def handle(self, *args, **options):
try:
file = args[0]
except IndexError:
raise CommandError('No file was specified')
try:
tree = ET.parse(file)
except IOError:
raise CommandError("%s could not be found" % file)
wp = 'http://wordpress.org/export/1.1/'
for item in tree.findall('channel/item'):
# Get the post's slug
slug = item.find('{%s}post_name' % (wp)).text
# Unpublished posts will not have anything in their name tag (which
# is used to hold the slug), so we have to create the slug from the
# title.
if slug is None:
slug = slugify(item.find('title').text)
print 'Importing post "%s"...' % slug
# If the post is already in the database, get it. Otherwise create
# it.
try:
post = Post.objects.get(slug=slug)
except:
post = Post()
post.title = item.find('title').text
post.slug = slug
post.body = item.find('{http://purl.org/rss/1.0/modules/content/}encoded').text
post.created = item.find('{%s}post_date' % (wp)).text
# If the post was published, set its status to public. Otherwise make
# it a draft.
if item.find('{%s}status' % (wp)).text == 'publish':
post.status = 2
else:
post.status = 1
# Unpublished posts will not have a timestamp associated
# with them. We'll set the creation date to now.
post.created = datetime.now()
# Set publish time to the creation time.
post.publish = post.created
# If the excerpt flag was set, do some auto excerpting magic.
if options['excerpt']:
# Partition the string at the Wordpress more quicktag.
partition = post.body.partition('<!--more-->')
# If the `more` tag was not found, Python will have returned
# a tuple with the full post body in the first item followed by
# two empty items. To make sure that the excerpt is only set if
# the post does actually contain a `more` quicktag, we'll check
# to see if the third tuple item is an empty string.
if partition[2]:
post.tease = partition[0]
# Post must be saved before we apply tags or comments.
post.save()
# Get all tags and categories. They look like this, respectively:
# <category domain="post_tag" nicename="a tag">a tag</category>
# <category domain="category" nicename="a category">a category</category>
descriptors = item.findall('category')
categories = []
for descriptor in descriptors:
if descriptor.attrib['domain'] == 'post_tag':
# Add the tag to the post
post.tags.add(descriptor.text)
if descriptor.attrib['domain'] == 'category':
category = descriptor.text
# If the category exists, add it to the model. Otherwise,
# create the category, then add it.
try:
cat = Category.objects.get(slug=slugify(category))
except:
cat = Category(title=category, slug=slugify(category))
cat.save()
post.categories.add(cat)
# Save the post again, this time with tags and categories.
post.save()
# Get and save the comments.
comments = item.findall('{%s}comment' % (wp))
for comment in comments:
# When I was importing my posts, I stumbled upon a comment that
# somehow had no author email associated with it. If that is
# the case, don't bother importing the comment.
email = comment.find('{%s}comment_author_email' % (wp)).text
if email is None:
continue
c = Comment()
c.user_name = comment.find('{%s}comment_author' % (wp)).text
c.user_email = comment.find('{%s}comment_author_email' % (wp)).text
c.comment = comment.find('{%s}comment_content' % (wp)).text
c.submit_date = comment.find('{%s}comment_date' % (wp)).text
c.content_type = ContentType.objects.get(app_label='blog', model='post')
c.object_pk = post.id
c.site_id = Site.objects.get_current().id
# Only attempt to assign a user URL to the new comment if the
# old comment has one.
user_url = comment.find('{%s}comment_author_url' % (wp)).text
if user_url:
c.user_url = user_url
c.save()
