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')
55 border-left: 2px solid #ccc;
65 .footnote-ref, .footnote-back { text-decoration: none;}
68 font-family: monospace;
74 STYLESHEET = os.path.join(os.path.expanduser('~/.mutt'),
76 if os.path.exists(STYLESHEET):
77 DEFAULT_CSS += open(STYLESHEET).read()
79 HTML_DOCUMENT = '''<!DOCTYPE html>
81 <meta http-equiv="Content-Type" content="text/html; charset=utf-8"/>
82 <meta charset="utf-8"/>
83 <meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes"/>
84 <title>HTML E-Mail</title>
85 </head><body class="email">
91 '<div class="signature"><span class="leader">-- </span>{sig}</div>'
94 def _preprocess_markdown(mdwn):
96 Preprocess Markdown for handling by the converter.
98 # convert hard line breaks within paragraphs to 2 trailing spaces, which
99 # is the markdown way of representing hard line breaks. Note how the
100 # regexp will not match between paragraphs.
101 ret = re.sub(r'(\S)\n(\s*\S)', r'\g<1> \n\g<2>', mdwn, re.MULTILINE)
106 def _identify_quotes_for_later(mdwn):
108 Email quoting such as:
111 On 1970-01-01, you said:
112 > The Flat Earth Society has members all around the globe.
115 isn't really properly handled by Markdown, so let's do our best to
116 identify the individual elements, and mark them, using a syntax similar to
117 what pandoc uses already in some cases. As pandoc won't actually use these
118 data (yet?), we call `self._reformat_quotes` later to use these markers
119 to slap the appropriate classes on the HTML tags.
122 def generate_lines_with_context(mdwn):
124 Iterates the input string line-wise, returning a triplet of
125 previous, current, and next line, the first and last of which
126 will be None on the first and last line of the input data
129 prev = cur = nxt = None
130 lines = iter(mdwn.splitlines())
136 yield prev, cur, None
139 for prev, cur, nxt in generate_lines_with_context(mdwn):
141 # The lead-in to a quote is a single line immediately preceding the
142 # quote, and ending with ':'. Note that there could be multiple of
144 if re.match(r'^.+:\s*$', cur) and nxt.startswith('>'):
145 ret.append(f'{{.quotelead}}{cur.strip()}')
146 # pandoc needs an empty line before the blockquote, so
147 # we enter one for the purpose of HTML rendition:
151 # The first blockquote after such a lead-in gets marked as the
153 elif prev and re.match(r'^.+:\s*$', prev) and cur.startswith('>'):
154 ret.append(re.sub(r'^(\s*>\s*)+(.+)',
155 r'\g<1>{.quoteinitial}\g<2>',
158 # All other occurrences of blockquotes get the "subsequent" marker:
159 elif cur.startswith('>') and not prev.startswith('>'):
160 ret.append(re.sub(r'^((?:\s*>\s*)+)(.+)',
161 r'\g<1>{.quotesubsequent}\g<2>',
164 else: # pass through everything else.
167 return '\n'.join(ret)
170 def _reformat_quotes(html):
172 Earlier in the pipeline, we marked email quoting, using markers, which we
173 now need to turn into HTML classes, so that we can use CSS to style them.
175 ret = html.replace('<p>{.quotelead}', '<p class="quotelead">')
176 ret = re.sub(r'<blockquote>\n((?:<blockquote>\n)*)<p>(?:\{\.quote(\w+)\})',
177 r'<blockquote class="quote \g<2>">\n\g<1><p>', ret, re.MULTILINE)
182 def _convert_with_pandoc(mdwn, inputfmt='markdown', outputfmt='html5',
183 ext_enabled=None, ext_disabled=None,
184 standalone=True, title="HTML E-Mail"):
186 Invoke pandoc to do the actual conversion of Markdown to HTML5.
189 ext_enabled = [ 'backtick_code_blocks',
200 'all_symbols_escapable',
201 'intraword_underscores',
210 'tex_math_double_backslash',
213 ext_disabled = [ 'tex_math_single_backslash',
218 enabled = '+'.join(ext_enabled)
219 disabled = '-'.join(ext_disabled)
220 inputfmt = f'{inputfmt}+{enabled}-{disabled}'
224 args.append('--standalone')
226 args.append(f'--metadata=pagetitle:"{title}"')
228 return pypandoc.convert_text(mdwn, format=inputfmt, to=outputfmt,
232 def _apply_styling(html):
234 Inline all styles defined and used into the individual HTML tags.
236 return pynliner.Pynliner().from_string(html).with_cssString(DEFAULT_CSS).run()
239 def _postprocess_html(html):
241 Postprocess the generated and styled HTML.
246 def convert_markdown_to_html(mdwn):
248 Converts the input Markdown to HTML, handling separately the body, as well
249 as an optional signature.
251 parts = re.split(r'^-- $', mdwn, 1, flags=re.MULTILINE)
260 body = _preprocess_markdown(body)
261 body = _identify_quotes_for_later(body)
262 html = _convert_with_pandoc(body, standalone=False)
263 html = _reformat_quotes(html)
266 sig = _preprocess_markdown(sig)
267 html += SIGNATURE_HTML.format(sig='<br/>'.join(sig.splitlines()))
269 html = HTML_DOCUMENT.format(htmlbody=html)
270 html = _apply_styling(html)
271 html = _postprocess_html(html)
278 Convert text on stdin to HTML, and print it to stdout, like mutt would
281 html = convert_markdown_to_html(sys.stdin.read())
283 # mutt expects the content type in the first line, so:
284 print(f'text/html\n\n{html}')
287 if __name__ == '__main__':