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 `~/.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')
54 border-left: 2px solid #eee;
60 border-left: 2px solid #666;
68 .quotechar { display: none; }
69 .footnote-ref, .footnote-back { text-decoration: none;}
72 font-family: monospace;
78 border-collapse: collapse;
79 border: 1px solid #999;
81 th, td { padding: 0.5em; }
85 .even { background: #eee; }
86 h1, h2, h3, h4, h5, h6 {
88 background-color: #eee;
91 h1 { font-size: 130%; }
92 h2 { font-size: 120%; }
93 h3 { font-size: 110%; }
94 h4 { font-size: 107%; }
95 h5 { font-size: 103%; }
96 h6 { font-size: 100%; }
97 p { padding: 0 0.5em; }
100 STYLESHEET = os.path.join(os.path.expanduser('~/.mutt'),
102 if os.path.exists(STYLESHEET):
103 DEFAULT_CSS += open(STYLESHEET).read()
105 HTML_DOCUMENT = '''<!DOCTYPE html>
107 <meta http-equiv="Content-Type" content="text/html; charset=utf-8"/>
108 <meta charset="utf-8"/>
109 <meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes"/>
110 <title>HTML E-Mail</title>
111 </head><body class="email">
117 '<div class="signature"><span class="leader">-- </span>{sig}</div>'
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)
140 def _identify_quotes_for_later(mdwn):
142 Email quoting such as:
145 On 1970-01-01, you said:
146 > The Flat Earth Society has members all around the globe.
149 isn't really properly handled by Markdown, so let's do our best to
150 identify the individual elements, and mark them, using a syntax similar to
151 what pandoc uses already in some cases. As pandoc won't actually use these
152 data (yet?), we call `self._reformat_quotes` later to use these markers
153 to slap the appropriate classes on the HTML tags.
156 def generate_lines_with_context(mdwn):
158 Iterates the input string line-wise, returning a triplet of
159 previous, current, and next line, the first and last of which
160 will be None on the first and last line of the input data
163 prev = cur = nxt = None
164 lines = iter(mdwn.splitlines())
170 yield prev, cur, None
173 for prev, cur, nxt in generate_lines_with_context(mdwn):
175 # The lead-in to a quote is a single line immediately preceding the
176 # quote, and ending with ':'. Note that there could be multiple of
178 if re.match(r'^.+:\s*$', cur) and nxt.startswith('>'):
179 ret.append(f'{{.quotelead}}{cur.strip()}')
180 # pandoc needs an empty line before the blockquote, so
181 # we enter one for the purpose of HTML rendition:
185 # The first blockquote after such a lead-in gets marked as the
187 elif prev and re.match(r'^.+:\s*$', prev) and cur.startswith('>'):
188 ret.append(re.sub(r'^(\s*>\s*)+(.+)',
189 r'\g<1>{.quoteinitial}\g<2>',
190 cur, flags=re.MULTILINE))
192 # All other occurrences of blockquotes get the "subsequent" marker:
193 elif cur.startswith('>') and prev and not prev.startswith('>'):
194 ret.append(re.sub(r'^((?:\s*>\s*)+)(.+)',
195 r'\g<1>{.quotesubsequent}\g<2>',
196 cur, flags=re.MULTILINE))
198 else: # pass through everything else.
201 return '\n'.join(ret)
204 def _reformat_quotes(html):
206 Earlier in the pipeline, we marked email quoting, using markers, which we
207 now need to turn into HTML classes, so that we can use CSS to style them.
209 ret = html.replace('<p>{.quotelead}', '<p class="quotelead">')
210 ret = re.sub(r'<blockquote>\n((?:<blockquote>\n)*)<p>(?:\{\.quote(\w+)\})',
211 r'<blockquote class="quote \g<2>">\n\g<1><p>', ret, flags=re.MULTILINE)
216 def _convert_with_pandoc(mdwn, inputfmt='markdown', outputfmt='html5',
217 ext_enabled=None, ext_disabled=None,
218 standalone=True, title="HTML E-Mail"):
220 Invoke pandoc to do the actual conversion of Markdown to HTML5.
223 ext_enabled = [ 'backtick_code_blocks',
234 'all_symbols_escapable',
235 'intraword_underscores',
244 'tex_math_double_backslash',
248 ext_disabled = [ 'tex_math_single_backslash',
254 enabled = '+'.join(ext_enabled)
255 disabled = '-'.join(ext_disabled)
256 inputfmt = f'{inputfmt}+{enabled}-{disabled}'
260 args.append('--standalone')
262 args.append(f'--metadata=pagetitle:"{title}"')
264 return pypandoc.convert_text(mdwn, format=inputfmt, to=outputfmt,
268 def _apply_styling(html):
270 Inline all styles defined and used into the individual HTML tags.
272 return pynliner.Pynliner().from_string(html).with_cssString(DEFAULT_CSS).run()
275 def _postprocess_html(html):
277 Postprocess the generated and styled HTML.
282 def convert_markdown_to_html(mdwn):
284 Converts the input Markdown to HTML, handling separately the body, as well
285 as an optional signature.
287 parts = re.split(r'^-- $', mdwn, 1, flags=re.MULTILINE)
296 body = _preprocess_markdown(body)
297 body = _identify_quotes_for_later(body)
298 html = _convert_with_pandoc(body, standalone=False)
299 html = _reformat_quotes(html)
302 sig = _preprocess_markdown(sig)
303 html += SIGNATURE_HTML.format(sig='<br/>'.join(sig.splitlines()))
305 html = HTML_DOCUMENT.format(htmlbody=html)
306 html = _apply_styling(html)
307 html = _postprocess_html(html)
314 Convert text on stdin to HTML, and print it to stdout, like mutt would
317 html = convert_markdown_to_html(sys.stdin.read())
319 # mutt expects the content type in the first line, so:
320 print(f'text/html\n\n{html}')
323 if __name__ == '__main__':