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

Add comment about better mkconf alternative, thanks flatcap
[etc/mutt.git] / .config / 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 `~/.config/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 pre { padding: 0 1em; }
99 '''
100
101 STYLESHEET = os.path.join(os.path.expanduser('~/.config/mutt'),
102                           'markdown2html.css')
103 if os.path.exists(STYLESHEET):
104     DEFAULT_CSS += open(STYLESHEET).read()
105
106 SIGNATURE_HTML = \
107         '<div class="signature"><span class="leader">-- </span>{sig}</div>'
108
109
110 def _preprocess_signature(sig):
111     '''
112     Preprocess the signature before markdown processing.
113     '''
114     return f"```\n{sig}\n```"
115
116 def _preprocess_markdown(mdwn):
117     '''
118     Preprocess Markdown for handling by the converter.
119     '''
120     # convert hard line breaks within paragraphs to 2 trailing spaces, which
121     # is the markdown way of representing hard line breaks. Note how the
122     # regexp will not match between paragraphs.
123     ret = re.sub(r'(\S)\n(\s*\S)', r'\g<1>  \n\g<2>', mdwn, flags=re.MULTILINE)
124
125     # Clients like Thunderbird need the leading '>' to be able to properly
126     # create nested quotes, so we duplicate the symbol, the first instance
127     # will tell pandoc to create a blockquote, while the second instance will
128     # be a <span> containing the character, along with a class that causes CSS
129     # to actually hide it from display. However, this does not work with the
130     # text-mode HTML2text converters, and so it's left commented for now.
131     #ret = re.sub(r'\n>', r'  \n>[>]{.quotechar}', ret, flags=re.MULTILINE)
132
133     # With the autolink_bare_uris extension, we do not need to put links into
134     # angle brackets to have them converted, so let's conserve the brackets
135     # when used around email addresses. Note that this needs a postprocessing
136     # hack because the pandoc autolink converted includes the ambersand
137     # (https://github.com/jgm/pandoc/issues/7398).
138     ret = re.sub(r'<([^@]+@\S+)>', r'&lt;\g<1> PANDOC_BUG_7398 &gt;', ret)
139     return ret
140
141
142 def _identify_quotes_for_later(mdwn):
143     '''
144     Email quoting such as:
145
146     ```
147     On 1970-01-01, you said:
148     > The Flat Earth Society has members all around the globe.
149     ```
150
151     isn't really properly handled by Markdown, so let's do our best to
152     identify the individual elements, and mark them, using a syntax similar to
153     what pandoc uses already in some cases. As pandoc won't actually use these
154     data (yet?), we call `self._reformat_quotes` later to use these markers
155     to slap the appropriate classes on the HTML tags.
156     '''
157
158     def generate_lines_with_context(mdwn):
159         '''
160         Iterates the input string line-wise, returning a triplet of
161         previous, current, and next line, the first and last of which
162         will be None on the first and last line of the input data
163         respectively.
164         '''
165         prev = cur = nxt = None
166         lines = iter(mdwn.splitlines())
167         cur = next(lines)
168         for nxt in lines:
169             yield prev, cur, nxt
170             prev = cur
171             cur = nxt
172         yield prev, cur, None
173
174     ret = []
175     for prev, cur, nxt in generate_lines_with_context(mdwn):
176
177         # The lead-in to a quote is a single line immediately preceding the
178         # quote, and ending with ':'. Note that there could be multiple of
179         # these:
180         if re.match(r'^[^>]+.*:\s*$', cur) and nxt.startswith('>'):
181             ret.append(f'{{.quotelead}}{cur.strip()}')
182             # pandoc needs an empty line before the blockquote, so
183             # we enter one for the purpose of HTML rendition:
184             ret.append('')
185             continue
186
187         # The first blockquote after such a lead-in gets marked as the
188         # "initial" quote:
189         elif prev and re.match(r'^[^>]+.*:\s*$', prev) and cur.startswith('>'):
190             ret.append(re.sub(r'^(\s*>\s*)+(.+)',
191                               r'\g<1>{.quoteinitial}\g<2>',
192                               cur, flags=re.MULTILINE))
193
194         # All other occurrences of blockquotes get the "subsequent" marker:
195         elif cur.startswith('>') and prev is not None and not prev.startswith('>'):
196             ret.append(re.sub(r'^((?:\s*>\s*)+)(.+)',
197                               r'\g<1>{.quotesubsequent}\g<2>',
198                               cur, flags=re.MULTILINE))
199
200         else: # pass through everything else.
201             ret.append(cur)
202
203     return '\n'.join(ret)
204
205
206 def _reformat_quotes(html):
207     '''
208     Earlier in the pipeline, we marked email quoting, using markers, which we
209     now need to turn into HTML classes, so that we can use CSS to style them.
210     '''
211     ret = html.replace('<p>{.quotelead}', '<p class="quotelead">')
212     ret = re.sub(r'<blockquote>\n((?:<blockquote>\n)*)<p>(?:\{\.quote(\w+)\})',
213                  r'<blockquote class="quote \g<2>">\n\g<1><p>', ret, flags=re.MULTILINE)
214     return ret
215
216
217
218 def _convert_with_pandoc(mdwn, inputfmt='markdown', outputfmt='html5',
219                          ext_enabled=None, ext_disabled=None,
220                          standalone=True, selfcontained=True, title=None):
221     '''
222     Invoke pandoc to do the actual conversion of Markdown to HTML5.
223     '''
224     if not ext_enabled:
225         ext_enabled = [ 'backtick_code_blocks',
226                        'line_blocks',
227                        'fancy_lists',
228                        'startnum',
229                        'definition_lists',
230                        'example_lists',
231                        'table_captions',
232                        'simple_tables',
233                        'multiline_tables',
234                        'grid_tables',
235                        'pipe_tables',
236                        'all_symbols_escapable',
237                        'intraword_underscores',
238                        'strikeout',
239                        'superscript',
240                        'subscript',
241                        'fenced_divs',
242                        'bracketed_spans',
243                        'footnotes',
244                        'inline_notes',
245                        'emoji',
246                        'tex_math_double_backslash',
247                        'autolink_bare_uris'
248                       ]
249     if not ext_disabled:
250         ext_disabled = [ 'tex_math_single_backslash',
251                          'tex_math_dollars',
252                          'smart',
253                          'raw_html'
254                        ]
255
256     enabled = '+'.join(ext_enabled)
257     disabled = '-'.join(ext_disabled)
258     inputfmt = f'{inputfmt}+{enabled}-{disabled}'
259
260     args = []
261     if standalone:
262         args.append('--standalone')
263     if selfcontained:
264         args.append('--self-contained')
265     if title:
266         args.append(f'--metadata=pagetitle:"{title}"')
267
268     return pypandoc.convert_text(mdwn, format=inputfmt, to=outputfmt,
269                                  extra_args=args)
270
271
272 def _apply_styling(html):
273     '''
274     Inline all styles defined and used into the individual HTML tags.
275     '''
276     return pynliner.Pynliner().from_string(html).with_cssString(DEFAULT_CSS).run()
277
278
279 def _postprocess_html(html):
280     '''
281     Postprocess the generated and styled HTML.
282     '''
283
284     # Preprocessing leaves a sentinel to work around
285     # https://github.com/jgm/pandoc/issues/7398, and so we need to remove it:
286     html = html.replace(' PANDOC_BUG_7398 ', '')
287     return html
288
289
290 def convert_markdown_to_html(mdwn):
291     '''
292     Converts the input Markdown to HTML, handling separately the body, as well
293     as an optional signature.
294     '''
295     parts = re.split(r'^-- \n', mdwn, 1, flags=re.MULTILINE)
296     body = parts[0]
297     if len(parts) == 2:
298         sig = parts[1]
299     else:
300         sig = None
301
302     html=''
303     if body:
304         body = _preprocess_markdown(body)
305         body = _identify_quotes_for_later(body)
306         html = _convert_with_pandoc(body, standalone=True, selfcontained=True,
307                                     title=None)
308         html = html.replace('<title>Untitled</title>\n','')
309         html = _reformat_quotes(html)
310
311     if sig:
312         sig = _preprocess_signature(sig)
313         sig = _preprocess_markdown(sig)
314         print(sig, file=sys.stderr, end='')
315         sig = _convert_with_pandoc(sig, standalone=False, selfcontained=False)
316         sig = SIGNATURE_HTML.format(sig=sig)
317         eob = html.find('</body>')
318         html = f'{html[:eob]}{sig}\n{html[eob:]}'
319
320     html = _apply_styling(html)
321     html = _postprocess_html(html)
322
323     return html
324
325
326 def main():
327     '''
328     Convert text on stdin to HTML, and print it to stdout, like mutt would
329     expect.
330     '''
331     html = convert_markdown_to_html(sys.stdin.read())
332     if html:
333         # mutt expects the content type in the first line, so:
334         print(f'text/html\n\n{html}')
335
336
337 if __name__ == '__main__':
338     main()