All patches and comments are welcome. Please squash your changes to logical
commits before using git-format-patch and git-send-email to
patches@git.madduck.net.
If you'd read over the Git project's submission guidelines and adhered to them,
I'd be especially grateful.
3 # markdown2html.py — simple Markdown-to-HTML converter for use with Mutt
5 # Mutt recently learnt [how to compose `multipart/alternative`
6 # emails][1]. This script assumes a message has been composed using Markdown
7 # (with a lot of pandoc extensions enabled), and translates it to `text/html`
8 # for Mutt to tie into such a `multipart/alternative` message.
10 # [1]: https://gitlab.com/muttmua/mutt/commit/0e566a03725b4ad789aa6ac1d17cdf7bf4e7e354)
14 # set send_multipart_alternative=yes
15 # set send_multipart_alternative_filter=/path/to/markdown2html.py
17 # Optionally, Custom CSS styles will be read from `~/.config/mutt/markdown2html.css`,
22 # - PyPandoc (and pandoc installed, or downloaded)
26 # - Pygments, if installed, then syntax highlighting is enabled
29 # https://git.madduck.net/etc/mutt.git/blob_plain/HEAD:/.mutt/markdown2html
31 # Copyright © 2019 martin f. krafft <madduck@madduck.net>
32 # Released under the GPL-2+ licence, just like Mutt itself.
42 from pygments.formatters import get_formatter_by_name
43 formatter = get_formatter_by_name('html', style='default')
44 DEFAULT_CSS = formatter.get_style_defs('.sourceCode')
55 border-left: 2px solid #eee;
61 border-left: 2px solid #666;
69 .quotechar { display: none; }
70 .footnote-ref, .footnote-back { text-decoration: none;}
73 font-family: monospace;
79 border-collapse: collapse;
80 border: 1px solid #999;
82 th, td { padding: 0.5em; }
86 .even { background: #eee; }
87 h1, h2, h3, h4, h5, h6 {
89 background-color: #eee;
92 h1 { font-size: 130%; }
93 h2 { font-size: 120%; }
94 h3 { font-size: 110%; }
95 h4 { font-size: 107%; }
96 h5 { font-size: 103%; }
97 h6 { font-size: 100%; }
98 p { padding: 0 0.5em; }
99 pre { padding: 0 1em; }
102 STYLESHEET = os.path.join(os.path.expanduser('~/.config/mutt'),
104 if os.path.exists(STYLESHEET):
105 DEFAULT_CSS += open(STYLESHEET).read()
109 <div class="signature"><span class="leader">-- </span><br/>
114 def _preprocess_signature(sig):
116 Preprocess the signature before markdown processing.
118 return f"```\n{sig}\n```"
120 def _preprocess_markdown(mdwn):
122 Preprocess Markdown for handling by the converter.
124 # convert hard line breaks within paragraphs to 2 trailing spaces, which
125 # is the markdown way of representing hard line breaks. Note how the
126 # regexp will not match between paragraphs.
127 ret = re.sub(r'(\S)\n(\s*\S)', r'\g<1> \n\g<2>', mdwn, flags=re.MULTILINE)
129 # Clients like Thunderbird need the leading '>' to be able to properly
130 # create nested quotes, so we duplicate the symbol, the first instance
131 # will tell pandoc to create a blockquote, while the second instance will
132 # be a <span> containing the character, along with a class that causes CSS
133 # to actually hide it from display. However, this does not work with the
134 # text-mode HTML2text converters, and so it's left commented for now.
135 #ret = re.sub(r'\n>', r' \n>[>]{.quotechar}', ret, flags=re.MULTILINE)
137 # With the autolink_bare_uris extension, we do not need to put links into
138 # angle brackets to have them converted, so let's conserve the brackets
139 # when used around email addresses. Note that this needs a postprocessing
140 # hack because the pandoc autolink converted includes the ambersand
141 # (https://github.com/jgm/pandoc/issues/7398).
142 ret = re.sub(r'<([^@]+@\S+)>', r'<\g<1> PANDOC_BUG_7398 >', ret)
146 def _identify_quotes_for_later(mdwn):
148 Email quoting such as:
151 On 1970-01-01, you said:
152 > The Flat Earth Society has members all around the globe.
155 isn't really properly handled by Markdown, so let's do our best to
156 identify the individual elements, and mark them, using a syntax similar to
157 what pandoc uses already in some cases. As pandoc won't actually use these
158 data (yet?), we call `self._reformat_quotes` later to use these markers
159 to slap the appropriate classes on the HTML tags.
162 def generate_lines_with_context(mdwn):
164 Iterates the input string line-wise, returning a triplet of
165 previous, current, and next line, the first and last of which
166 will be None on the first and last line of the input data
169 prev = cur = nxt = None
170 lines = iter(mdwn.splitlines())
176 yield prev, cur, None
179 for prev, cur, nxt in generate_lines_with_context(mdwn):
181 # The lead-in to a quote is a single line immediately preceding the
182 # quote, and ending with ':'. Note that there could be multiple of
184 if re.match(r'^[^>]+.*:\s*$', cur) and nxt.startswith('>'):
185 ret.append(f'{{.quotelead}}{cur.strip()}')
186 # pandoc needs an empty line before the blockquote, so
187 # we enter one for the purpose of HTML rendition:
191 # The first blockquote after such a lead-in gets marked as the
193 elif prev and re.match(r'^[^>]+.*:\s*$', prev) and cur.startswith('>'):
194 ret.append(re.sub(r'^(\s*>\s*)+(.+)',
195 r'\g<1>{.quoteinitial}\g<2>',
196 cur, flags=re.MULTILINE))
198 # All other occurrences of blockquotes get the "subsequent" marker:
199 elif cur.startswith('>') and prev is not None and not prev.startswith('>'):
200 ret.append(re.sub(r'^((?:\s*>\s*)+)(.+)',
201 r'\g<1>{.quotesubsequent}\g<2>',
202 cur, flags=re.MULTILINE))
204 else: # pass through everything else.
207 return '\n'.join(ret)
210 def _reformat_quotes(html):
212 Earlier in the pipeline, we marked email quoting, using markers, which we
213 now need to turn into HTML classes, so that we can use CSS to style them.
215 ret = html.replace('<p>{.quotelead}', '<p class="quotelead">')
216 ret = re.sub(r'<blockquote>\n((?:<blockquote>\n)*)<p>(?:\{\.quote(\w+)\})',
217 r'<blockquote class="quote \g<2>">\n\g<1><p>', ret, flags=re.MULTILINE)
222 def _convert_with_pandoc(mdwn, inputfmt='markdown', outputfmt='html5',
223 ext_enabled=None, ext_disabled=None,
224 standalone=True, selfcontained=True, title='Untitled'):
226 Invoke pandoc to do the actual conversion of Markdown to HTML5.
229 ext_enabled = [ 'backtick_code_blocks',
240 'all_symbols_escapable',
241 'intraword_underscores',
250 'tex_math_double_backslash',
254 ext_disabled = [ 'tex_math_single_backslash',
258 'yaml_metadata_block'
261 enabled = '+'.join(ext_enabled)
262 disabled = '-'.join(ext_disabled)
263 inputfmt = f'{inputfmt}+{enabled}-{disabled}'
265 args = ['--metadata=document-css:false']
267 args.append('--standalone')
269 args.append('--self-contained')
271 args.append(f'--metadata=title:{title}')
273 return pypandoc.convert_text(mdwn, format=inputfmt, to=outputfmt,
277 def _apply_styling(html):
279 Inline all styles defined and used into the individual HTML tags.
281 return pynliner.Pynliner().from_string(html).with_cssString(DEFAULT_CSS).run()
284 def _postprocess_html(html):
286 Postprocess the generated and styled HTML.
289 # Preprocessing leaves a sentinel to work around
290 # https://github.com/jgm/pandoc/issues/7398, and so we need to remove it:
291 html = html.replace(' PANDOC_BUG_7398 ', '')
295 def convert_markdown_to_html(mdwn):
297 Converts the input Markdown to HTML, handling separately the body, as well
298 as an optional signature.
300 parts = re.split(r'^-- \n', mdwn, 1, flags=re.MULTILINE)
309 body = _preprocess_markdown(body)
310 body = _identify_quotes_for_later(body)
311 html = _convert_with_pandoc(body, standalone=True, selfcontained=True,
313 html = _reformat_quotes(html)
316 sig = _preprocess_signature(sig)
317 sig = _preprocess_markdown(sig)
318 sig = _convert_with_pandoc(sig, standalone=False, selfcontained=False,
320 sig = SIGNATURE_HTML.format(sig=sig)
321 eob = html.find('</body>')
322 html = f'{html[:eob]}{sig}\n{html[eob:]}'
324 html = _apply_styling(html)
325 html = _postprocess_html(html)
332 Convert text on stdin to HTML, and print it to stdout, like mutt would
335 with open(sys.argv[1], 'r') if len(sys.argv) > 1 else sys.stdin as f:
336 html = convert_markdown_to_html(f.read())
338 # mutt expects the content type in the first line, so:
339 print(f'text/html\n\n{html}')
342 if __name__ == '__main__':