def render_content(obj,
include_location=True,
synthesize_content=True,
render_attachments=False):
"""Renders the content of an ActivityStreams object as HTML.
Includes tags, mentions, and non-note/article attachments. (Note/article
attachments are converted to mf2 children in object_to_json and then rendered
in json_to_html.)
Note that the returned HTML is included in Atom as well as HTML documents,
so it *must* be HTML4 / XHTML, not HTML5! All tags must be closed, etc.
Args:
obj: decoded JSON ActivityStreams object
include_location: whether to render location, if provided
synthesize_content: whether to generate synthetic content if the object
doesn't have its own, e.g. 'likes this.' or 'shared this.'
Returns:
string, rendered HTML
"""
content = obj.get('content', '')
# extract tags. preserve order but de-dupe, ie don't include a tag more than
# once.
seen_ids = set()
mentions = []
tags = {} # maps string objectType to list of tag objects
for t in obj.get('tags', []):
id = t.get('id')
if id and id in seen_ids:
continue
seen_ids.add(id)
if 'startIndex' in t and 'length' in t:
mentions.append(t)
else:
tags.setdefault(source.object_type(t), []).append(t)
# linkify embedded mention tags inside content.
if mentions:
mentions.sort(key=lambda t: t['startIndex'])
last_end = 0
orig = util.WideUnicode(content)
content = util.WideUnicode('')
for tag in mentions:
start = tag['startIndex']
end = start + tag['length']
content = util.WideUnicode(
'%s%s<a href="%s">%s</a>' %
(content, orig[last_end:start], tag['url'], orig[start:end]))
last_end = end
content += orig[last_end:]
if not obj.get('content_is_html'):
# convert newlines to <br>s
# do this *after* linkifying tags so we don't have to shuffle indices over
content = content.replace('\n', '<br />\n')
# linkify embedded links. ignore the "mention" tags that we added ourselves.
# TODO: fix the bug in test_linkify_broken() in webutil/util_test.py, then
# uncomment this.
# if content:
# content = util.linkify(content)
# attachments, e.g. links (aka articles)
# TODO: use oEmbed? http://oembed.com/ , http://code.google.com/p/python-oembed/
if render_attachments:
atts = [
a for a in obj.get('attachments', [])
if a.get('objectType') not in ('note', 'article')
]
content += _render_attachments(atts + tags.pop('article', []), obj)
# generate share/like contexts if the activity does not have content
# of its own
obj_type = source.object_type(obj)
for as_type, verb in (('favorite', 'Favorites'), ('like', 'Likes'),
('share', 'Shared')):
if (not synthesize_content or obj_type != as_type
or 'object' not in obj or 'content' in obj):
continue
targets = get_list(obj, 'object')
if not targets:
continue
for target in targets:
# sometimes likes don't have enough content to render anything
# interesting
if 'url' in target and set(target) <= set(['url', 'objectType']):
content += '<a href="%s">%s this.</a>' % (target.get('url'),
verb.lower())
else:
author = target.get('author', target.get('actor', {}))
# special case for twitter RT's
if obj_type == 'share' and 'url' in obj and re.search(
'^https?://(?:www\.|mobile\.)?twitter\.com/',
obj.get('url')):
content += 'RT <a href="%s">@%s</a> ' % (target.get(
'url', '#'), author.get('username'))
else:
# image looks bad in the simplified rendering
author = {k: v for k, v in author.items() if k != 'image'}
content += '%s <a href="%s">%s</a> by %s' % (
verb,
target.get('url', '#'),
target.get('displayName', target.get(
'title', 'a post')),
hcard_to_html(
object_to_json(author,
default_object_type='person')),
)
content += render_content(
target,
include_location=include_location,
synthesize_content=synthesize_content)
# only include the first context in the content (if there are
# others, they'll be included as separate properties)
break
break
if render_attachments and obj.get('verb') == 'share':
atts = [
a for a in obj.get('object', {}).get('attachments', [])
if a.get('objectType') not in ('note', 'article')
]
content += _render_attachments(atts, obj)
# location
loc = obj.get('location')
if include_location and loc:
content += '\n<p>%s</p>' % hcard_to_html(object_to_json(
loc, default_object_type='place'),
parent_props=['p-location'])
# these are rendered manually in json_to_html()
for type in 'like', 'share', 'react', 'person':
tags.pop(type, None)
# render the rest
content += tags_to_html(tags.pop('hashtag', []), 'p-category')
content += tags_to_html(tags.pop('mention', []), 'u-mention')
content += tags_to_html(sum(tags.values(), []), 'tag')
return content
def status_to_object(self, status):
"""Converts a status to an object.
Args:
status: dict, a decoded JSON status
Returns:
an ActivityStreams object dict, ready to be JSON-encoded
"""
id = status.get('id')
if not id:
return {}
obj = {
'objectType': 'note',
'id': self.tag_uri(id),
'url': status.get('url'),
'published': status.get('created_at'),
'author': self.user_to_actor(status.get('account') or {}),
'attachments': [],
}
reblog = status.get('reblog')
base_status = reblog if reblog else status
# media! into attachments.
for media in status.get('media_attachments', []):
type = media.get('type')
att = {
'id': self.tag_uri(media.get('id')),
'objectType': MEDIA_TYPES.get(type),
'displayName': media.get('description'),
}
url = media.get('url')
if type == 'image':
att['image'] = {'url': url}
elif type in ('gifv', 'video'):
att.update({
'stream': {
'url': url
},
'image': {
'url': media.get('preview_url')
},
})
obj['attachments'].append(att)
if obj['attachments']:
first = obj['attachments'][0]
if first['objectType'] == 'video':
obj['stream'] = first.get('stream')
else:
obj['image'] = first.get('image')
# tags
obj['tags'] = [{
'objectType': 'person',
'id': self.tag_uri(t.get('id')),
'url': t.get('url'),
'displayName': t.get('username'),
} for t in status.get('mentions', [])] + [{
'objectType': 'hashtag',
'url': t.get('url'),
'displayName': t.get('name'),
} for t in status.get('tags', [])]
card = status.get('card')
if card:
obj['tags'].append({
'objectType': 'article',
'url': card.get('url'),
'displayName': card.get('title'),
'content': card.get('description'),
'image': {
'url': card.get('image')
},
})
# content: insert images for custom emoji
# https://docs.joinmastodon.org/api/entities/#emoji
content = base_status.get('content') or ''
for emoji in base_status.get('emojis', []):
shortcode = emoji.get('shortcode')
url = emoji.get('url')
if shortcode and url:
content = re.sub(
r'(^|[^\w]):%s:([^\w]|$)' % shortcode,
r'\1<img alt="%s" src="%s" style="height: 1em">\2' %
(shortcode, url), content)
# content: add 'Boosted @username:'******'content'):
reblog_account = reblog.get('account')
content = 'Boosted <a href="%s">@%s</a>: ' % (
(reblog_account.get('url'),
reblog_account.get('username'))) + content
obj['content'] = util.WideUnicode(content)
# inReplyTo
reply_to_id = status.get('in_reply_to_id')
if reply_to_id:
obj['inReplyTo'] = [{
'id':
self.tag_uri(reply_to_id),
# Mastodon's in_reply_to_id is str, Pixelfed's is int.
'url':
urllib.parse.urljoin(self.instance,
'/web/statuses/' + str(reply_to_id)),
}]
# to (ie visibility)
visibility = status.get('visibility')
if visibility:
obj['to'] = [{
'objectType': 'group',
'alias': '@' + visibility,
}]
return self.postprocess_object(obj)
def render_content(obj, include_location=True, synthesize_content=True,
render_attachments=False, render_image=False,
white_space_pre=True):
"""Renders the content of an ActivityStreams object as HTML.
Includes tags, mentions, and non-note/article attachments. (Note/article
attachments are converted to mf2 children in object_to_json and then rendered
in json_to_html.)
Note that the returned HTML is included in Atom as well as HTML documents,
so it *must* be HTML4 / XHTML, not HTML5! All tags must be closed, etc.
Args:
obj: decoded JSON ActivityStreams object
include_location: boolean, whether to render location, if provided
synthesize_content: boolean, whether to generate synthetic content if the
object doesn't have its own, e.g. 'likes this.' or 'shared this.'
render_attachments: boolean, whether to render attachments, eg links,
images, audio, and video
render_image: boolean, whether to render the object's image(s)
white_space_pre: boolean, whether to wrap in CSS white-space: pre. If False,
newlines will be converted to <br> tags instead. Background:
https://indiewebcamp.com/note#Indieweb_whitespace_thinking
Returns:
string, rendered HTML
"""
content = obj.get('content', '')
# extract tags. preserve order but de-dupe, ie don't include a tag more than
# once.
seen_ids = set()
mentions = []
tags = {} # maps string objectType to list of tag objects
for t in obj.get('tags', []):
id = t.get('id')
if id and id in seen_ids:
continue
seen_ids.add(id)
if 'startIndex' in t and 'length' in t and 'url' in t:
mentions.append(t)
else:
tags.setdefault(source.object_type(t), []).append(t)
# linkify embedded mention tags inside content.
if mentions:
mentions.sort(key=lambda t: t['startIndex'])
last_end = 0
orig = util.WideUnicode(content)
content = util.WideUnicode('')
for tag in mentions:
start = tag['startIndex']
end = start + tag['length']
content = util.WideUnicode('%s%s<a href="%s">%s</a>' % (
content, orig[last_end:start], tag['url'], orig[start:end]))
last_end = end
content += orig[last_end:]
# is whitespace in this content meaningful? standard heuristic: if there are
# no HTML tags in it, and it has a newline, then assume yes.
# https://indiewebcamp.com/note#Indieweb_whitespace_thinking
# https://github.com/snarfed/granary/issues/80
if content and not obj.get('content_is_html') and '\n' in content:
if white_space_pre:
content = '<div style="white-space: pre">%s</div>' % content
else:
content = content.replace('\n', '<br />\n')
# linkify embedded links. ignore the "mention" tags that we added ourselves.
# TODO: fix the bug in test_linkify_broken() in webutil/tests/test_util.py, then
# uncomment this.
# if content:
# content = util.linkify(content)
# the image field. may be multiply valued.
rendered_urls = set()
if render_image:
urls = get_urls(obj, 'image')
content += _render_attachments([{
'objectType': 'image',
'image': {'url': url},
} for url in urls], obj)
rendered_urls = set(urls)
# attachments, e.g. links (aka articles)
# TODO: use oEmbed? http://oembed.com/ , http://code.google.com/p/python-oembed/
if render_attachments:
atts = [a for a in obj.get('attachments', [])
if a.get('objectType') not in ('note', 'article')
and get_url(a, 'image') not in rendered_urls]
content += _render_attachments(atts + tags.pop('article', []), obj)
# generate share/like contexts if the activity does not have content
# of its own
obj_type = source.object_type(obj)
for as_type, verb in (
('favorite', 'Favorites'), ('like', 'Likes'), ('share', 'Shared')):
if (not synthesize_content or obj_type != as_type or 'object' not in obj or
'content' in obj):
continue
targets = get_list(obj, 'object')
if not targets:
continue
for target in targets:
# sometimes likes don't have enough content to render anything
# interesting
if 'url' in target and set(target) <= set(['url', 'objectType']):
content += '<a href="%s">%s this.</a>' % (
target.get('url'), verb.lower())
else:
author = target.get('author', target.get('actor', {}))
# special case for twitter RT's
if obj_type == 'share' and 'url' in obj and re.search(
'^https?://(?:www\.|mobile\.)?twitter\.com/', obj.get('url')):
content += 'RT <a href="%s">@%s</a> ' % (
target.get('url', '#'), author.get('username'))
else:
# image looks bad in the simplified rendering
author = {k: v for k, v in author.items() if k != 'image'}
content += '%s <a href="%s">%s</a> by %s' % (
verb, target.get('url', '#'),
target.get('displayName', target.get('title', 'a post')),
hcard_to_html(object_to_json(author, default_object_type='person')),
)
content += render_content(target, include_location=include_location,
synthesize_content=synthesize_content,
white_space_pre=white_space_pre)
# only include the first context in the content (if there are
# others, they'll be included as separate properties)
break
break
if render_attachments and obj.get('verb') == 'share':
atts = [att for att in itertools.chain.from_iterable(
o.get('attachments', []) for o in util.get_list(obj, 'object'))
if att.get('objectType') not in ('note', 'article')]
content += _render_attachments(atts, obj)
# location
loc = obj.get('location')
if include_location and loc:
content += '\n<p>%s</p>' % hcard_to_html(
object_to_json(loc, default_object_type='place'),
parent_props=['p-location'])
# these are rendered manually in json_to_html()
for type in 'like', 'share', 'react', 'person':
tags.pop(type, None)
# render the rest
content += tags_to_html(tags.pop('hashtag', []), 'p-category')
content += tags_to_html(tags.pop('mention', []), 'u-mention', visible=False)
content += tags_to_html(sum(tags.values(), []), 'tag')
return content
