input type="text" %(attrs)s/>
-            <script type="text/javascript">//<!--
-            %(js)s//--></script>
-            ''' % {
-                'attrs' : flatatt(final_attrs),
-                'js' : self.render_js(final_attrs['id']),
-            }
-        
-        return mark_safe(html)
-
-
-class JQueryAutoCompleteField(forms.CharField):
-    def __init__(self, source, options=None, *args, **kwargs):
-        if 'widget' not in kwargs:
-            kwargs['widget'] = JQueryAutoCompleteWidget(source, options)
-        
-        super(JQueryAutoCompleteField, self).__init__(*args, **kwargs)
-
+        super(EbookField, self).contribute_to_class(cls, name)
+
+        def has(model_instance):
+            return bool(getattr(model_instance, self.attname, None))
+        has.__doc__ = None
+        has.__name__ = "has_%s" % self.attname
+        has.short_description = self.name
+        has.boolean = True
+        setattr(cls, 'has_%s' % self.attname, has)
+
+
+class BuildEbook(Task):
+    formats = {}
+
+    @classmethod
+    def register(cls, format_name):
+        """A decorator for registering subclasses for particular formats."""
+        def wrapper(builder):
+            cls.formats[format_name] = builder
+            return builder
+        return wrapper
+
+    @classmethod
+    def for_format(cls, format_name):
+        """Returns a celery task suitable for specified format."""
+        return cls.formats.get(format_name, BuildEbookTask)
+
+    @staticmethod
+    def transform(wldoc, fieldfile):
+        """Transforms an librarian.WLDocument into an librarian.OutputFile.
+
+        By default, it just calls relevant wldoc.as_??? method.
+
+        """
+        return getattr(wldoc, "as_%s" % fieldfile.field.format_name)()
+
+    def run(self, obj, field_name):
+        """Just run `build` on FieldFile, can't pass it directly to Celery."""
+        return self.build(getattr(obj, field_name))
+
+    def build(self, fieldfile):
+        book = fieldfile.instance
+        out = self.transform(book.wldocument(), fieldfile)
+        fieldfile.save(None, File(open(out.get_filename())), save=False)
+        if book.pk is not None:
+            type(book).objects.filter(pk=book.pk).update(**{
+                fieldfile.field.attname: fieldfile
+            })
+        if fieldfile.field.format_name in app_settings.FORMAT_ZIPS:
+            remove_zip(app_settings.FORMAT_ZIPS[fieldfile.field.format_name])
+# Don't decorate BuildEbook, because we want to subclass it.
+BuildEbookTask = task(BuildEbook, ignore_result=True)
+
+
+@BuildEbook.register('txt')
+@task(ignore_result=True)
+class BuildTxt(BuildEbook):
+    @staticmethod
+    def transform(wldoc, fieldfile):
+        return wldoc.as_text()
+
+
+@BuildEbook.register('pdf')
+@task(ignore_result=True)
+class BuildPdf(BuildEbook):
+    @staticmethod
+    def transform(wldoc, fieldfile):
+        return wldoc.as_pdf(morefloats=settings.LIBRARIAN_PDF_MOREFLOATS,
+            cover=True)
+
+    def build(self, fieldfile):
+        BuildEbook.build(self, fieldfile)
+        clear_cache(fieldfile.instance.slug)
+
+
+@BuildEbook.register('epub')
+@task(ignore_result=True)
+class BuildEpub(BuildEbook):
+    @staticmethod
+    def transform(wldoc, fieldfile):
+        return wldoc.as_epub(cover=True)
+
+
+@BuildEbook.register('html')
+@task(ignore_result=True)
+class BuildHtml(BuildEbook):
+    def build(self, fieldfile):
+        from django.core.files.base import ContentFile
+        from fnpdjango.utils.text.slughifi import slughifi
+        from sortify import sortify
+        from librarian import html
+        from catalogue.models import Fragment, Tag
+
+        book = fieldfile.instance
+
+        meta_tags = list(book.tags.filter(
+            category__in=('author', 'epoch', 'genre', 'kind')))
+        book_tag = book.book_tag()
+
+        html_output = self.transform(
+                        book.wldocument(parse_dublincore=False),
+                        fieldfile)
+        if html_output:
+            fieldfile.save(None, ContentFile(html_output.get_string()),
+                    save=False)
+            type(book).objects.filter(pk=book.pk).update(**{
+                fieldfile.field.attname: fieldfile
+            })
+
+            # get ancestor l-tags for adding to new fragments
+            ancestor_tags = []
+            p = book.parent
+            while p:
+                ancestor_tags.append(p.book_tag())
+                p = p.parent
+
+            # Delete old fragments and create them from scratch
+            book.fragments.all().delete()
+            # Extract fragments
+            closed_fragments, open_fragments = html.extract_fragments(fieldfile.path)
+            for fragment in closed_fragments.values():
+                try:
+                    theme_names = [s.strip() for s in fragment.themes.split(',')]
+                except AttributeError:
+                    continue
+                themes = []
+                for theme_name in theme_names:
+                    if not theme_name:
+                        continue
+                    tag, created = Tag.objects.get_or_create(
+                                        slug=slughifi(theme_name),
+                                        category='theme')
+                    if created:
+                        tag.name = theme_name
+                        tag.sort_key = sortify(theme_name.lower())
+                        tag.save()
+                    themes.append(tag)
+                if not themes:
+                    continue
+
+                text = fragment.to_string()
+                short_text = truncate_html_words(text, 15)
+                if text == short_text:
+                    short_text = ''
+                new_fragment = Fragment.objects.create(anchor=fragment.id, 
+                        book=book, text=text, short_text=short_text)
+
+                new_fragment.save()
+                new_fragment.tags = set(meta_tags + themes + [book_tag] + ancestor_tags)
+            book.html_built.send(sender=book)
+            return True
+        return False
+
+
+class OverwritingFieldFile(FieldFile):
+    """
+        Deletes the old file before saving the new one.
+    """
+
+    def save(self, name, content, *args, **kwargs):
+        leave = kwargs.pop('leave', None)
+        # delete if there's a file already and there's a new one coming
+        if not leave and self and (not hasattr(content, 'path') or
+                                   content.path != self.path):
+            self.delete(save=False)
+        return super(OverwritingFieldFile, self).save(
+                name, content, *args, **kwargs)
+
+
+class OverwritingFileField(models.FileField):
+    attr_class = OverwritingFieldFile
+
+
+try:
+    # check for south
+    from south.modelsinspector import add_introspection_rules
+except ImportError:
+    pass
+else:
+    add_introspection_rules([
+        (
+            [EbookField],
+            [],
+            {'format_name': ('format_name', {})}
+        )
+    ], ["^catalogue\.fields\.EbookField"])
+    add_introspection_rules([], ["^catalogue\.fields\.OverwritingFileField"])