def soup2markup(self, soup: Any) -> List[Any]:
# Ensure a string is provided, in case the soup finds none
# This could occur if eg. an image is removed or not shown
markup = [''] # type: List[Union[str, Tuple[Optional[str], Any]]]
if soup is None: # This is not iterable, so return promptly
return markup
unrendered_tags = { # In pairs of 'tag_name': 'text'
# TODO: Some of these could be implemented
'br': '', # No indicator of absence
'hr': 'RULER',
'img': 'IMAGE',
}
unrendered_div_classes = { # In pairs of 'div_class': 'text'
# TODO: Support embedded content & twitter preview?
'message_embed': 'EMBEDDED CONTENT',
'inline-preview-twitter': 'TWITTER PREVIEW',
'message_inline_ref': '', # Duplicate of other content
'message_inline_image': '', # Duplicate of other content
}
unrendered_template = '[{} NOT RENDERED]'
for element in soup:
if isinstance(element, NavigableString):
# NORMAL STRINGS
if hasattr(self, 'bq_len') and element == '\n' and \
self.bq_len > 0:
self.bq_len -= 1
continue
markup.append(element)
elif (element.name == 'div' and element.attrs and
any(cls in element.attrs.get('class', [])
for cls in unrendered_div_classes)):
# UNRENDERED DIV CLASSES
matching_class = (set(unrendered_div_classes) &
set(element.attrs.get('class')))
text = unrendered_div_classes[matching_class.pop()]
if text:
markup.append(unrendered_template.format(text))
elif (element.name == 'img' and
element.attrs.get('class', []) == ['emoji']):
# CUSTOM EMOJIS AND ZULIP_EXTRA_EMOJI
emoji_name = element.attrs.get('title', [])
markup.append(('msg_emoji', ":"+emoji_name+":"))
elif element.name in unrendered_tags:
# UNRENDERED SIMPLE TAGS
text = unrendered_tags[element.name]
if text:
markup.append(unrendered_template.format(text))
elif element.name in ('p', 'ul', 'del'):
# PARAGRAPH, LISTS, STRIKE-THROUGH
markup.extend(self.soup2markup(element))
elif (element.name == 'span' and element.attrs and
'emoji' in element.attrs.get('class', [])):
# EMOJI
markup.append(('msg_emoji', element.text))
elif (element.name == 'span' and element.attrs and
('katex-display' in element.attrs.get('class', []) or
'katex' in element.attrs.get('class', []))):
# MATH TEXT
markup.append(element.text)
elif element.name == 'span' and element.attrs and\
('user-mention' in element.attrs.get('class', []) or
'user-group-mention' in element.attrs.get('class', [])):
# USER MENTIONS & USER-GROUP MENTIONS
markup.append(('msg_mention', element.text))
elif element.name == 'a':
# LINKS
link = element.attrs['href']
text = element.img['src'] if element.img else element.text
parsed_link = urlparse(link)
if not parsed_link.scheme: # => relative link
# Prepend org url to convert it to an absolute link
link = urljoin(self.model.server_url, link)
if link == text:
# If the link and text are same
# usually the case when user just pastes
# a link then just display the link
markup.append(('msg_link', text))
else:
markup.append(
('msg_link', '[' + text + ']' + '(' + link + ')'))
elif element.name == 'blockquote':
# BLOCKQUOTE TEXT
markup.append((
'msg_quote', self.soup2markup(element)
))
elif element.name == 'code':
# CODE (INLINE?)
markup.append((
'msg_code', element.text
))
elif element.name == 'div' and element.attrs and\
'codehilite' in element.attrs.get('class', []):
# CODE (BLOCK?)
markup.append((
'msg_code', element.text
))
elif element.name in ('strong', 'em'):
# BOLD & ITALIC
markup.append(('msg_bold', element.text))
elif element.name == 'li':
# LISTS
# TODO: Support nested lists
markup.append(' * ')
markup.extend(self.soup2markup(element))
elif element.name == 'table':
markup.extend(render_table(element))
else:
markup.extend(self.soup2markup(element))
return markup
def soup2markup(self, soup: Any, **state: Any) -> List[Any]:
# Ensure a string is provided, in case the soup finds none
# This could occur if eg. an image is removed or not shown
markup = [''] # type: List[Union[str, Tuple[Optional[str], Any]]]
if soup is None: # This is not iterable, so return promptly
return markup
unrendered_tags = { # In pairs of 'tag_name': 'text'
# TODO: Some of these could be implemented
'br': '', # No indicator of absence
'hr': 'RULER',
'img': 'IMAGE',
}
unrendered_div_classes = { # In pairs of 'div_class': 'text'
# TODO: Support embedded content & twitter preview?
'message_embed': 'EMBEDDED CONTENT',
'inline-preview-twitter': 'TWITTER PREVIEW',
'message_inline_ref': '', # Duplicate of other content
'message_inline_image': '', # Duplicate of other content
}
unrendered_template = '[{} NOT RENDERED]'
for element in soup:
if isinstance(element, NavigableString):
# NORMAL STRINGS
if (hasattr(self, 'bq_len') and element == '\n'
and self.bq_len > 0):
self.bq_len -= 1
continue
markup.append(element)
elif (element.name == 'div' and element.attrs
and any(cls in element.attrs.get('class', [])
for cls in unrendered_div_classes)):
# UNRENDERED DIV CLASSES
matching_class = (set(unrendered_div_classes)
& set(element.attrs.get('class')))
text = unrendered_div_classes[matching_class.pop()]
if text:
markup.append(unrendered_template.format(text))
elif (element.name == 'img'
and element.attrs.get('class', []) == ['emoji']):
# CUSTOM EMOJIS AND ZULIP_EXTRA_EMOJI
emoji_name = element.attrs.get('title', [])
markup.append(('msg_emoji', ":" + emoji_name + ":"))
elif element.name in unrendered_tags:
# UNRENDERED SIMPLE TAGS
text = unrendered_tags[element.name]
if text:
markup.append(unrendered_template.format(text))
elif element.name in ('p', 'del'):
# PARAGRAPH, STRIKE-THROUGH
markup.extend(self.soup2markup(element))
elif (element.name == 'span' and element.attrs
and 'emoji' in element.attrs.get('class', [])):
# EMOJI
markup.append(('msg_emoji', element.text))
elif (element.name == 'span' and element.attrs
and ('katex-display' in element.attrs.get('class', [])
or 'katex' in element.attrs.get('class', []))):
# MATH TEXT
markup.append(element.text)
elif (element.name == 'span' and element.attrs
and ('user-group-mention' in element.attrs.get('class', [])
or 'user-mention' in element.attrs.get('class', []))):
# USER MENTIONS & USER-GROUP MENTIONS
markup.append(('msg_mention', element.text))
elif element.name == 'a':
# LINKS
# Use rstrip to avoid anomalies and edge cases like
# https://google.com vs https://google.com/.
link = element.attrs['href'].rstrip('/')
text = element.img['src'] if element.img else element.text
text = text.rstrip('/')
parsed_link = urlparse(link)
if not parsed_link.scheme: # => relative link
# Prepend org url to convert it to an absolute link
link = urljoin(self.model.server_url, link)
text = text if text else link
show_footlink = True
# Only use the last segment if the text is redundant.
# NOTE: The 'without scheme' excerpt is to deal with the case
# where a user puts a link without any scheme and the server
# uses http as the default scheme but keeps the text as-is.
# For instance, see how example.com/some/path becomes
# <a href="http://example.com">example.com/some/path</a>.
link_without_scheme, text_without_scheme = [
data.split('://')[1] if '://' in data else data
for data in [link, text]
] # Split on '://' is for cases where text == link.
if link_without_scheme == text_without_scheme:
segment = text.split('/')[-1]
# Replace text with its last segment if the segment has
# something significant than simply the 'domain name'.
if segment != text_without_scheme:
text = segment
else:
# Do not show as a footlink as the text is sufficient
# to represent the link.
show_footlink = False
# Detect duplicate links to save screen real estate.
if link not in self.message_links:
self.message_links[link] = (text,
len(self.message_links) + 1,
show_footlink)
else:
# Append the text if its link already exist with a
# different text.
saved_text, saved_link_index, saved_footlink_status = (
self.message_links[link])
if saved_text != text:
self.message_links[link] = (
'{}, {}'.format(saved_text, text),
saved_link_index,
show_footlink or saved_footlink_status,
)
markup.extend([
('msg_link', text),
' ',
('msg_link_index',
'[{}]'.format(self.message_links[link][1])),
])
elif element.name == 'blockquote':
# BLOCKQUOTE TEXT
markup.append(('msg_quote', self.soup2markup(element)))
elif element.name == 'code':
# CODE (INLINE?)
markup.append(('msg_code', element.text))
elif (element.name == 'div' and element.attrs
and 'codehilite' in element.attrs.get('class', [])):
# CODE (BLOCK?)
markup.append(('msg_code', element.text))
elif element.name in ('strong', 'em'):
# BOLD & ITALIC
markup.append(('msg_bold', element.text))
elif element.name in ('ul', 'ol'):
# LISTS (UL & OL)
for part in element.contents:
if part == '\n':
part.replace_with('')
if 'indent_level' not in state:
state['indent_level'] = 1
state['list_start'] = True
else:
state['indent_level'] += 1
state['list_start'] = False
if element.name == 'ol':
start_number = int(element.attrs.get('start', 1))
state['list_index'] = start_number
markup.extend(self.soup2markup(element, **state))
del state['list_index'] # reset at end of this list
else:
if 'list_index' in state:
del state['list_index'] # this is unordered
markup.extend(self.soup2markup(element, **state))
del state['indent_level'] # reset indents after any list
elif element.name == 'li':
# LIST ITEMS (LI)
for part in element.contents:
if part == '\n':
part.replace_with('')
if not state.get('list_start', False):
markup.append('\n')
indent = state.get('indent_level', 1)
if 'list_index' in state:
markup.append('{}{}. '.format(' ' * indent,
state['list_index']))
state['list_index'] += 1
else:
chars = [
'\N{BULLET}',
'\N{RING OPERATOR}', # small hollow
'\N{HYPHEN}',
]
markup.append('{}{} '.format(' ' * indent,
chars[(indent - 1) % 3]))
state['list_start'] = False
markup.extend(self.soup2markup(element, **state))
elif element.name == 'table':
markup.extend(render_table(element))
else:
markup.extend(self.soup2markup(element))
return markup