]> git.madduck.net Git - etc/neomutt.git/blob - .config/neomutt/buildmimetree.py

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:

fbc8040061877b74cc0e9fbc4a8e605d21a8a33c
[etc/neomutt.git] / .config / neomutt / buildmimetree.py
1 #!/usr/bin/python3
2 #
3 # NeoMutt helper script to create multipart/* emails with Markdown → HTML
4 # alternative conversion, and handling of inline images, using NeoMutt's
5 # ability to manually craft MIME trees, but automating this process.
6 #
7 # Configuration:
8 #   neomuttrc (needs to be a single line):
9 #     set my_mdwn_extensions="extra,admonition,codehilite,sane_lists,smarty"
10 #     macro compose B "\
11 #       <enter-command> source '$my_confdir/buildmimetree.py \
12 #       --tempdir $tempdir --extensions $my_mdwn_extensions \
13 #       --css-file $my_confdir/htmlmail.css |'<enter>\
14 #       <enter-command> sourc e \$my_mdwn_postprocess_cmd_file<enter>\
15 #     " "Convert message into a modern MIME tree with inline images"
16 #
17 #     (Yes, we need to call source twice, as mutt only starts to process output
18 #     from a source command when the command exits, and since we need to react
19 #     to the output, we need to be invoked again, using a $my_ variable to pass
20 #     information)
21 #
22 # Requirements:
23 #   - python3
24 #   - python3-markdown
25 #   - python3-beautifulsoup4
26 # Optional:
27 #   - pytest
28 #   - Pynliner, provides --css-file and thus inline styling of HTML output
29 #   - Pygments, then syntax highlighting for fenced code is enabled
30 #
31 # Latest version:
32 #   https://git.madduck.net/etc/neomutt.git/blob_plain/HEAD:/.config/neomutt/buildmimetree.py
33 #
34 # Copyright © 2023 martin f. krafft <madduck@madduck.net>
35 # Released under the GPL-2+ licence, just like Mutt itself.
36 #
37
38 import sys
39 import os
40 import os.path
41 import pathlib
42 import markdown
43 import tempfile
44 import argparse
45 import re
46 import mimetypes
47 import bs4
48 from collections import namedtuple, OrderedDict
49 from markdown.extensions import Extension
50 from markdown.inlinepatterns import ImageInlineProcessor, IMAGE_LINK_RE
51 from email.utils import make_msgid
52 from urllib import request
53
54
55 def parse_cli_args(*args, **kwargs):
56     parser = argparse.ArgumentParser(
57         description=(
58             "NeoMutt helper to turn text/markdown email parts "
59             "into full-fledged MIME trees"
60         )
61     )
62     parser.epilog = (
63         "Copyright © 2023 martin f. krafft <madduck@madduck.net>.\n"
64         "Released under the MIT licence"
65     )
66
67     parser.add_argument(
68         "--extensions",
69         metavar="EXT[,EXT[,EXT]]",
70         type=str,
71         default="",
72         help="Markdown extension to use (comma-separated list)",
73     )
74
75     if _PYNLINER:
76         parser.add_argument(
77             "--css-file",
78             metavar="FILE",
79             type=pathlib.Path,
80             default=os.devnull,
81             help="CSS file to merge with the final HTML",
82         )
83     else:
84         parser.set_defaults(css_file=None)
85
86     parser.add_argument(
87         "--related-to-html-only",
88         action="store_true",
89         help="Make related content be sibling to HTML parts only",
90     )
91
92     def positive_integer(value):
93         try:
94             if int(value) > 0:
95                 return int(value)
96
97         except ValueError:
98             pass
99
100         raise ValueError("Must be a positive integer")
101
102     parser.add_argument(
103         "--max-number-other-attachments",
104         metavar="INTEGER",
105         type=positive_integer,
106         default=20,
107         help="Maximum number of other attachments to expect",
108     )
109
110     parser.add_argument(
111         "--only-build",
112         "--just-build",
113         action="store_true",
114         help="Only build, don't send the message",
115     )
116
117     parser.add_argument(
118         "--tempdir",
119         metavar="DIR",
120         type=pathlib.Path,
121         help="Specify temporary directory to use for attachments",
122     )
123
124     parser.add_argument(
125         "--debug-commands",
126         action="store_true",
127         help="Turn on debug logging of commands generated to stderr",
128     )
129
130     parser.add_argument(
131         "--debug-walk",
132         action="store_true",
133         help="Turn on debugging to stderr of the MIME tree walk",
134     )
135
136     subp = parser.add_subparsers(help="Sub-command parsers", dest="mode")
137     massage_p = subp.add_parser(
138         "massage", help="Massaging phase (internal use)"
139     )
140
141     massage_p.add_argument(
142         "--write-commands-to",
143         "-o",
144         metavar="FILE",
145         dest="cmdpath",
146         type=pathlib.Path,
147         required=True,
148         help="Temporary file path to write commands to",
149     )
150
151     massage_p.add_argument(
152         "MAILDRAFT",
153         nargs="?",
154         type=pathlib.Path,
155         help="If provided, the script is invoked as editor on the mail draft",
156     )
157
158     return parser.parse_args(*args, **kwargs)
159
160
161 # [ IMAGE HANDLING ] ##########################################################
162
163
164 InlineImageInfo = namedtuple(
165     "InlineImageInfo", ["cid", "desc"], defaults=[None]
166 )
167
168
169 class ImageRegistry:
170     def __init__(self):
171         self._images = OrderedDict()
172
173     def register(self, path, description=None):
174         path = os.path.expanduser(path)
175         if path.startswith("/"):
176             path = f"file://{path}"
177         cid = make_msgid()[1:-1]
178         self._images[path] = InlineImageInfo(cid, description)
179         return cid
180
181     def __iter__(self):
182         return self._images.__iter__()
183
184     def __getitem__(self, idx):
185         return self._images.__getitem__(idx)
186
187     def __len__(self):
188         return self._images.__len__()
189
190     def items(self):
191         return self._images.items()
192
193     def __repr__(self):
194         return f"<ImageRegistry(items={len(self._images)})>"
195
196     def __str__(self):
197         return self._images.__str__()
198
199
200 class InlineImageExtension(Extension):
201     class RelatedImageInlineProcessor(ImageInlineProcessor):
202         def __init__(self, re, md, registry):
203             super().__init__(re, md)
204             self._registry = registry
205
206         def handleMatch(self, m, data):
207             el, start, end = super().handleMatch(m, data)
208             if "src" in el.attrib:
209                 src = el.attrib["src"]
210                 if "://" not in src or src.startswith("file://"):
211                     # We only inline local content
212                     cid = self._registry.register(
213                         el.attrib["src"],
214                         el.attrib.get("title", el.attrib.get("alt")),
215                     )
216                     el.attrib["src"] = f"cid:{cid}"
217             return el, start, end
218
219     def __init__(self, registry):
220         super().__init__()
221         self._image_registry = registry
222
223     INLINE_PATTERN_NAME = "image_link"
224
225     def extendMarkdown(self, md):
226         md.registerExtension(self)
227         inline_image_proc = self.RelatedImageInlineProcessor(
228             IMAGE_LINK_RE, md, self._image_registry
229         )
230         md.inlinePatterns.register(
231             inline_image_proc, InlineImageExtension.INLINE_PATTERN_NAME, 150
232         )
233
234
235 def markdown_with_inline_image_support(
236     text,
237     *,
238     mdwn=None,
239     image_registry=None,
240     extensions=None,
241     extension_configs=None,
242 ):
243     registry = (
244         image_registry if image_registry is not None else ImageRegistry()
245     )
246     inline_image_handler = InlineImageExtension(registry=registry)
247     extensions = extensions or []
248     extensions.append(inline_image_handler)
249     mdwn = markdown.Markdown(
250         extensions=extensions, extension_configs=extension_configs
251     )
252
253     htmltext = mdwn.convert(text)
254
255     def replace_image_with_cid(matchobj):
256         for m in (matchobj.group(1), f"file://{matchobj.group(1)}"):
257             if m in registry:
258                 return f"(cid:{registry[m].cid}"
259         return matchobj.group(0)
260
261     text = re.sub(r"\(([^)\s]+)", replace_image_with_cid, text)
262     return text, htmltext, registry, mdwn
263
264
265 # [ CSS STYLING ] #############################################################
266
267
268 try:
269     import pynliner
270
271     _PYNLINER = True
272
273 except ImportError:
274     _PYNLINER = False
275
276 try:
277     from pygments.formatters import get_formatter_by_name
278
279     _CODEHILITE_CLASS = "codehilite"
280
281     _PYGMENTS_CSS = get_formatter_by_name(
282         "html", style="default"
283     ).get_style_defs(f".{_CODEHILITE_CLASS}")
284
285 except ImportError:
286     _PYGMENTS_CSS = None
287
288
289 def apply_styling(html, css):
290     return (
291         pynliner.Pynliner()
292         .from_string(html)
293         .with_cssString("\n".join(s for s in [_PYGMENTS_CSS, css] if s))
294         .run()
295     )
296
297
298 # [ PARTS GENERATION ] ########################################################
299
300
301 class Part(
302     namedtuple(
303         "Part",
304         ["type", "subtype", "path", "desc", "cid", "orig"],
305         defaults=[None, None, False],
306     )
307 ):
308     def __str__(self):
309         ret = f"<{self.type}/{self.subtype}>"
310         if self.cid:
311             ret = f"{ret} cid:{self.cid}"
312         if self.orig:
313             ret = f"{ret} ORIGINAL"
314         return ret
315
316
317 class Multipart(
318     namedtuple("Multipart", ["subtype", "children", "desc"], defaults=[None])
319 ):
320     def __str__(self):
321         return f"<multipart/{self.subtype}> children={len(self.children)}"
322
323     def __hash__(self):
324         return hash(str(self.subtype) + "".join(str(self.children)))
325
326
327 def filereader_fn(path, mode="r", **kwargs):
328     with open(path, mode, **kwargs) as in_f:
329         return in_f.read()
330
331
332 def filewriter_fn(path, content, mode="w", **kwargs):
333     with open(path, mode, **kwargs) as out_f:
334         out_f.write(content)
335
336
337 def collect_inline_images(
338     image_registry, *, tempdir=None, filewriter_fn=filewriter_fn
339 ):
340     relparts = []
341     for path, info in image_registry.items():
342         if path.startswith("cid:"):
343             continue
344
345         data = request.urlopen(path)
346
347         mimetype = data.headers["Content-Type"]
348         ext = mimetypes.guess_extension(mimetype)
349         tempfilename = tempfile.mkstemp(prefix="img", suffix=ext, dir=tempdir)
350         path = pathlib.Path(tempfilename[1])
351
352         filewriter_fn(path, data.read(), "w+b")
353
354         desc = (
355             f'Inline image: "{info.desc}"'
356             if info.desc
357             else f"Inline image {str(len(relparts)+1)}"
358         )
359         relparts.append(
360             Part(*mimetype.split("/"), path, cid=info.cid, desc=desc)
361         )
362
363     return relparts
364
365
366 EMAIL_SIG_SEP = "\n-- \n"
367 HTML_SIG_MARKER = "=htmlsig "
368
369
370 def make_html_doc(body, sig=None):
371     ret = (
372         "<!DOCTYPE html>\n"
373         "<html>\n"
374         "<head>\n"
375         '<meta http-equiv="content-type" content="text/html; charset=UTF-8">\n'  # noqa: E501
376         '<meta name="viewport" content="width=device-width, initial-scale=1.0">\n'  # noqa: E501
377         "</head>\n"
378         "<body>\n"
379         f"{body}\n"
380     )
381
382     if sig:
383         nl = "\n"
384         ret = (
385             f'{ret}<div id="signature"><span class="sig_separator">{EMAIL_SIG_SEP.strip(nl)}</span>\n'  # noqa: E501
386             f"{sig}\n"
387             "</div>"
388         )
389
390     return f"{ret}\n  </body>\n</html>"
391
392
393 def make_text_mail(text, sig=None):
394     return EMAIL_SIG_SEP.join((text, sig)) if sig else text
395
396
397 def extract_signature(text, *, filereader_fn=filereader_fn):
398     parts = text.split(EMAIL_SIG_SEP, 1)
399     if len(parts) == 1:
400         return text, None, None
401
402     lines = parts[1].splitlines()
403     if lines[0].startswith(HTML_SIG_MARKER):
404         path = pathlib.Path(re.split(r" +", lines.pop(0), maxsplit=1)[1])
405         textsig = "\n".join(lines)
406
407         sig_input = filereader_fn(path.expanduser())
408         soup = bs4.BeautifulSoup(sig_input, "html.parser")
409
410         style = str(soup.style.extract()) if soup.style else ""
411         for sig_selector in (
412             "#signature",
413             "#signatur",
414             "#emailsig",
415             ".signature",
416             ".signatur",
417             ".emailsig",
418             "body",
419             "div",
420         ):
421             sig = soup.select_one(sig_selector)
422             if sig:
423                 break
424
425         if not sig:
426             return parts[0], textsig, style + sig_input
427
428         if sig.attrs.get("id") == "signature":
429             sig = "".join(str(c) for c in sig.children)
430
431         return parts[0], textsig, style + str(sig)
432
433     return parts[0], parts[1], None
434
435
436 def convert_markdown_to_html(
437     origtext,
438     draftpath,
439     *,
440     related_to_html_only=False,
441     css=None,
442     filewriter_fn=filewriter_fn,
443     filereader_fn=filereader_fn,
444     tempdir=None,
445     extensions=None,
446     extension_configs=None,
447 ):
448     # TODO extension_configs need to be handled differently
449     extension_configs = extension_configs or {}
450     extension_configs.setdefault("pymdownx.highlight", {})
451     extension_configs["pymdownx.highlight"]["css_class"] = _CODEHILITE_CLASS
452
453     origtext, textsig, htmlsig = extract_signature(
454         origtext, filereader_fn=filereader_fn
455     )
456
457     (
458         origtext,
459         htmltext,
460         image_registry,
461         mdwn,
462     ) = markdown_with_inline_image_support(
463         origtext, extensions=extensions, extension_configs=extension_configs
464     )
465
466     if htmlsig:
467         if not textsig:
468             # TODO: decide what to do if there is no plain-text version
469             raise NotImplementedError("HTML signature but no text alternative")
470
471         soup = bs4.BeautifulSoup(htmlsig, "html.parser")
472         for img in soup.find_all("img"):
473             uri = img.attrs["src"]
474             desc = img.attrs.get("title", img.attrs.get("alt"))
475             cid = image_registry.register(uri, desc)
476             img.attrs["src"] = f"cid:{cid}"
477
478         htmlsig = str(soup)
479
480     elif textsig:
481         (
482             textsig,
483             htmlsig,
484             image_registry,
485             mdwn,
486         ) = markdown_with_inline_image_support(
487             textsig,
488             extensions=extensions,
489             extension_configs=extension_configs,
490             image_registry=image_registry,
491             mdwn=mdwn,
492         )
493
494     origtext = make_text_mail(origtext, textsig)
495
496     filewriter_fn(draftpath, origtext, encoding="utf-8")
497     textpart = Part(
498         "text", "plain", draftpath, "Plain-text version", orig=True
499     )
500
501     htmltext = make_html_doc(htmltext, htmlsig)
502     htmltext = apply_styling(htmltext, css)
503
504     htmlpath = draftpath.with_suffix(".html")
505     filewriter_fn(
506         htmlpath, htmltext, encoding="utf-8", errors="xmlcharrefreplace"
507     )
508     htmlpart = Part("text", "html", htmlpath, "HTML version")
509
510     imgparts = collect_inline_images(
511         image_registry, tempdir=tempdir, filewriter_fn=filewriter_fn
512     )
513
514     if related_to_html_only:
515         # If there are inline image part, they will be contained within a
516         # multipart/related part along with the HTML part only
517         if imgparts:
518             # replace htmlpart with a multipart/related container of the HTML
519             # parts and the images
520             htmlpart = Multipart(
521                 "relative", [htmlpart] + imgparts, "Group of related content"
522             )
523
524         return Multipart(
525             "alternative", [textpart, htmlpart], "Group of alternative content"
526         )
527
528     else:
529         # If there are inline image part, they will be siblings to the
530         # multipart/alternative tree within a multipart/related part
531         altpart = Multipart(
532             "alternative", [textpart, htmlpart], "Group of alternative content"
533         )
534         if imgparts:
535             return Multipart(
536                 "relative", [altpart] + imgparts, "Group of related content"
537             )
538         else:
539             return altpart
540
541
542 class MIMETreeDFWalker:
543     def __init__(self, *, visitor_fn=None, debug=False):
544         self._visitor_fn = visitor_fn or self._echovisit
545         self._debug = debug
546
547     def _echovisit(self, node, ancestry, debugprint):
548         debugprint(f"node={node} ancestry={ancestry}")
549
550     def walk(self, root, *, visitor_fn=None):
551         """
552         Recursive function to implement a depth-dirst walk of the MIME-tree
553         rooted at `root`.
554         """
555         if isinstance(root, list):
556             if len(root) > 1:
557                 root = Multipart("mixed", children=root)
558             else:
559                 root = root[0]
560
561         self._walk(
562             root,
563             ancestry=[],
564             descendents=[],
565             visitor_fn=visitor_fn or self._visitor_fn,
566         )
567
568     def _walk(self, node, *, ancestry, descendents, visitor_fn):
569         # Let's start by enumerating the parts at the current level. At the
570         # root level, ancestry will be the empty list, and we expect a
571         # multipart/* container at this level. Later, e.g. within a
572         # mutlipart/alternative container, the subtree will just be the
573         # alternative parts, while the top of the ancestry will be the
574         # multipart/alternative container, which we will process after the
575         # following loop.
576
577         lead = f"{'│ '*len(ancestry)}"
578         if isinstance(node, Multipart):
579             self.debugprint(
580                 f"{lead}├{node} ancestry={[s.subtype for s in ancestry]}"
581             )
582
583             # Depth-first, so push the current container onto the ancestry
584             # stack, then descend …
585             ancestry.append(node)
586             self.debugprint(lead + "│ " * 2)
587             for child in node.children:
588                 self._walk(
589                     child,
590                     ancestry=ancestry,
591                     descendents=descendents,
592                     visitor_fn=visitor_fn,
593                 )
594             assert ancestry.pop() == node
595             sibling_descendents = descendents
596             descendents.extend(node.children)
597
598         else:
599             self.debugprint(f"{lead}├{node}")
600             sibling_descendents = descendents
601
602         if False and ancestry:
603             self.debugprint(lead[:-1] + " │")
604
605         if visitor_fn:
606             visitor_fn(
607                 node, ancestry, sibling_descendents, debugprint=self.debugprint
608             )
609
610     def debugprint(self, s, **kwargs):
611         if self._debug:
612             print(s, file=sys.stderr, **kwargs)
613
614
615 # [ RUN MODES ] ###############################################################
616
617
618 class MuttCommands:
619     """
620     Stupid class to interface writing out Mutt commands. This is quite a hack
621     to deal with the fact that Mutt runs "push" commands in reverse order, so
622     all of a sudden, things become very complicated when mixing with "real"
623     commands.
624
625     Hence we keep two sets of commands, and one set of pushes. Commands are
626     added to the first until a push is added, after which commands are added to
627     the second set of commands.
628
629     On flush(), the first set is printed, followed by the pushes in reverse,
630     and then the second set is printed. All 3 sets are then cleared.
631     """
632
633     def __init__(self, out_f=sys.stdout, *, debug=False):
634         self._cmd1, self._push, self._cmd2 = [], [], []
635         self._out_f = out_f
636         self._debug = debug
637
638     def cmd(self, s):
639         self.debugprint(s)
640         if self._push:
641             self._cmd2.append(s)
642         else:
643             self._cmd1.append(s)
644
645     def push(self, s):
646         s = s.replace('"', r"\"")
647         s = f'push "{s}"'
648         self.debugprint(s)
649         self._push.insert(0, s)
650
651     def flush(self):
652         print(
653             "\n".join(self._cmd1 + self._push + self._cmd2), file=self._out_f
654         )
655         self._cmd1, self._push, self._cmd2 = [], [], []
656
657     def debugprint(self, s, **kwargs):
658         if self._debug:
659             print(s, file=sys.stderr, **kwargs)
660
661
662 def do_setup(
663     *,
664     out_f=sys.stdout,
665     temppath=None,
666     tempdir=None,
667     debug_commands=False,
668 ):
669     temppath = temppath or pathlib.Path(
670         tempfile.mkstemp(prefix="muttmdwn-", dir=tempdir)[1]
671     )
672     cmds = MuttCommands(out_f, debug=debug_commands)
673
674     editor = f"{' '.join(sys.argv)} massage --write-commands-to {temppath}"
675
676     cmds.cmd('set my_editor="$editor"')
677     cmds.cmd('set my_edit_headers="$edit_headers"')
678     cmds.cmd(f'set editor="{editor}"')
679     cmds.cmd("unset edit_headers")
680     cmds.cmd(f"set my_mdwn_postprocess_cmd_file={temppath}")
681     cmds.push("<first-entry><edit-file>")
682     cmds.flush()
683
684
685 def do_massage(
686     draft_f,
687     draftpath,
688     cmd_f,
689     *,
690     extensions=None,
691     css_f=None,
692     converter=convert_markdown_to_html,
693     related_to_html_only=True,
694     only_build=False,
695     max_other_attachments=20,
696     tempdir=None,
697     debug_commands=False,
698     debug_walk=False,
699 ):
700     # Here's the big picture: we're being invoked as the editor on the email
701     # draft, and whatever commands we write to the file given as cmdpath will
702     # be run by the second source command in the macro definition.
703
704     # Let's start by cleaning up what the setup did (see above), i.e. we
705     # restore the $editor and $edit_headers variables, and also unset the
706     # variable used to identify the command file we're currently writing
707     # to.
708     cmds = MuttCommands(cmd_f, debug=debug_commands)
709     cmds.cmd('set editor="$my_editor"')
710     cmds.cmd('set edit_headers="$my_edit_headers"')
711     cmds.cmd("unset my_editor")
712     cmds.cmd("unset my_edit_headers")
713
714     # let's flush those commands, as there'll be a lot of pushes from now
715     # on, which need to be run in reverse order
716     cmds.flush()
717
718     extensions = extensions.split(",") if extensions else []
719     tree = converter(
720         draft_f.read(),
721         draftpath,
722         css=css_f.read() if css_f else None,
723         related_to_html_only=related_to_html_only,
724         tempdir=tempdir,
725         extensions=extensions,
726     )
727
728     mimetree = MIMETreeDFWalker(debug=debug_walk)
729
730     state = dict(pos=1, tags={}, parts=1)
731
732     def visitor_fn(item, ancestry, descendents, *, debugprint=None):
733         """
734         Visitor function called for every node (part) of the MIME tree,
735         depth-first, and responsible for telling NeoMutt how to assemble
736         the tree.
737         """
738         KILL_LINE = r"\Ca\Ck"
739
740         if isinstance(item, Part):
741             # We've hit a leaf-node, i.e. an alternative or a related part
742             # with actual content.
743
744             # Let's add the part
745             if item.orig:
746                 # The original source already exists in the NeoMutt tree, but
747                 # the underlying file may have been modified, so we need to
748                 # update the encoding, but that's it:
749                 cmds.push("<first-entry>")
750                 cmds.push("<update-encoding>")
751
752                 # We really just need to be able to assume that at this point,
753                 # NeoMutt is at position 1, and that we've processed only this
754                 # part so far. Nevermind about actual attachments, we can
755                 # safely ignore those as they stay at the end.
756                 assert state["pos"] == 1
757                 assert state["parts"] == 1
758             else:
759                 # … whereas all other parts need to be added, and they're all
760                 # considered to be temporary and inline:
761                 cmds.push(f"<attach-file>{item.path}<enter>")
762                 cmds.push("<toggle-unlink><toggle-disposition>")
763
764                 # This added a part at the end of the list of parts, and that's
765                 # just how many parts we've seen so far, so it's position in
766                 # the NeoMutt compose list is the count of parts
767                 state["parts"] += 1
768                 state["pos"] = state["parts"]
769
770             # If the item (including the original) comes with additional
771             # information, then we might just as well update the NeoMutt
772             # tree now:
773             if item.cid:
774                 cmds.push(f"<edit-content-id>{KILL_LINE}{item.cid}<enter>")
775
776             # Now for the biggest hack in this script, which is to handle
777             # attachments, such as PDFs, that aren't related or alternatives.
778             # The problem is that when we add an inline image, it always gets
779             # appended to the list, i.e. inserted *after* other attachments.
780             # Since we don't know the number of attachments, we also cannot
781             # infer the postition of the new attachment. Therefore, we bubble
782             # it all the way to the top, only to then move it down again:
783             if state["pos"] > 1:  # skip for the first part
784                 for i in range(max_other_attachments):
785                     # could use any number here, but has to be larger than the
786                     # number of possible attachments. The performance
787                     # difference of using a high number is negligible.
788                     # Bubble up the new part
789                     cmds.push("<move-up>")
790
791                 # As we push the part to the right position in the list (i.e.
792                 # the last of the subset of attachments this script added), we
793                 # must handle the situation that subtrees are skipped by
794                 # NeoMutt. Hence, the actual number of positions to move down
795                 # is decremented by the number of descendents so far
796                 # encountered.
797                 for i in range(1, state["pos"] - len(descendents)):
798                     cmds.push("<move-down>")
799
800         elif isinstance(item, Multipart):
801             # This node has children, but we already visited them (see
802             # above). The tags dictionary of State should contain a list of
803             # their positions in the NeoMutt compose window, so iterate those
804             # and tag the parts there:
805             n_tags = len(state["tags"][item])
806             for tag in state["tags"][item]:
807                 cmds.push(f"<jump>{tag}<enter><tag-entry>")
808
809             if item.subtype == "alternative":
810                 cmds.push("<group-alternatives>")
811             elif item.subtype in ("relative", "related"):
812                 cmds.push("<group-related>")
813             elif item.subtype == "multilingual":
814                 cmds.push("<group-multilingual>")
815             else:
816                 raise NotImplementedError(
817                     f"Handling of multipart/{item.subtype} is not implemented"
818                 )
819
820             state["pos"] -= n_tags - 1
821             state["parts"] += 1
822
823         else:
824             # We should never get here
825             raise RuntimeError(f"Type {type(item)} is unexpected: {item}")
826
827         # If the item has a description, we might just as well add it
828         if item.desc:
829             cmds.push(f"<edit-description>{KILL_LINE}{item.desc}<enter>")
830
831         if ancestry:
832             # If there's an ancestry, record the current (assumed) position in
833             # the NeoMutt compose window as needed-to-tag by our direct parent
834             # (i.e. the last item of the ancestry)
835             state["tags"].setdefault(ancestry[-1], []).append(state["pos"])
836
837             lead = "│ " * (len(ancestry) + 1) + "* "
838             debugprint(
839                 f"{lead}ancestry={[a.subtype for a in ancestry]}\n"
840                 f"{lead}descendents={[d.subtype for d in descendents]}\n"
841                 f"{lead}children_positions={state['tags'][ancestry[-1]]}\n"
842                 f"{lead}pos={state['pos']}, parts={state['parts']}"
843             )
844
845     # -----------------
846     # End of visitor_fn
847
848     # Let's walk the tree and visit every node with our fancy visitor
849     # function
850     mimetree.walk(tree, visitor_fn=visitor_fn)
851
852     if not only_build:
853         cmds.push("<send-message>")
854
855     # Finally, cleanup. Since we're responsible for removing the temporary
856     # file, how's this for a little hack?
857     try:
858         filename = cmd_f.name
859     except AttributeError:
860         filename = "pytest_internal_file"
861     cmds.cmd(f"source 'rm -f {filename}|'")
862     cmds.cmd("unset my_mdwn_postprocess_cmd_file")
863     cmds.flush()
864
865
866 # [ CLI ENTRY ] ###############################################################
867
868 if __name__ == "__main__":
869     args = parse_cli_args()
870
871     if args.mode is None:
872         do_setup(
873             tempdir=args.tempdir,
874             debug_commands=args.debug_commands,
875         )
876
877     elif args.mode == "massage":
878         with open(args.MAILDRAFT, "r") as draft_f, open(
879             args.cmdpath, "w"
880         ) as cmd_f, open(args.css_file, "r") as css_f:
881             do_massage(
882                 draft_f,
883                 args.MAILDRAFT,
884                 cmd_f,
885                 extensions=args.extensions,
886                 css_f=css_f,
887                 related_to_html_only=args.related_to_html_only,
888                 max_other_attachments=args.max_number_other_attachments,
889                 only_build=args.only_build,
890                 tempdir=args.tempdir,
891                 debug_commands=args.debug_commands,
892                 debug_walk=args.debug_walk,
893             )
894
895
896 # [ TESTS ] ###################################################################
897
898 try:
899     import pytest
900     from io import StringIO
901
902     class Tests:
903         @pytest.fixture
904         def const1(self):
905             return "CONSTANT STRING 1"
906
907         @pytest.fixture
908         def const2(self):
909             return "CONSTANT STRING 2"
910
911         # NOTE: tests using the capsys fixture must specify sys.stdout to the
912         # functions they call, else old stdout is used and not captured
913
914         @pytest.mark.muttctrl
915         def test_MuttCommands_cmd(self, const1, const2, capsys):
916             "Assert order of commands"
917             cmds = MuttCommands(out_f=sys.stdout)
918             cmds.cmd(const1)
919             cmds.cmd(const2)
920             cmds.flush()
921             captured = capsys.readouterr()
922             assert captured.out == "\n".join((const1, const2, ""))
923
924         @pytest.mark.muttctrl
925         def test_MuttCommands_push(self, const1, const2, capsys):
926             "Assert reverse order of pushes"
927             cmds = MuttCommands(out_f=sys.stdout)
928             cmds.push(const1)
929             cmds.push(const2)
930             cmds.flush()
931             captured = capsys.readouterr()
932             assert (
933                 captured.out
934                 == ('"\npush "'.join(("", const2, const1, "")))[2:-6]
935             )
936
937         @pytest.mark.muttctrl
938         def test_MuttCommands_push_escape(self, const1, const2, capsys):
939             cmds = MuttCommands(out_f=sys.stdout)
940             cmds.push(f'"{const1}"')
941             cmds.flush()
942             captured = capsys.readouterr()
943             assert f'"\\"{const1}\\""' in captured.out
944
945         @pytest.mark.muttctrl
946         def test_MuttCommands_cmd_push_mixed(self, const1, const2, capsys):
947             "Assert reverse order of pushes"
948             cmds = MuttCommands(out_f=sys.stdout)
949             lines = ["000", "001", "010", "011", "100", "101", "110", "111"]
950             for i in range(2):
951                 cmds.cmd(lines[4 * i + 0])
952                 cmds.cmd(lines[4 * i + 1])
953                 cmds.push(lines[4 * i + 2])
954                 cmds.push(lines[4 * i + 3])
955             cmds.flush()
956
957             captured = capsys.readouterr()
958             lines_out = captured.out.splitlines()
959             assert lines[0] in lines_out[0]
960             assert lines[1] in lines_out[1]
961             assert lines[7] in lines_out[2]
962             assert lines[6] in lines_out[3]
963             assert lines[3] in lines_out[4]
964             assert lines[2] in lines_out[5]
965             assert lines[4] in lines_out[6]
966             assert lines[5] in lines_out[7]
967
968         @pytest.fixture
969         def mime_tree_related_to_alternative(self):
970             return Multipart(
971                 "relative",
972                 children=[
973                     Multipart(
974                         "alternative",
975                         children=[
976                             Part(
977                                 "text",
978                                 "plain",
979                                 "part.txt",
980                                 desc="Plain",
981                                 orig=True,
982                             ),
983                             Part("text", "html", "part.html", desc="HTML"),
984                         ],
985                         desc="Alternative",
986                     ),
987                     Part(
988                         "text", "png", "logo.png", cid="logo.png", desc="Logo"
989                     ),
990                 ],
991                 desc="Related",
992             )
993
994         @pytest.fixture
995         def mime_tree_related_to_html(self):
996             return Multipart(
997                 "alternative",
998                 children=[
999                     Part(
1000                         "text",
1001                         "plain",
1002                         "part.txt",
1003                         desc="Plain",
1004                         orig=True,
1005                     ),
1006                     Multipart(
1007                         "relative",
1008                         children=[
1009                             Part("text", "html", "part.html", desc="HTML"),
1010                             Part(
1011                                 "text",
1012                                 "png",
1013                                 "logo.png",
1014                                 cid="logo.png",
1015                                 desc="Logo",
1016                             ),
1017                         ],
1018                         desc="Related",
1019                     ),
1020                 ],
1021                 desc="Alternative",
1022             )
1023
1024         @pytest.mark.treewalk
1025         def test_MIMETreeDFWalker_depth_first_walk(
1026             self, mime_tree_related_to_alternative
1027         ):
1028             mimetree = MIMETreeDFWalker()
1029
1030             items = []
1031
1032             def visitor_fn(item, ancestry, descendents, debugprint):
1033                 items.append((item, len(ancestry), len(descendents)))
1034
1035             mimetree.walk(
1036                 mime_tree_related_to_alternative, visitor_fn=visitor_fn
1037             )
1038             assert len(items) == 5
1039             assert items[0][0].subtype == "plain"
1040             assert items[0][1] == 2
1041             assert items[0][2] == 0
1042             assert items[1][0].subtype == "html"
1043             assert items[1][1] == 2
1044             assert items[1][2] == 0
1045             assert items[2][0].subtype == "alternative"
1046             assert items[2][1] == 1
1047             assert items[2][2] == 2
1048             assert items[3][0].subtype == "png"
1049             assert items[3][1] == 1
1050             assert items[3][2] == 2
1051             assert items[4][0].subtype == "relative"
1052             assert items[4][1] == 0
1053             assert items[4][2] == 4
1054
1055         @pytest.mark.treewalk
1056         def test_MIMETreeDFWalker_list_to_mixed(self, const1):
1057             mimetree = MIMETreeDFWalker()
1058             items = []
1059
1060             def visitor_fn(item, ancestry, descendents, debugprint):
1061                 items.append(item)
1062
1063             p = Part("text", "plain", const1)
1064             mimetree.walk([p], visitor_fn=visitor_fn)
1065             assert items[-1].subtype == "plain"
1066             mimetree.walk([p, p], visitor_fn=visitor_fn)
1067             assert items[-1].subtype == "mixed"
1068
1069         @pytest.mark.treewalk
1070         def test_MIMETreeDFWalker_visitor_in_constructor(
1071             self, mime_tree_related_to_alternative
1072         ):
1073             items = []
1074
1075             def visitor_fn(item, ancestry, descendents, debugprint):
1076                 items.append(item)
1077
1078             mimetree = MIMETreeDFWalker(visitor_fn=visitor_fn)
1079             mimetree.walk(mime_tree_related_to_alternative)
1080             assert len(items) == 5
1081
1082         @pytest.fixture
1083         def string_io(self, const1, text=None):
1084             return StringIO(text or const1)
1085
1086         @pytest.mark.massage
1087         def test_do_massage_basic(self, const1, string_io, capsys):
1088             def converter(
1089                 drafttext,
1090                 draftpath,
1091                 css,
1092                 related_to_html_only,
1093                 extensions,
1094                 tempdir,
1095             ):
1096                 return Part("text", "plain", draftpath, orig=True)
1097
1098             do_massage(
1099                 draft_f=string_io,
1100                 draftpath=const1,
1101                 cmd_f=sys.stdout,
1102                 converter=converter,
1103             )
1104
1105             captured = capsys.readouterr()
1106             lines = captured.out.splitlines()
1107             assert '="$my_editor"' in lines.pop(0)
1108             assert '="$my_edit_headers"' in lines.pop(0)
1109             assert "unset my_editor" == lines.pop(0)
1110             assert "unset my_edit_headers" == lines.pop(0)
1111             assert "send-message" in lines.pop(0)
1112             assert "update-encoding" in lines.pop(0)
1113             assert "first-entry" in lines.pop(0)
1114             assert "source 'rm -f " in lines.pop(0)
1115             assert "unset my_mdwn_postprocess_cmd_file" == lines.pop(0)
1116
1117         @pytest.mark.massage
1118         def test_do_massage_fulltree(
1119             self, string_io, const1, mime_tree_related_to_alternative, capsys
1120         ):
1121             def converter(
1122                 drafttext,
1123                 draftpath,
1124                 css,
1125                 related_to_html_only,
1126                 extensions,
1127                 tempdir,
1128             ):
1129                 return mime_tree_related_to_alternative
1130
1131             max_attachments = 5
1132             do_massage(
1133                 draft_f=string_io,
1134                 draftpath=const1,
1135                 cmd_f=sys.stdout,
1136                 max_other_attachments=max_attachments,
1137                 converter=converter,
1138             )
1139
1140             captured = capsys.readouterr()
1141             lines = captured.out.splitlines()[4:-2]
1142             assert "first-entry" in lines.pop()
1143             assert "update-encoding" in lines.pop()
1144             assert "Plain" in lines.pop()
1145             assert "part.html" in lines.pop()
1146             assert "toggle-unlink" in lines.pop()
1147             for i in range(max_attachments):
1148                 assert "move-up" in lines.pop()
1149             assert "move-down" in lines.pop()
1150             assert "HTML" in lines.pop()
1151             assert "jump>1" in lines.pop()
1152             assert "jump>2" in lines.pop()
1153             assert "group-alternatives" in lines.pop()
1154             assert "Alternative" in lines.pop()
1155             assert "logo.png" in lines.pop()
1156             assert "toggle-unlink" in lines.pop()
1157             assert "content-id" in lines.pop()
1158             for i in range(max_attachments):
1159                 assert "move-up" in lines.pop()
1160             assert "move-down" in lines.pop()
1161             assert "Logo" in lines.pop()
1162             assert "jump>1" in lines.pop()
1163             assert "jump>4" in lines.pop()
1164             assert "group-related" in lines.pop()
1165             assert "Related" in lines.pop()
1166             assert "send-message" in lines.pop()
1167             assert len(lines) == 0
1168
1169         @pytest.fixture
1170         def fake_filewriter(self):
1171             class FileWriter:
1172                 def __init__(self):
1173                     self._writes = []
1174
1175                 def __call__(self, path, content, mode="w", **kwargs):
1176                     self._writes.append((path, content))
1177
1178                 def pop(self, index=-1):
1179                     return self._writes.pop(index)
1180
1181             return FileWriter()
1182
1183         @pytest.fixture
1184         def markdown_non_converter(self, const1, const2):
1185             return lambda s, text: f"{const1}{text}{const2}"
1186
1187         @pytest.mark.converter
1188         def test_converter_tree_basic(self, const1, const2, fake_filewriter):
1189             path = pathlib.Path(const2)
1190             tree = convert_markdown_to_html(
1191                 const1, path, filewriter_fn=fake_filewriter
1192             )
1193
1194             assert tree.subtype == "alternative"
1195             assert len(tree.children) == 2
1196             assert tree.children[0].subtype == "plain"
1197             assert tree.children[0].path == path
1198             assert tree.children[0].orig
1199             assert tree.children[1].subtype == "html"
1200             assert tree.children[1].path == path.with_suffix(".html")
1201
1202         def test_converter_writes(
1203             self,
1204             const1,
1205             const2,
1206             fake_filewriter,
1207             monkeypatch,
1208             markdown_non_converter,
1209         ):
1210             path = pathlib.Path(const2)
1211
1212             with monkeypatch.context() as m:
1213                 m.setattr(markdown.Markdown, "convert", markdown_non_converter)
1214                 convert_markdown_to_html(
1215                     const1, path, filewriter_fn=fake_filewriter
1216                 )
1217
1218             assert (path, const1) == fake_filewriter.pop(0)
1219             written = fake_filewriter.pop(0)
1220             assert path.with_suffix(".html") == written[0]
1221             assert const1 in written[1]
1222
1223         @pytest.mark.imgproc
1224         def test_markdown_inline_image_processor(self):
1225             imgpath1 = "file:/path/to/image.png"
1226             imgpath2 = "file:///path/to/image.png?url=params"
1227             imgpath3 = "/path/to/image.png"
1228             text = f"""![inline local image]({imgpath1})
1229                        ![image inlined
1230                          with newline]({imgpath2})
1231                        ![image local path]({imgpath3})"""
1232             text, html, images, mdwn = markdown_with_inline_image_support(text)
1233
1234             # local paths have been normalised to URLs:
1235             imgpath3 = f"file://{imgpath3}"
1236
1237             assert 'src="cid:' in html
1238             assert "](cid:" in text
1239             assert len(images) == 3
1240             assert imgpath1 in images
1241             assert imgpath2 in images
1242             assert imgpath3 in images
1243             assert images[imgpath1].cid != images[imgpath2].cid
1244             assert images[imgpath1].cid != images[imgpath3].cid
1245             assert images[imgpath2].cid != images[imgpath3].cid
1246
1247         @pytest.mark.imgproc
1248         def test_markdown_inline_image_processor_title_to_desc(self, const1):
1249             imgpath = "file:///path/to/image.png"
1250             text = f'![inline local image]({imgpath} "{const1}")'
1251             text, html, images, mdwn = markdown_with_inline_image_support(text)
1252             assert images[imgpath].desc == const1
1253
1254         @pytest.mark.imgproc
1255         def test_markdown_inline_image_processor_alt_to_desc(self, const1):
1256             imgpath = "file:///path/to/image.png"
1257             text = f"![{const1}]({imgpath})"
1258             text, html, images, mdwn = markdown_with_inline_image_support(text)
1259             assert images[imgpath].desc == const1
1260
1261         @pytest.mark.imgproc
1262         def test_markdown_inline_image_processor_title_over_alt_desc(
1263             self, const1, const2
1264         ):
1265             imgpath = "file:///path/to/image.png"
1266             text = f'![{const1}]({imgpath} "{const2}")'
1267             text, html, images, mdwn = markdown_with_inline_image_support(text)
1268             assert images[imgpath].desc == const2
1269
1270         @pytest.mark.imgproc
1271         def test_markdown_inline_image_not_external(self):
1272             imgpath = "https://path/to/image.png"
1273             text = f"![inline image]({imgpath})"
1274             text, html, images, mdwn = markdown_with_inline_image_support(text)
1275
1276             assert 'src="cid:' not in html
1277             assert "](cid:" not in text
1278             assert len(images) == 0
1279
1280         @pytest.mark.imgproc
1281         def test_markdown_inline_image_local_file(self):
1282             imgpath = "/path/to/image.png"
1283             text = f"![inline image]({imgpath})"
1284             text, html, images, mdwn = markdown_with_inline_image_support(text)
1285
1286             for k, v in images.items():
1287                 assert k == f"file://{imgpath}"
1288                 break
1289
1290         @pytest.mark.imgproc
1291         def test_markdown_inline_image_expanduser(self):
1292             imgpath = pathlib.Path("~/image.png")
1293             text = f"![inline image]({imgpath})"
1294             text, html, images, mdwn = markdown_with_inline_image_support(text)
1295
1296             for k, v in images.items():
1297                 assert k == f"file://{imgpath.expanduser()}"
1298                 break
1299
1300         @pytest.fixture
1301         def test_png(self):
1302             return (
1303                 "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAE"
1304                 "AAAABCAAAAAA6fptVAAAACklEQVQI12P4DwABAQEAG7buVgAA"
1305             )
1306
1307         @pytest.mark.imgproc
1308         def test_markdown_inline_image_processor_base64(self, test_png):
1309             text = f"![1px white inlined]({test_png})"
1310             text, html, images, mdwn = markdown_with_inline_image_support(text)
1311
1312             assert 'src="cid:' in html
1313             assert "](cid:" in text
1314             assert len(images) == 1
1315             assert test_png in images
1316
1317         @pytest.mark.converter
1318         def test_converter_tree_inline_image_base64(
1319             self, test_png, const1, fake_filewriter
1320         ):
1321             text = f"![inline base64 image]({test_png})"
1322             path = pathlib.Path(const1)
1323             tree = convert_markdown_to_html(
1324                 text,
1325                 path,
1326                 filewriter_fn=fake_filewriter,
1327                 related_to_html_only=False,
1328             )
1329             assert tree.subtype == "relative"
1330             assert tree.children[0].subtype == "alternative"
1331             assert tree.children[1].subtype == "png"
1332             written = fake_filewriter.pop()
1333             assert tree.children[1].path == written[0]
1334             assert written[1] == request.urlopen(test_png).read()
1335
1336         @pytest.mark.converter
1337         def test_converter_tree_inline_image_base64_related_to_html(
1338             self, test_png, const1, fake_filewriter
1339         ):
1340             text = f"![inline base64 image]({test_png})"
1341             path = pathlib.Path(const1)
1342             tree = convert_markdown_to_html(
1343                 text,
1344                 path,
1345                 filewriter_fn=fake_filewriter,
1346                 related_to_html_only=True,
1347             )
1348             assert tree.subtype == "alternative"
1349             assert tree.children[1].subtype == "relative"
1350             assert tree.children[1].children[1].subtype == "png"
1351             written = fake_filewriter.pop()
1352             assert tree.children[1].children[1].path == written[0]
1353             assert written[1] == request.urlopen(test_png).read()
1354
1355         @pytest.mark.converter
1356         def test_converter_tree_inline_image_cid(
1357             self, const1, fake_filewriter
1358         ):
1359             text = f"![inline base64 image](cid:{const1})"
1360             path = pathlib.Path(const1)
1361             tree = convert_markdown_to_html(
1362                 text,
1363                 path,
1364                 filewriter_fn=fake_filewriter,
1365                 related_to_html_only=False,
1366             )
1367             assert len(tree.children) == 2
1368             assert tree.children[0].cid != const1
1369             assert tree.children[0].type != "image"
1370             assert tree.children[1].cid != const1
1371             assert tree.children[1].type != "image"
1372
1373         @pytest.mark.imgcoll
1374         def test_inline_image_collection(
1375             self, test_png, const1, const2, fake_filewriter
1376         ):
1377             test_images = {test_png: InlineImageInfo(cid=const1, desc=const2)}
1378             relparts = collect_inline_images(
1379                 test_images, filewriter_fn=fake_filewriter
1380             )
1381
1382             written = fake_filewriter.pop()
1383             assert b"PNG" in written[1]
1384
1385             assert relparts[0].subtype == "png"
1386             assert relparts[0].path == written[0]
1387             assert relparts[0].cid == const1
1388             assert relparts[0].desc.endswith(const2)
1389
1390         if _PYNLINER:
1391
1392             @pytest.mark.styling
1393             def test_apply_stylesheet(self):
1394                 html = "<p>Hello, world!</p>"
1395                 css = "p { color:red }"
1396                 out = apply_styling(html, css)
1397                 assert 'p style="color' in out
1398
1399             @pytest.mark.styling
1400             def test_massage_styling_to_converter(self, string_io, const1):
1401                 css = "p { color:red }"
1402                 css_f = StringIO(css)
1403                 out_f = StringIO()
1404                 css_applied = []
1405
1406                 def converter(
1407                     drafttext,
1408                     draftpath,
1409                     css,
1410                     related_to_html_only,
1411                     extensions,
1412                     tempdir,
1413                 ):
1414                     css_applied.append(css)
1415                     return Part("text", "plain", draftpath, orig=True)
1416
1417                 do_massage(
1418                     draft_f=string_io,
1419                     draftpath=const1,
1420                     cmd_f=out_f,
1421                     css_f=css_f,
1422                     converter=converter,
1423                 )
1424                 assert css_applied[0] == css
1425
1426             @pytest.mark.converter
1427             def test_converter_apply_styles(
1428                 self, const1, fake_filewriter, monkeypatch
1429             ):
1430                 path = pathlib.Path(const1)
1431                 text = "Hello, world!"
1432                 css = "p { color:red }"
1433                 with monkeypatch.context() as m:
1434                     m.setattr(
1435                         markdown.Markdown,
1436                         "convert",
1437                         lambda s, t: f"<p>{t}</p>",
1438                     )
1439                     convert_markdown_to_html(
1440                         text, path, css=css, filewriter_fn=fake_filewriter
1441                     )
1442                 assert "color: red" in fake_filewriter.pop()[1]
1443
1444         if _PYGMENTS_CSS:
1445
1446             @pytest.mark.styling
1447             def test_apply_stylesheet_pygments(self):
1448                 html = (
1449                     f'<div class="{_CODEHILITE_CLASS}">'
1450                     "<pre>def foo():\n    return</pre></div>"
1451                 )
1452                 out = apply_styling(html, _PYGMENTS_CSS)
1453                 assert f'{_CODEHILITE_CLASS}" style="' in out
1454
1455         @pytest.mark.massage
1456         def test_mime_tree_relative_within_alternative(
1457             self, string_io, const1, capsys, mime_tree_related_to_html
1458         ):
1459             def converter(
1460                 drafttext,
1461                 draftpath,
1462                 css,
1463                 related_to_html_only,
1464                 extensions,
1465                 tempdir,
1466             ):
1467                 return mime_tree_related_to_html
1468
1469             do_massage(
1470                 draft_f=string_io,
1471                 draftpath=const1,
1472                 cmd_f=sys.stdout,
1473                 converter=converter,
1474             )
1475
1476             captured = capsys.readouterr()
1477             lines = captured.out.splitlines()[4:-2]
1478             assert "first-entry" in lines.pop()
1479             assert "update-encoding" in lines.pop()
1480             assert "Plain" in lines.pop()
1481             assert "part.html" in lines.pop()
1482             assert "toggle-unlink" in lines.pop()
1483             assert "move-up" in lines.pop()
1484             while True:
1485                 top = lines.pop()
1486                 if "move-up" not in top:
1487                     break
1488             assert "move-down" in top
1489             assert "HTML" in lines.pop()
1490             assert "logo.png" in lines.pop()
1491             assert "toggle-unlink" in lines.pop()
1492             assert "content-id" in lines.pop()
1493             assert "move-up" in lines.pop()
1494             while True:
1495                 top = lines.pop()
1496                 if "move-up" not in top:
1497                     break
1498             assert "move-down" in top
1499             assert "move-down" in lines.pop()
1500             assert "Logo" in lines.pop()
1501             assert "jump>2" in lines.pop()
1502             assert "jump>3" in lines.pop()
1503             assert "group-related" in lines.pop()
1504             assert "Related" in lines.pop()
1505             assert "jump>1" in lines.pop()
1506             assert "jump>2" in lines.pop()
1507             assert "group-alternative" in lines.pop()
1508             assert "Alternative" in lines.pop()
1509             assert "send-message" in lines.pop()
1510             assert len(lines) == 0
1511
1512         @pytest.mark.massage
1513         def test_mime_tree_nested_trees_does_not_break_positioning(
1514             self, string_io, const1, capsys
1515         ):
1516             def converter(
1517                 drafttext,
1518                 draftpath,
1519                 css,
1520                 related_to_html_only,
1521                 extensions,
1522                 tempdir,
1523             ):
1524                 return Multipart(
1525                     "relative",
1526                     children=[
1527                         Multipart(
1528                             "alternative",
1529                             children=[
1530                                 Part(
1531                                     "text",
1532                                     "plain",
1533                                     "part.txt",
1534                                     desc="Plain",
1535                                     orig=True,
1536                                 ),
1537                                 Multipart(
1538                                     "alternative",
1539                                     children=[
1540                                         Part(
1541                                             "text",
1542                                             "plain",
1543                                             "part.txt",
1544                                             desc="Nested plain",
1545                                         ),
1546                                         Part(
1547                                             "text",
1548                                             "html",
1549                                             "part.html",
1550                                             desc="Nested HTML",
1551                                         ),
1552                                     ],
1553                                     desc="Nested alternative",
1554                                 ),
1555                             ],
1556                             desc="Alternative",
1557                         ),
1558                         Part(
1559                             "text",
1560                             "png",
1561                             "logo.png",
1562                             cid="logo.png",
1563                             desc="Logo",
1564                         ),
1565                     ],
1566                     desc="Related",
1567                 )
1568
1569             do_massage(
1570                 draft_f=string_io,
1571                 draftpath=const1,
1572                 cmd_f=sys.stdout,
1573                 converter=converter,
1574             )
1575
1576             captured = capsys.readouterr()
1577             lines = captured.out.splitlines()
1578             while "logo.png" not in lines.pop():
1579                 pass
1580             lines.pop()
1581             assert "content-id" in lines.pop()
1582             assert "move-up" in lines.pop()
1583             while True:
1584                 top = lines.pop()
1585                 if "move-up" not in top:
1586                     break
1587             assert "move-down" in top
1588             # Due to the nested trees, the number of descendents of the sibling
1589             # actually needs to be considered, not just the nieces. So to move
1590             # from position 1 to position 6, it only needs one <move-down>
1591             # because that jumps over the entire sibling tree. Thus what
1592             # follows next must not be another <move-down>
1593             assert "Logo" in lines.pop()
1594
1595         @pytest.mark.sig
1596         def test_signature_extraction_no_signature(self, const1):
1597             assert (const1, None, None) == extract_signature(const1)
1598
1599         @pytest.mark.sig
1600         def test_signature_extraction_just_text(self, const1, const2):
1601             origtext, textsig, htmlsig = extract_signature(
1602                 f"{const1}{EMAIL_SIG_SEP}{const2}"
1603             )
1604             assert origtext == const1
1605             assert textsig == const2
1606             assert htmlsig is None
1607
1608         @pytest.mark.sig
1609         def test_signature_extraction_html(self, const1, const2):
1610             path = pathlib.Path("somepath")
1611             sigconst = "HTML signature from {path} but as a string"
1612
1613             def filereader_fn(path):
1614                 return (
1615                     f'<div id="signature">{sigconst.format(path=path)}</div>'
1616                 )
1617
1618             origtext, textsig, htmlsig = extract_signature(
1619                 f"{const1}{EMAIL_SIG_SEP}{HTML_SIG_MARKER} {path}\n{const2}",
1620                 filereader_fn=filereader_fn,
1621             )
1622             assert origtext == const1
1623             assert textsig == const2
1624             assert htmlsig == sigconst.format(path=path)
1625
1626         @pytest.mark.sig
1627         def test_signature_extraction_file_not_found(self, const1):
1628             path = pathlib.Path("/does/not/exist")
1629             with pytest.raises(FileNotFoundError):
1630                 origtext, textsig, htmlsig = extract_signature(
1631                     f"{const1}{EMAIL_SIG_SEP}{HTML_SIG_MARKER}{path}\n{const1}"
1632                 )
1633
1634         @pytest.mark.imgproc
1635         def test_image_registry(self, const1):
1636             reg = ImageRegistry()
1637             cid = reg.register(const1)
1638             assert "@" in cid
1639             assert not cid.startswith("<")
1640             assert not cid.endswith(">")
1641             assert const1 in reg
1642
1643         @pytest.mark.imgproc
1644         def test_image_registry_file_uri(self, const1):
1645             reg = ImageRegistry()
1646             reg.register("/some/path")
1647             for path in reg:
1648                 assert path.startswith("file://")
1649                 break
1650
1651         @pytest.mark.converter
1652         @pytest.mark.sig
1653         def test_converter_signature_handling(
1654             self, const1, fake_filewriter, monkeypatch
1655         ):
1656             path = pathlib.Path(const1)
1657
1658             mailparts = (
1659                 "This is the mail body\n",
1660                 f"{EMAIL_SIG_SEP}",
1661                 "This is a plain-text signature only",
1662             )
1663
1664             def filereader_fn(path):
1665                 return ""
1666
1667             with monkeypatch.context() as m:
1668                 m.setattr(markdown.Markdown, "convert", lambda s, t: t)
1669                 convert_markdown_to_html(
1670                     "".join(mailparts),
1671                     path,
1672                     filewriter_fn=fake_filewriter,
1673                     filereader_fn=filereader_fn,
1674                 )
1675
1676             soup = bs4.BeautifulSoup(fake_filewriter.pop()[1], "html.parser")
1677             body = soup.body.contents
1678
1679             assert mailparts[0] in body.pop(0)
1680
1681             sig = soup.select_one("#signature")
1682             assert sig == body.pop(0)
1683
1684             sep = sig.select_one("span.sig_separator")
1685             assert sep == sig.contents[0]
1686             assert f"\n{sep.text}\n" == EMAIL_SIG_SEP
1687
1688             assert mailparts[2] in sig.contents[1]
1689
1690         @pytest.mark.converter
1691         @pytest.mark.sig
1692         def test_converter_signature_handling_htmlsig(
1693             self, const1, fake_filewriter, monkeypatch
1694         ):
1695             path = pathlib.Path(const1)
1696
1697             mailparts = (
1698                 "This is the mail body",
1699                 f"{EMAIL_SIG_SEP}",
1700                 f"{HTML_SIG_MARKER}{path}\n",
1701                 "This is the plain-text version",
1702             )
1703
1704             htmlsig = "HTML Signature from {path}"
1705
1706             def filereader_fn(path):
1707                 return f'<div id="signature">{htmlsig.format(path=path)}</div>'
1708
1709             def mdwn_fn(t):
1710                 return t.upper()
1711
1712             with monkeypatch.context() as m:
1713                 m.setattr(
1714                     markdown.Markdown, "convert", lambda s, t: mdwn_fn(t)
1715                 )
1716                 convert_markdown_to_html(
1717                     "".join(mailparts),
1718                     path,
1719                     filewriter_fn=fake_filewriter,
1720                     filereader_fn=filereader_fn,
1721                 )
1722
1723             soup = bs4.BeautifulSoup(fake_filewriter.pop()[1], "html.parser")
1724             sig = soup.select_one("#signature")
1725             sig.span.extract()
1726
1727             assert HTML_SIG_MARKER not in sig.text
1728             assert htmlsig.format(path=path) == sig.text.strip()
1729
1730             plaintext = fake_filewriter.pop()[1]
1731             assert plaintext.endswith(EMAIL_SIG_SEP + mailparts[-1])
1732
1733         @pytest.mark.converter
1734         @pytest.mark.sig
1735         def test_converter_signature_handling_htmlsig_with_image(
1736             self, const1, fake_filewriter, monkeypatch, test_png
1737         ):
1738             path = pathlib.Path(const1)
1739
1740             mailparts = (
1741                 "This is the mail body",
1742                 f"{EMAIL_SIG_SEP}",
1743                 f"{HTML_SIG_MARKER}{path}\n",
1744                 "This is the plain-text version",
1745             )
1746
1747             htmlsig = (
1748                 "HTML Signature from {path} with image\n"
1749                 f'<img src="{test_png}">\n'
1750             )
1751
1752             def filereader_fn(path):
1753                 return f'<div id="signature">{htmlsig.format(path=path)}</div>'
1754
1755             def mdwn_fn(t):
1756                 return t.upper()
1757
1758             with monkeypatch.context() as m:
1759                 m.setattr(
1760                     markdown.Markdown, "convert", lambda s, t: mdwn_fn(t)
1761                 )
1762                 convert_markdown_to_html(
1763                     "".join(mailparts),
1764                     path,
1765                     filewriter_fn=fake_filewriter,
1766                     filereader_fn=filereader_fn,
1767                 )
1768
1769             assert fake_filewriter.pop()[0].suffix == ".png"
1770
1771             soup = bs4.BeautifulSoup(fake_filewriter.pop()[1], "html.parser")
1772             assert soup.img.attrs["src"].startswith("cid:")
1773
1774         @pytest.mark.converter
1775         @pytest.mark.sig
1776         def test_converter_signature_handling_textsig_with_image(
1777             self, const1, fake_filewriter, test_png
1778         ):
1779             mailparts = (
1780                 "This is the mail body",
1781                 f"{EMAIL_SIG_SEP}",
1782                 "This is the plain-text version with image\n",
1783                 f"![Inline]({test_png})",
1784
1785             )
1786             tree = convert_markdown_to_html
1787                 "".join(mailparts),
1788                 pathlib.Path(const1),
1789                 filewriter_fn=fake_filewriter,
1790             )
1791
1792             assert tree.subtype == "relative"
1793             assert tree.children[0].subtype == "alternative"
1794             assert tree.children[1].subtype == "png"
1795             written = fake_filewriter.pop()
1796             assert tree.children[1].path == written[0]
1797             assert written[1] == request.urlopen(test_png).read()
1798
1799         def test_converter_attribution_to_admonition(self, fake_filewriter):
1800
1801
1802 except ImportError:
1803     pass