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 #666;
63 .quotechar { display: none; }
64 .footnote-ref, .footnote-back { text-decoration: none;}
67 font-family: monospace;
73 border-collapse: collapse;
74 border: 1px solid #999;
76 th, td { padding: 0.5em; }
80 .even { background: #eee; }
81 h1, h2, h3, h4, h5, h6 {
83 background-color: #eee;
86 h1 { font-size: 130%; }
87 h2 { font-size: 120%; }
88 h3 { font-size: 110%; }
89 h4 { font-size: 107%; }
90 h5 { font-size: 103%; }
91 h6 { font-size: 100%; }
92 p { padding: 0 0.5em; }
95 STYLESHEET = os.path.join(os.path.expanduser('~/.mutt'),
97 if os.path.exists(STYLESHEET):
98 DEFAULT_CSS += open(STYLESHEET).read()
100 HTML_DOCUMENT = '''<!DOCTYPE html>
102 <meta http-equiv="Content-Type" content="text/html; charset=utf-8"/>
103 <meta charset="utf-8"/>
104 <meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes"/>
105 <title>HTML E-Mail</title>
106 </head><body class="email">
112 '<div class="signature"><span class="leader">-- </span>{sig}</div>'
115 def _preprocess_markdown(mdwn):
117 Preprocess Markdown for handling by the converter.
119 # convert hard line breaks within paragraphs to 2 trailing spaces, which
120 # is the markdown way of representing hard line breaks. Note how the
121 # regexp will not match between paragraphs.
122 ret = re.sub(r'(\S)\n(\s*\S)', r'\g<1> \n\g<2>', mdwn, flags=re.MULTILINE)
124 # Clients like Thunderbird need the leading '>' to be able to properly
125 # create nested quotes, so we duplicate the symbol, the first instance
126 # will tell pandoc to create a blockquote, while the second instance will
127 # be a <span> containing the character, along with a class that causes CSS
128 # to actually hide it from display. However, this does not work with the
129 # text-mode HTML2text converters, and so it's left commented for now.
130 #ret = re.sub(r'\n>', r' \n>[>]{.quotechar}', ret, flags=re.MULTILINE)
135 def _identify_quotes_for_later(mdwn):
137 Email quoting such as:
140 On 1970-01-01, you said:
141 > The Flat Earth Society has members all around the globe.
144 isn't really properly handled by Markdown, so let's do our best to
145 identify the individual elements, and mark them, using a syntax similar to
146 what pandoc uses already in some cases. As pandoc won't actually use these
147 data (yet?), we call `self._reformat_quotes` later to use these markers
148 to slap the appropriate classes on the HTML tags.
151 def generate_lines_with_context(mdwn):
153 Iterates the input string line-wise, returning a triplet of
154 previous, current, and next line, the first and last of which
155 will be None on the first and last line of the input data
158 prev = cur = nxt = None
159 lines = iter(mdwn.splitlines())
165 yield prev, cur, None
168 for prev, cur, nxt in generate_lines_with_context(mdwn):
170 # The lead-in to a quote is a single line immediately preceding the
171 # quote, and ending with ':'. Note that there could be multiple of
173 if re.match(r'^.+:\s*$', cur) and nxt.startswith('>'):
174 ret.append(f'{{.quotelead}}{cur.strip()}')
175 # pandoc needs an empty line before the blockquote, so
176 # we enter one for the purpose of HTML rendition:
180 # The first blockquote after such a lead-in gets marked as the
182 elif prev and re.match(r'^.+:\s*$', prev) and cur.startswith('>'):
183 ret.append(re.sub(r'^(\s*>\s*)+(.+)',
184 r'\g<1>{.quoteinitial}\g<2>',
185 cur, flags=re.MULTILINE))
187 # All other occurrences of blockquotes get the "subsequent" marker:
188 elif cur.startswith('>') and prev and not prev.startswith('>'):
189 ret.append(re.sub(r'^((?:\s*>\s*)+)(.+)',
190 r'\g<1>{.quotesubsequent}\g<2>',
191 cur, flags=re.MULTILINE))
193 else: # pass through everything else.
196 return '\n'.join(ret)
199 def _reformat_quotes(html):
201 Earlier in the pipeline, we marked email quoting, using markers, which we
202 now need to turn into HTML classes, so that we can use CSS to style them.
204 ret = html.replace('<p>{.quotelead}', '<p class="quotelead">')
205 ret = re.sub(r'<blockquote>\n((?:<blockquote>\n)*)<p>(?:\{\.quote(\w+)\})',
206 r'<blockquote class="quote \g<2>">\n\g<1><p>', ret, flags=re.MULTILINE)
211 def _convert_with_pandoc(mdwn, inputfmt='markdown', outputfmt='html5',
212 ext_enabled=None, ext_disabled=None,
213 standalone=True, title="HTML E-Mail"):
215 Invoke pandoc to do the actual conversion of Markdown to HTML5.
218 ext_enabled = [ 'backtick_code_blocks',
229 'all_symbols_escapable',
230 'intraword_underscores',
239 'tex_math_double_backslash',
243 ext_disabled = [ 'tex_math_single_backslash',
249 enabled = '+'.join(ext_enabled)
250 disabled = '-'.join(ext_disabled)
251 inputfmt = f'{inputfmt}+{enabled}-{disabled}'
255 args.append('--standalone')
257 args.append(f'--metadata=pagetitle:"{title}"')
259 return pypandoc.convert_text(mdwn, format=inputfmt, to=outputfmt,
263 def _apply_styling(html):
265 Inline all styles defined and used into the individual HTML tags.
267 return pynliner.Pynliner().from_string(html).with_cssString(DEFAULT_CSS).run()
270 def _postprocess_html(html):
272 Postprocess the generated and styled HTML.
277 def convert_markdown_to_html(mdwn):
279 Converts the input Markdown to HTML, handling separately the body, as well
280 as an optional signature.
282 parts = re.split(r'^-- $', mdwn, 1, flags=re.MULTILINE)
291 body = _preprocess_markdown(body)
292 body = _identify_quotes_for_later(body)
293 html = _convert_with_pandoc(body, standalone=False)
294 html = _reformat_quotes(html)
297 sig = _preprocess_markdown(sig)
298 html += SIGNATURE_HTML.format(sig='<br/>'.join(sig.splitlines()))
300 html = HTML_DOCUMENT.format(htmlbody=html)
301 html = _apply_styling(html)
302 html = _postprocess_html(html)
309 Convert text on stdin to HTML, and print it to stdout, like mutt would
312 html = convert_markdown_to_html(sys.stdin.read())
314 # mutt expects the content type in the first line, so:
315 print(f'text/html\n\n{html}')
318 if __name__ == '__main__':