2018-09-07 00:09:51 +02:00
|
|
|
import re
|
|
|
|
|
|
|
|
from markdown.extensions import Extension
|
|
|
|
from markdown.preprocessors import Preprocessor
|
|
|
|
from typing import Any, Dict, Optional, List, Tuple
|
|
|
|
import markdown
|
|
|
|
|
|
|
|
START_TABBED_SECTION_REGEX = re.compile(r'^\{start_tabs\}$')
|
|
|
|
END_TABBED_SECTION_REGEX = re.compile(r'^\{end_tabs\}$')
|
|
|
|
TAB_CONTENT_REGEX = re.compile(r'^\{tab\|\s*(.+?)\s*\}$')
|
|
|
|
|
|
|
|
CODE_SECTION_TEMPLATE = """
|
|
|
|
<div class="code-section" markdown="1">
|
|
|
|
{nav_bar}
|
|
|
|
<div class="blocks">
|
|
|
|
{blocks}
|
|
|
|
</div>
|
|
|
|
</div>
|
|
|
|
""".strip()
|
|
|
|
|
|
|
|
NAV_BAR_TEMPLATE = """
|
|
|
|
<ul class="nav">
|
|
|
|
{tabs}
|
|
|
|
</ul>
|
|
|
|
""".strip()
|
|
|
|
|
|
|
|
NAV_LIST_ITEM_TEMPLATE = """
|
|
|
|
<li data-language="{data_language}">{name}</li>
|
|
|
|
""".strip()
|
|
|
|
|
|
|
|
DIV_TAB_CONTENT_TEMPLATE = """
|
|
|
|
<div data-language="{data_language}" markdown="1">
|
|
|
|
{content}
|
|
|
|
</div>
|
|
|
|
""".strip()
|
|
|
|
|
2018-09-15 21:09:35 +02:00
|
|
|
# If adding new entries here, also check if you need to update
|
|
|
|
# tabbed-instructions.js
|
2018-09-07 00:09:51 +02:00
|
|
|
TAB_DISPLAY_NAMES = {
|
|
|
|
'desktop-web': 'Desktop/Web',
|
|
|
|
'ios': 'iOS',
|
|
|
|
'android': 'Android',
|
2018-09-15 21:35:25 +02:00
|
|
|
'mac': 'macOS',
|
|
|
|
'windows': 'Windows',
|
|
|
|
'linux': 'Linux',
|
2018-09-17 16:27:32 +02:00
|
|
|
'python': 'Python',
|
|
|
|
'js': 'JavaScript',
|
|
|
|
'curl': 'curl',
|
|
|
|
'zulip-send': 'zulip-send',
|
2018-10-07 01:19:00 +02:00
|
|
|
|
|
|
|
'cloud': 'HipChat Cloud',
|
|
|
|
'server': 'HipChat Server or Data Center',
|
2018-09-07 00:09:51 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
class TabbedSectionsGenerator(Extension):
|
|
|
|
def extendMarkdown(self, md: markdown.Markdown, md_globals: Dict[str, Any]) -> None:
|
|
|
|
md.preprocessors.add(
|
|
|
|
'tabbed_sections', TabbedSectionsPreprocessor(md, self.getConfigs()), '_end')
|
|
|
|
|
|
|
|
class TabbedSectionsPreprocessor(Preprocessor):
|
|
|
|
def __init__(self, md: markdown.Markdown, config: Dict[str, Any]) -> None:
|
|
|
|
super(TabbedSectionsPreprocessor, self).__init__(md)
|
|
|
|
|
|
|
|
def run(self, lines: List[str]) -> List[str]:
|
|
|
|
tab_section = self.parse_tabs(lines)
|
|
|
|
while tab_section:
|
|
|
|
nav_bar = self.generate_nav_bar(tab_section)
|
|
|
|
content_blocks = self.generate_content_blocks(tab_section, lines)
|
|
|
|
rendered_tabs = CODE_SECTION_TEMPLATE.format(
|
|
|
|
nav_bar=nav_bar, blocks=content_blocks)
|
|
|
|
|
|
|
|
start = tab_section['start_tabs_index']
|
|
|
|
end = tab_section['end_tabs_index'] + 1
|
|
|
|
lines = lines[:start] + [rendered_tabs] + lines[end:]
|
|
|
|
tab_section = self.parse_tabs(lines)
|
|
|
|
return lines
|
|
|
|
|
|
|
|
def generate_content_blocks(self, tab_section: Dict[str, Any], lines: List[str]) -> str:
|
|
|
|
tab_content_blocks = []
|
|
|
|
for index, tab in enumerate(tab_section['tabs']):
|
|
|
|
start_index = tab['start'] + 1
|
|
|
|
try:
|
|
|
|
# If there are more tabs, we can use the starting index
|
|
|
|
# of the next tab as the ending index of the previous one
|
|
|
|
end_index = tab_section['tabs'][index + 1]['start']
|
|
|
|
except IndexError:
|
|
|
|
# Otherwise, just use the end of the entire section
|
|
|
|
end_index = tab_section['end_tabs_index']
|
|
|
|
|
|
|
|
content = '\n'.join(lines[start_index:end_index]).strip()
|
|
|
|
tab_content_block = DIV_TAB_CONTENT_TEMPLATE.format(
|
|
|
|
data_language=tab['tab_name'],
|
|
|
|
# Wrapping the content in two newlines is necessary here.
|
|
|
|
# If we don't do this, the inner Markdown does not get
|
|
|
|
# rendered properly.
|
|
|
|
content='\n{}\n'.format(content))
|
|
|
|
tab_content_blocks.append(tab_content_block)
|
|
|
|
return '\n'.join(tab_content_blocks)
|
|
|
|
|
|
|
|
def generate_nav_bar(self, tab_section: Dict[str, Any]) -> str:
|
|
|
|
li_elements = []
|
|
|
|
for tab in tab_section['tabs']:
|
|
|
|
li = NAV_LIST_ITEM_TEMPLATE.format(
|
|
|
|
data_language=tab.get('tab_name'),
|
|
|
|
name=TAB_DISPLAY_NAMES.get(tab.get('tab_name')))
|
|
|
|
li_elements.append(li)
|
|
|
|
return NAV_BAR_TEMPLATE.format(tabs='\n'.join(li_elements))
|
|
|
|
|
|
|
|
def parse_tabs(self, lines: List[str]) -> Optional[Dict[str, Any]]:
|
|
|
|
block = {} # type: Dict[str, Any]
|
|
|
|
for index, line in enumerate(lines):
|
|
|
|
start_match = START_TABBED_SECTION_REGEX.search(line)
|
|
|
|
if start_match:
|
|
|
|
block['start_tabs_index'] = index
|
|
|
|
|
|
|
|
tab_content_match = TAB_CONTENT_REGEX.search(line)
|
|
|
|
if tab_content_match:
|
|
|
|
block.setdefault('tabs', [])
|
|
|
|
tab = {'start': index,
|
|
|
|
'tab_name': tab_content_match.group(1)}
|
|
|
|
block['tabs'].append(tab)
|
|
|
|
|
|
|
|
end_match = END_TABBED_SECTION_REGEX.search(line)
|
|
|
|
if end_match:
|
|
|
|
block['end_tabs_index'] = index
|
|
|
|
break
|
|
|
|
return block
|
|
|
|
|
|
|
|
def makeExtension(*args: Any, **kwargs: str) -> TabbedSectionsGenerator:
|
2018-12-20 08:28:40 +01:00
|
|
|
return TabbedSectionsGenerator(**kwargs)
|