]> git.madduck.net Git - etc/mutt.git/blob - .mutt/markdown2html

madduck's git repository

Every one of the projects in this repository is available at the canonical URL git://git.madduck.net/madduck/pub/<projectpath> — see each project's metadata for the exact URL.

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.

SSH access, as well as push access can be individually arranged.

If you use my repositories frequently, consider adding the following snippet to ~/.gitconfig and using the third clone URL listed for each project:

[url "git://git.madduck.net/madduck/"]
  insteadOf = madduck:

8930c88cc91d3bdb15383f1c709b69978b63a559
[etc/mutt.git] / .mutt / markdown2html
1 #!/usr/bin/python3
2 #
3 # markdown2html.py — simple Markdown-to-HTML converter for use with Mutt
4 #
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.
9 #
10 # [1]: https://gitlab.com/muttmua/mutt/commit/0e566a03725b4ad789aa6ac1d17cdf7bf4e7e354)
11 #
12 # Configuration:
13 #   muttrc:
14 #     set send_multipart_alternative=yes
15 #     set send_multipart_alternative_filter=/path/to/markdown2html.py
16 #
17 # Optionally, Custom CSS styles will be read from `~/.mutt/markdown2html.css`,
18 # if present.
19 #
20 # Requirements:
21 #   - python3
22 #   - PyPandoc (and pandoc installed, or downloaded)
23 #   - Pynliner
24 #
25 # Optional:
26 #   - Pygments, if installed, then syntax highlighting is enabled
27 #
28 # Latest version:
29 #   https://git.madduck.net/etc/mutt.git/blob_plain/HEAD:/.mutt/markdown2html
30 #
31 # Copyright © 2019 martin f. krafft <madduck@madduck.net>
32 # Released under the GPL-2+ licence, just like Mutt itself.
33 #
34
35 import pypandoc
36 import pynliner
37 import re
38 import os
39 import sys
40
41 try:
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')
45
46 except ImportError:
47     DEFAULT_CSS = ""
48
49
50 DEFAULT_CSS += '''
51 .block {
52     padding: 0 0.5em;
53     margin: 0;
54     border-left: 2px solid #eee;
55 }
56 .quote, blockquote {
57     padding: 0 0.5em;
58     margin: 0;
59     font-style: italic;
60     border-left: 2px solid #666;
61     color: #666;
62     font-size: 80%;
63 }
64 .quotelead {
65     margin-bottom: -1em;
66     font-size: 80%;
67 }
68 .quotechar { display: none; }
69 .footnote-ref, .footnote-back { text-decoration: none;}
70 .signature {
71     color: #999;
72     font-family: monospace;
73     white-space: pre;
74     margin: 1em 0 0 0;
75     font-size: 80%;
76 }
77 table, th, td {
78     border-collapse: collapse;
79     border: 1px solid #999;
80 }
81 th, td { padding: 0.5em; }
82 .header {
83     background: #eee;
84 }
85 .even { background: #eee; }
86 h1, h2, h3, h4, h5, h6 {
87     color: #666;
88     background-color: #eee;
89     padding-left: 0.5em
90 }
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; }
98 '''
99
100 STYLESHEET = os.path.join(os.path.expanduser('~/.mutt'),
101                           'markdown2html.css')
102 if os.path.exists(STYLESHEET):
103     DEFAULT_CSS += open(STYLESHEET).read()
104
105 HTML_DOCUMENT = '''<!DOCTYPE html>
106 <html><head>
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 </head><body class="email">
111 {htmlbody}
112 </body></html>'''
113
114
115 SIGNATURE_HTML = \
116         '<div class="signature"><span class="leader">-- </span>{sig}</div>'
117
118
119 def _preprocess_signature(sig):
120     '''
121     Preprocess the signature before markdown processing.
122     '''
123     return sig
124
125 def _preprocess_markdown(mdwn):
126     '''
127     Preprocess Markdown for handling by the converter.
128     '''
129     # convert hard line breaks within paragraphs to 2 trailing spaces, which
130     # is the markdown way of representing hard line breaks. Note how the
131     # regexp will not match between paragraphs.
132     ret = re.sub(r'(\S)\n(\s*\S)', r'\g<1>  \n\g<2>', mdwn, flags=re.MULTILINE)
133
134     # Clients like Thunderbird need the leading '>' to be able to properly
135     # create nested quotes, so we duplicate the symbol, the first instance
136     # will tell pandoc to create a blockquote, while the second instance will
137     # be a <span> containing the character, along with a class that causes CSS
138     # to actually hide it from display. However, this does not work with the
139     # text-mode HTML2text converters, and so it's left commented for now.
140     #ret = re.sub(r'\n>', r'  \n>[>]{.quotechar}', ret, flags=re.MULTILINE)
141
142     # With the autolink_bare_uris extension, we do not need to put links into
143     # angle brackets to have them converted, so let's conserve the brackets
144     # when used around email addresses. Note that this needs a postprocessing
145     # hack because the pandoc autolink converted includes the ambersand
146     # (https://github.com/jgm/pandoc/issues/7398).
147     ret = re.sub(r'<([^@]+@.+\.[^>]+)>', r'&lt;\g<1> -PANDOC_BUG_7398-&gt;', ret)
148
149     return ret
150
151
152 def _identify_quotes_for_later(mdwn):
153     '''
154     Email quoting such as:
155
156     ```
157     On 1970-01-01, you said:
158     > The Flat Earth Society has members all around the globe.
159     ```
160
161     isn't really properly handled by Markdown, so let's do our best to
162     identify the individual elements, and mark them, using a syntax similar to
163     what pandoc uses already in some cases. As pandoc won't actually use these
164     data (yet?), we call `self._reformat_quotes` later to use these markers
165     to slap the appropriate classes on the HTML tags.
166     '''
167
168     def generate_lines_with_context(mdwn):
169         '''
170         Iterates the input string line-wise, returning a triplet of
171         previous, current, and next line, the first and last of which
172         will be None on the first and last line of the input data
173         respectively.
174         '''
175         prev = cur = nxt = None
176         lines = iter(mdwn.splitlines())
177         cur = next(lines)
178         for nxt in lines:
179             yield prev, cur, nxt
180             prev = cur
181             cur = nxt
182         yield prev, cur, None
183
184     ret = []
185     for prev, cur, nxt in generate_lines_with_context(mdwn):
186
187         # The lead-in to a quote is a single line immediately preceding the
188         # quote, and ending with ':'. Note that there could be multiple of
189         # these:
190         if re.match(r'^.+:\s*$', cur) and nxt.startswith('>'):
191             ret.append(f'{{.quotelead}}{cur.strip()}')
192             # pandoc needs an empty line before the blockquote, so
193             # we enter one for the purpose of HTML rendition:
194             ret.append('')
195             continue
196
197         # The first blockquote after such a lead-in gets marked as the
198         # "initial" quote:
199         elif prev and re.match(r'^.+:\s*$', prev) and cur.startswith('>'):
200             ret.append(re.sub(r'^(\s*>\s*)+(.+)',
201                               r'\g<1>{.quoteinitial}\g<2>',
202                               cur, flags=re.MULTILINE))
203
204         # All other occurrences of blockquotes get the "subsequent" marker:
205         elif cur.startswith('>') and prev and not prev.startswith('>'):
206             ret.append(re.sub(r'^((?:\s*>\s*)+)(.+)',
207                               r'\g<1>{.quotesubsequent}\g<2>',
208                               cur, flags=re.MULTILINE))
209
210         else: # pass through everything else.
211             ret.append(cur)
212
213     return '\n'.join(ret)
214
215
216 def _reformat_quotes(html):
217     '''
218     Earlier in the pipeline, we marked email quoting, using markers, which we
219     now need to turn into HTML classes, so that we can use CSS to style them.
220     '''
221     ret = html.replace('<p>{.quotelead}', '<p class="quotelead">')
222     ret = re.sub(r'<blockquote>\n((?:<blockquote>\n)*)<p>(?:\{\.quote(\w+)\})',
223                  r'<blockquote class="quote \g<2>">\n\g<1><p>', ret, flags=re.MULTILINE)
224     return ret
225
226
227
228 def _convert_with_pandoc(mdwn, inputfmt='markdown', outputfmt='html5',
229                          ext_enabled=None, ext_disabled=None,
230                          standalone=True, selfcontained=True, title=None):
231     '''
232     Invoke pandoc to do the actual conversion of Markdown to HTML5.
233     '''
234     if not ext_enabled:
235         ext_enabled = [ 'backtick_code_blocks',
236                        'line_blocks',
237                        'fancy_lists',
238                        'startnum',
239                        'definition_lists',
240                        'example_lists',
241                        'table_captions',
242                        'simple_tables',
243                        'multiline_tables',
244                        'grid_tables',
245                        'pipe_tables',
246                        'all_symbols_escapable',
247                        'intraword_underscores',
248                        'strikeout',
249                        'superscript',
250                        'subscript',
251                        'fenced_divs',
252                        'bracketed_spans',
253                        'footnotes',
254                        'inline_notes',
255                        'emoji',
256                        'tex_math_double_backslash',
257                        'autolink_bare_uris'
258                       ]
259     if not ext_disabled:
260         ext_disabled = [ 'tex_math_single_backslash',
261                          'tex_math_dollars',
262                          'smart',
263                          'raw_html'
264                        ]
265
266     enabled = '+'.join(ext_enabled)
267     disabled = '-'.join(ext_disabled)
268     inputfmt = f'{inputfmt}+{enabled}-{disabled}'
269
270     args = []
271     if standalone:
272         args.append('--standalone')
273     if selfcontained:
274         args.append('--self-contained')
275     if title:
276         args.append(f'--metadata=pagetitle:"{title}"')
277
278     return pypandoc.convert_text(mdwn, format=inputfmt, to=outputfmt,
279                                  extra_args=args)
280
281
282 def _apply_styling(html):
283     '''
284     Inline all styles defined and used into the individual HTML tags.
285     '''
286     return pynliner.Pynliner().from_string(html).with_cssString(DEFAULT_CSS).run()
287
288
289 def _postprocess_html(html):
290     '''
291     Postprocess the generated and styled HTML.
292     '''
293
294     # Preprocessing leaves a sentinel to work around
295     # https://github.com/jgm/pandoc/issues/7398, and so we need to remove it:
296     html = html.replace('</a> -PANDOC_BUG_7398-&gt;', '</a>&gt;')
297     return html
298
299
300 def convert_markdown_to_html(mdwn):
301     '''
302     Converts the input Markdown to HTML, handling separately the body, as well
303     as an optional signature.
304     '''
305     parts = re.split(r'^-- $', mdwn, 1, flags=re.MULTILINE)
306     body = parts[0]
307     if len(parts) == 2:
308         sig = parts[1]
309     else:
310         sig = None
311
312     html=''
313     if body:
314         body = _preprocess_markdown(body)
315         body = _identify_quotes_for_later(body)
316         html = _convert_with_pandoc(body, standalone=True, selfcontained=True)
317         html = _reformat_quotes(html)
318
319     if sig:
320         sig = _preprocess_signature(sig)
321         sig = _preprocess_markdown(sig)
322         sig = _convert_with_pandoc(sig, standalone=False, selfcontained=False)
323         sig = SIGNATURE_HTML.format(sig='<br/>'.join(sig.splitlines()))
324         eob = html.find('</body>')
325         html = f'{html[:eob]}{sig}\n{html[eob:]}'
326
327     html = HTML_DOCUMENT.format(htmlbody=html)
328     html = _apply_styling(html)
329     html = _postprocess_html(html)
330
331     return html
332
333
334 def main():
335     '''
336     Convert text on stdin to HTML, and print it to stdout, like mutt would
337     expect.
338     '''
339     html = convert_markdown_to_html(sys.stdin.read())
340     if html:
341         # mutt expects the content type in the first line, so:
342         print(f'text/html\n\n{html}')
343
344
345 if __name__ == '__main__':
346     main()