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 .quotechar { display: none; }
66 .footnote-ref, .footnote-back { text-decoration: none;}
69 font-family: monospace;
75 border-collapse: collapse;
76 border: 1px solid #999;
78 th, td { padding: 0.5em; }
82 .even { background: #eee; }
85 STYLESHEET = os.path.join(os.path.expanduser('~/.mutt'),
87 if os.path.exists(STYLESHEET):
88 DEFAULT_CSS += open(STYLESHEET).read()
90 HTML_DOCUMENT = '''<!DOCTYPE html>
92 <meta http-equiv="Content-Type" content="text/html; charset=utf-8"/>
93 <meta charset="utf-8"/>
94 <meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes"/>
95 <title>HTML E-Mail</title>
96 </head><body class="email">
102 '<div class="signature"><span class="leader">-- </span>{sig}</div>'
105 def _preprocess_markdown(mdwn):
107 Preprocess Markdown for handling by the converter.
109 # convert hard line breaks within paragraphs to 2 trailing spaces, which
110 # is the markdown way of representing hard line breaks. Note how the
111 # regexp will not match between paragraphs.
112 ret = re.sub(r'(\S)\n(\s*\S)', r'\g<1> \n\g<2>', mdwn, flags=re.MULTILINE)
114 # Clients like Thunderbird need the leading '>' to be able to properly
115 # create nested quotes, so we duplicate the symbol, the first instance
116 # will tell pandoc to create a blockquote, while the second instance will
117 # be a <span> containing the character, along with a class that causes CSS
118 # to actually hide it from display. However, this does not work with the
119 # text-mode HTML2text converters, and so it's left commented for now.
120 #ret = re.sub(r'\n>', r' \n>[>]{.quotechar}', ret, flags=re.MULTILINE)
125 def _identify_quotes_for_later(mdwn):
127 Email quoting such as:
130 On 1970-01-01, you said:
131 > The Flat Earth Society has members all around the globe.
134 isn't really properly handled by Markdown, so let's do our best to
135 identify the individual elements, and mark them, using a syntax similar to
136 what pandoc uses already in some cases. As pandoc won't actually use these
137 data (yet?), we call `self._reformat_quotes` later to use these markers
138 to slap the appropriate classes on the HTML tags.
141 def generate_lines_with_context(mdwn):
143 Iterates the input string line-wise, returning a triplet of
144 previous, current, and next line, the first and last of which
145 will be None on the first and last line of the input data
148 prev = cur = nxt = None
149 lines = iter(mdwn.splitlines())
155 yield prev, cur, None
158 for prev, cur, nxt in generate_lines_with_context(mdwn):
160 # The lead-in to a quote is a single line immediately preceding the
161 # quote, and ending with ':'. Note that there could be multiple of
163 if re.match(r'^.+:\s*$', cur) and nxt.startswith('>'):
164 ret.append(f'{{.quotelead}}{cur.strip()}')
165 # pandoc needs an empty line before the blockquote, so
166 # we enter one for the purpose of HTML rendition:
170 # The first blockquote after such a lead-in gets marked as the
172 elif prev and re.match(r'^.+:\s*$', prev) and cur.startswith('>'):
173 ret.append(re.sub(r'^(\s*>\s*)+(.+)',
174 r'\g<1>{.quoteinitial}\g<2>',
175 cur, flags=re.MULTILINE))
177 # All other occurrences of blockquotes get the "subsequent" marker:
178 elif cur.startswith('>') and prev and not prev.startswith('>'):
179 ret.append(re.sub(r'^((?:\s*>\s*)+)(.+)',
180 r'\g<1>{.quotesubsequent}\g<2>',
181 cur, flags=re.MULTILINE))
183 else: # pass through everything else.
186 return '\n'.join(ret)
189 def _reformat_quotes(html):
191 Earlier in the pipeline, we marked email quoting, using markers, which we
192 now need to turn into HTML classes, so that we can use CSS to style them.
194 ret = html.replace('<p>{.quotelead}', '<p class="quotelead">')
195 ret = re.sub(r'<blockquote>\n((?:<blockquote>\n)*)<p>(?:\{\.quote(\w+)\})',
196 r'<blockquote class="quote \g<2>">\n\g<1><p>', ret, flags=re.MULTILINE)
201 def _convert_with_pandoc(mdwn, inputfmt='markdown', outputfmt='html5',
202 ext_enabled=None, ext_disabled=None,
203 standalone=True, title="HTML E-Mail"):
205 Invoke pandoc to do the actual conversion of Markdown to HTML5.
208 ext_enabled = [ 'backtick_code_blocks',
219 'all_symbols_escapable',
220 'intraword_underscores',
229 'tex_math_double_backslash',
232 ext_disabled = [ 'tex_math_single_backslash',
238 enabled = '+'.join(ext_enabled)
239 disabled = '-'.join(ext_disabled)
240 inputfmt = f'{inputfmt}+{enabled}-{disabled}'
244 args.append('--standalone')
246 args.append(f'--metadata=pagetitle:"{title}"')
248 return pypandoc.convert_text(mdwn, format=inputfmt, to=outputfmt,
252 def _apply_styling(html):
254 Inline all styles defined and used into the individual HTML tags.
256 return pynliner.Pynliner().from_string(html).with_cssString(DEFAULT_CSS).run()
259 def _postprocess_html(html):
261 Postprocess the generated and styled HTML.
266 def convert_markdown_to_html(mdwn):
268 Converts the input Markdown to HTML, handling separately the body, as well
269 as an optional signature.
271 parts = re.split(r'^-- $', mdwn, 1, flags=re.MULTILINE)
280 body = _preprocess_markdown(body)
281 body = _identify_quotes_for_later(body)
282 html = _convert_with_pandoc(body, standalone=False)
283 html = _reformat_quotes(html)
286 sig = _preprocess_markdown(sig)
287 html += SIGNATURE_HTML.format(sig='<br/>'.join(sig.splitlines()))
289 html = HTML_DOCUMENT.format(htmlbody=html)
290 html = _apply_styling(html)
291 html = _postprocess_html(html)
298 Convert text on stdin to HTML, and print it to stdout, like mutt would
301 html = convert_markdown_to_html(sys.stdin.read())
303 # mutt expects the content type in the first line, so:
304 print(f'text/html\n\n{html}')
307 if __name__ == '__main__':