]> 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:

move mutt config to ~/.config
[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 sig
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'<([^@]+@.+\.[^>]+)>', r'&lt;\g<1> -PANDOC_BUG_7398-&gt;', ret)
139
140     return ret
141
142
143 def _identify_quotes_for_later(mdwn):
144     '''
145     Email quoting such as:
146
147     ```
148     On 1970-01-01, you said:
149     > The Flat Earth Society has members all around the globe.
150     ```
151
152     isn't really properly handled by Markdown, so let's do our best to
153     identify the individual elements, and mark them, using a syntax similar to
154     what pandoc uses already in some cases. As pandoc won't actually use these
155     data (yet?), we call `self._reformat_quotes` later to use these markers
156     to slap the appropriate classes on the HTML tags.
157     '''
158
159     def generate_lines_with_context(mdwn):
160         '''
161         Iterates the input string line-wise, returning a triplet of
162         previous, current, and next line, the first and last of which
163         will be None on the first and last line of the input data
164         respectively.
165         '''
166         prev = cur = nxt = None
167         lines = iter(mdwn.splitlines())
168         cur = next(lines)
169         for nxt in lines:
170             yield prev, cur, nxt
171             prev = cur
172             cur = nxt
173         yield prev, cur, None
174
175     ret = []
176     for prev, cur, nxt in generate_lines_with_context(mdwn):
177
178         # The lead-in to a quote is a single line immediately preceding the
179         # quote, and ending with ':'. Note that there could be multiple of
180         # these:
181         if re.match(r'^[^>]+.*:\s*$', cur) and nxt.startswith('>'):
182             ret.append(f'{{.quotelead}}{cur.strip()}')
183             # pandoc needs an empty line before the blockquote, so
184             # we enter one for the purpose of HTML rendition:
185             ret.append('')
186             continue
187
188         # The first blockquote after such a lead-in gets marked as the
189         # "initial" quote:
190         elif prev and re.match(r'^[^>]+.*:\s*$', prev) and cur.startswith('>'):
191             ret.append(re.sub(r'^(\s*>\s*)+(.+)',
192                               r'\g<1>{.quoteinitial}\g<2>',
193                               cur, flags=re.MULTILINE))
194
195         # All other occurrences of blockquotes get the "subsequent" marker:
196         elif cur.startswith('>') and prev is not None and not prev.startswith('>'):
197             ret.append(re.sub(r'^((?:\s*>\s*)+)(.+)',
198                               r'\g<1>{.quotesubsequent}\g<2>',
199                               cur, flags=re.MULTILINE))
200
201         else: # pass through everything else.
202             ret.append(cur)
203
204     return '\n'.join(ret)
205
206
207 def _reformat_quotes(html):
208     '''
209     Earlier in the pipeline, we marked email quoting, using markers, which we
210     now need to turn into HTML classes, so that we can use CSS to style them.
211     '''
212     ret = html.replace('<p>{.quotelead}', '<p class="quotelead">')
213     ret = re.sub(r'<blockquote>\n((?:<blockquote>\n)*)<p>(?:\{\.quote(\w+)\})',
214                  r'<blockquote class="quote \g<2>">\n\g<1><p>', ret, flags=re.MULTILINE)
215     return ret
216
217
218
219 def _convert_with_pandoc(mdwn, inputfmt='markdown', outputfmt='html5',
220                          ext_enabled=None, ext_disabled=None,
221                          standalone=True, selfcontained=True, title=None):
222     '''
223     Invoke pandoc to do the actual conversion of Markdown to HTML5.
224     '''
225     if not ext_enabled:
226         ext_enabled = [ 'backtick_code_blocks',
227                        'line_blocks',
228                        'fancy_lists',
229                        'startnum',
230                        'definition_lists',
231                        'example_lists',
232                        'table_captions',
233                        'simple_tables',
234                        'multiline_tables',
235                        'grid_tables',
236                        'pipe_tables',
237                        'all_symbols_escapable',
238                        'intraword_underscores',
239                        'strikeout',
240                        'superscript',
241                        'subscript',
242                        'fenced_divs',
243                        'bracketed_spans',
244                        'footnotes',
245                        'inline_notes',
246                        'emoji',
247                        'tex_math_double_backslash',
248                        'autolink_bare_uris'
249                       ]
250     if not ext_disabled:
251         ext_disabled = [ 'tex_math_single_backslash',
252                          'tex_math_dollars',
253                          'smart',
254                          'raw_html'
255                        ]
256
257     enabled = '+'.join(ext_enabled)
258     disabled = '-'.join(ext_disabled)
259     inputfmt = f'{inputfmt}+{enabled}-{disabled}'
260
261     args = []
262     if standalone:
263         args.append('--standalone')
264     if selfcontained:
265         args.append('--self-contained')
266     if title:
267         args.append(f'--metadata=pagetitle:"{title}"')
268
269     return pypandoc.convert_text(mdwn, format=inputfmt, to=outputfmt,
270                                  extra_args=args)
271
272
273 def _apply_styling(html):
274     '''
275     Inline all styles defined and used into the individual HTML tags.
276     '''
277     return pynliner.Pynliner().from_string(html).with_cssString(DEFAULT_CSS).run()
278
279
280 def _postprocess_html(html):
281     '''
282     Postprocess the generated and styled HTML.
283     '''
284
285     # Preprocessing leaves a sentinel to work around
286     # https://github.com/jgm/pandoc/issues/7398, and so we need to remove it:
287     html = html.replace('</a> -PANDOC_BUG_7398-&gt;', '</a>&gt;')
288     return html
289
290
291 def convert_markdown_to_html(mdwn):
292     '''
293     Converts the input Markdown to HTML, handling separately the body, as well
294     as an optional signature.
295     '''
296     parts = re.split(r'^-- $', mdwn, 1, flags=re.MULTILINE)
297     body = parts[0]
298     if len(parts) == 2:
299         sig = parts[1]
300     else:
301         sig = None
302
303     html=''
304     if body:
305         body = _preprocess_markdown(body)
306         body = _identify_quotes_for_later(body)
307         html = _convert_with_pandoc(body, standalone=True, selfcontained=True,
308                                     title=None)
309         html = html.replace('<title>Untitled</title>\n','')
310         html = _reformat_quotes(html)
311
312     if sig:
313         sig = _preprocess_signature(sig)
314         sig = _preprocess_markdown(sig)
315         sig = _convert_with_pandoc(sig, standalone=False, selfcontained=False)
316         sig = SIGNATURE_HTML.format(sig='<br/>'.join(sig.splitlines()))
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()