X-Git-Url: https://git.madduck.net/etc/neomutt.git/blobdiff_plain/a2978b6ed0902fda8c5ab38029428ae2637ce4a6..af5fa262ab9c9bcb6a398a7557cd9732597df355:/.config/neomutt/buildmimetree.py?ds=sidebyside diff --git a/.config/neomutt/buildmimetree.py b/.config/neomutt/buildmimetree.py index f7d7c56..e41ca61 100755 --- a/.config/neomutt/buildmimetree.py +++ b/.config/neomutt/buildmimetree.py @@ -23,8 +23,8 @@ # - python3-markdown # Optional: # - pytest -# - Pynliner -# - Pygments, if installed, then syntax highlighting is enabled +# - Pynliner, provides --css-file and thus inline styling of HTML output +# - Pygments, then syntax highlighting for fenced code is enabled # # Latest version: # https://git.madduck.net/etc/neomutt.git/blob_plain/HEAD:/.config/neomutt/buildmimetree.py @@ -63,9 +63,19 @@ def parse_cli_args(*args, **kwargs): "--extensions", type=str, default="", - help="Markdown extension to use (comma-separated list)" + help="Markdown extension to use (comma-separated list)", ) + if _PYNLINER: + parser.add_argument( + "--css-file", + type=str, + default="", + help="CSS file to merge with the final HTML", + ) + else: + parser.set_defaults(css_file=None) + parser.add_argument( "--only-build", action="store_true", @@ -85,12 +95,16 @@ def parse_cli_args(*args, **kwargs): ) subp = parser.add_subparsers(help="Sub-command parsers", dest="mode") - massage_p = subp.add_parser("massage", help="Massaging phase (internal use)") + massage_p = subp.add_parser( + "massage", help="Massaging phase (internal use)" + ) massage_p.add_argument( "--write-commands-to", + "-o", metavar="PATH", dest="cmdpath", + required=True, help="Temporary file path to write commands to", ) @@ -158,11 +172,15 @@ class InlineImageExtension(Extension): return self._images -def markdown_with_inline_image_support(text, *, extensions=None): +def markdown_with_inline_image_support( + text, *, extensions=None, extension_configs=None +): inline_image_handler = InlineImageExtension() extensions = extensions or [] extensions.append(inline_image_handler) - mdwn = markdown.Markdown(extensions=extensions) + mdwn = markdown.Markdown( + extensions=extensions, extension_configs=extension_configs + ) htmltext = mdwn.convert(text) images = inline_image_handler.get_images() @@ -177,6 +195,38 @@ def markdown_with_inline_image_support(text, *, extensions=None): return text, htmltext, images +# [ CSS STYLING ] ############################################################# + +try: + import pynliner + + _PYNLINER = True + +except ImportError: + _PYNLINER = False + +try: + from pygments.formatters import get_formatter_by_name + + _CODEHILITE_CLASS = "codehilite" + + _PYGMENTS_CSS = get_formatter_by_name( + "html", style="default" + ).get_style_defs(f".{_CODEHILITE_CLASS}") + +except ImportError: + _PYGMENTS_CSS = None + + +def apply_styling(html, css): + return ( + pynliner.Pynliner() + .from_string(html) + .with_cssString("\n".join(s for s in [_PYGMENTS_CSS, css] if s)) + .run() + ) + + # [ PARTS GENERATION ] ######################################################## @@ -202,6 +252,9 @@ class Multipart( def __str__(self): return f" children={len(self.children)}" + def __hash__(self): + return hash(str(self.subtype) + "".join(str(self.children))) + def filewriter_fn(path, content, mode="w", **kwargs): with open(path, mode, **kwargs) as out_f: @@ -213,6 +266,9 @@ def collect_inline_images( ): relparts = [] for path, info in images.items(): + if path.startswith("cid:"): + continue + data = request.urlopen(path) mimetype = data.headers["Content-Type"] @@ -223,7 +279,12 @@ def collect_inline_images( filewriter_fn(path, data.read(), "w+b") relparts.append( - Part(*mimetype.split("/"), path, cid=info.cid, desc=f"Image: {info.desc}") + Part( + *mimetype.split("/"), + path, + cid=info.cid, + desc=f"Image: {info.desc}", + ) ) return relparts @@ -233,12 +294,19 @@ def convert_markdown_to_html( origtext, draftpath, *, + cssfile=None, filewriter_fn=filewriter_fn, tempdir=None, extensions=None, + extension_configs=None, ): + # TODO extension_configs need to be handled differently + extension_configs = extension_configs or {} + extension_configs.setdefault("pymdownx.highlight", {}) + extension_configs["pymdownx.highlight"]["css_class"] = _CODEHILITE_CLASS + origtext, htmltext, images = markdown_with_inline_image_support( - origtext, extensions=extensions + origtext, extensions=extensions, extension_configs=extension_configs ) filewriter_fn(draftpath, origtext, encoding="utf-8") @@ -246,6 +314,8 @@ def convert_markdown_to_html( "text", "plain", draftpath, "Plain-text version", orig=True ) + htmltext = apply_styling(htmltext, cssfile) + htmlpath = draftpath.with_suffix(".html") filewriter_fn( htmlpath, htmltext, encoding="utf-8", errors="xmlcharrefreplace" @@ -269,56 +339,64 @@ def convert_markdown_to_html( class MIMETreeDFWalker: def __init__(self, *, visitor_fn=None, debug=False): - self._visitor_fn = visitor_fn + self._visitor_fn = visitor_fn or self._echovisit self._debug = debug + def _echovisit(self, node, ancestry, debugprint): + debugprint(f"node={node} ancestry={ancestry}") + def walk(self, root, *, visitor_fn=None): """ Recursive function to implement a depth-dirst walk of the MIME-tree rooted at `root`. """ - if isinstance(root, list): - root = Multipart("mixed", children=root) + if len(root) > 1: + root = Multipart("mixed", children=root) + else: + root = root[0] self._walk( root, - stack=[], + ancestry=[], visitor_fn=visitor_fn or self._visitor_fn, ) - def _walk(self, node, *, stack, visitor_fn): + def _walk(self, node, *, ancestry, visitor_fn): # Let's start by enumerating the parts at the current level. At the - # root level, stack will be the empty list, and we expect a multipart/* - # container at this level. Later, e.g. within a mutlipart/alternative - # container, the subtree will just be the alternative parts, while the - # top of the stack will be the multipart/alternative container, which - # we will process after the following loop. - - lead = f"{'| '*len(stack)}|-" + # root level, ancestry will be the empty list, and we expect a + # multipart/* container at this level. Later, e.g. within a + # mutlipart/alternative container, the subtree will just be the + # alternative parts, while the top of the ancestry will be the + # multipart/alternative container, which we will process after the + # following loop. + + lead = f"{'│ '*len(ancestry)}" if isinstance(node, Multipart): self.debugprint( - f"{lead}{node} parents={[s.subtype for s in stack]}" + f"{lead}├{node} ancestry={[s.subtype for s in ancestry]}" ) - # Depth-first, so push the current container onto the stack, - # then descend … - stack.append(node) - self.debugprint("| " * (len(stack) + 1)) + # Depth-first, so push the current container onto the ancestry + # stack, then descend … + ancestry.append(node) + self.debugprint(lead + "│ " * 2) for child in node.children: self._walk( child, - stack=stack, + ancestry=ancestry, visitor_fn=visitor_fn, ) - self.debugprint("| " * len(stack)) - assert stack.pop() == node + assert ancestry.pop() == node else: - self.debugprint(f"{lead}{node}") + self.debugprint(f"{lead}├{node}") + + if False and ancestry: + self.debugprint(lead[:-1] + " │") if visitor_fn: - visitor_fn(node, stack, debugprint=self.debugprint) + visitor_fn(node, ancestry, debugprint=self.debugprint) def debugprint(self, s, **kwargs): if self._debug: @@ -401,6 +479,7 @@ def do_massage( cmd_f, *, extensions=None, + cssfile=None, converter=convert_markdown_to_html, only_build=False, tempdir=None, @@ -426,17 +505,25 @@ def do_massage( cmds.flush() extensions = extensions.split(",") if extensions else [] - tree = converter(draft_f.read(), draftpath, tempdir=tempdir, extensions=extensions) + tree = converter( + draft_f.read(), + draftpath, + cssfile=cssfile, + tempdir=tempdir, + extensions=extensions, + ) mimetree = MIMETreeDFWalker(debug=debug_walk) - def visitor_fn(item, stack, *, debugprint=None): + state = dict(pos=1, tags={}, parts=1) + + def visitor_fn(item, ancestry, *, debugprint=None): """ Visitor function called for every node (part) of the MIME tree, depth-first, and responsible for telling NeoMutt how to assemble the tree. """ - KILL_LINE=r'\Ca\Ck' + KILL_LINE = r"\Ca\Ck" if isinstance(item, Part): # We've hit a leaf-node, i.e. an alternative or a related part @@ -447,13 +534,27 @@ def do_massage( # The original source already exists in the NeoMutt tree, but # the underlying file may have been modified, so we need to # update the encoding, but that's it: + cmds.push("") cmds.push("") + + # We really just need to be able to assume that at this point, + # NeoMutt is at position 1, and that we've processed only this + # part so far. Nevermind about actual attachments, we can + # safely ignore those as they stay at the end. + assert state["pos"] == 1 + assert state["parts"] == 1 else: # … whereas all other parts need to be added, and they're all # considered to be temporary and inline: cmds.push(f"{item.path}") cmds.push("") + # This added a part at the end of the list of parts, and that's + # just how many parts we've seen so far, so it's position in + # the NeoMutt compose list is the count of parts + state["parts"] += 1 + state["pos"] = state["parts"] + # If the item (including the original) comes with additional # information, then we might just as well update the NeoMutt # tree now: @@ -462,29 +563,47 @@ def do_massage( elif isinstance(item, Multipart): # This node has children, but we already visited them (see - # above), and so they have been tagged in NeoMutt's compose - # window. Now it's just a matter of telling NeoMutt to do the - # appropriate grouping: + # above). The tags dictionary of State should contain a list of + # their positions in the NeoMutt compose window, so iterate those + # and tag the parts there: + for tag in state["tags"][item]: + cmds.push(f"{tag}") + if item.subtype == "alternative": cmds.push("") elif item.subtype in ("relative", "related"): cmds.push("") elif item.subtype == "multilingual": cmds.push("") + else: + raise NotImplementedError( + f"Handling of multipart/{item.subtype} is not implemented" + ) + + state["pos"] -= len(state["tags"][item]) - 1 + state["parts"] += 1 + del state["tags"][item] else: # We should never get here - assert not "is valid part" + raise RuntimeError(f"Type {type(item)} is unexpected: {item}") # If the item has a description, we might just as well add it if item.desc: cmds.push(f"{KILL_LINE}{item.desc}") - # Finally, if we're at non-root level, tag the new container, - # as it might itself be part of a container, to be processed - # one level up: - if stack: - cmds.push("") + if ancestry: + # If there's an ancestry, record the current (assumed) position in + # the NeoMutt compose window as needed-to-tag by our direct parent + # (i.e. the last item of the ancestry) + state["tags"].setdefault(ancestry[-1], []).append(state["pos"]) + + lead = "│ " * (len(ancestry) + 1) + "* " + debugprint( + f"{lead}ancestry={[a.subtype for a in ancestry]}\n" + f"{lead}children_positions={state['tags'][ancestry[-1]]}\n" + f"{lead}pos={state['pos']}, parts={state['parts']}" + ) # ----------------- # End of visitor_fn @@ -527,6 +646,7 @@ if __name__ == "__main__": pathlib.Path(args.MAILDRAFT), cmd_f, extensions=args.extensions, + cssfile=args.css_file, only_build=args.only_build, tempdir=args.tempdir, debug_commands=args.debug_commands, @@ -642,14 +762,17 @@ try: assert items[4][0].subtype == "relative" assert items[4][1] == 0 - def test_MIMETreeDFWalker_list_to_mixed(self, basic_mime_tree): + def test_MIMETreeDFWalker_list_to_mixed(self, const1): mimetree = MIMETreeDFWalker() items = [] - def visitor_fn(item, stack, debugprint): + def visitor_fn(item, ancestry, debugprint): items.append(item) - mimetree.walk([basic_mime_tree], visitor_fn=visitor_fn) + p = Part("text", "plain", const1) + mimetree.walk([p], visitor_fn=visitor_fn) + assert items[-1].subtype == "plain" + mimetree.walk([p, p], visitor_fn=visitor_fn) assert items[-1].subtype == "mixed" def test_MIMETreeDFWalker_visitor_in_constructor( @@ -657,7 +780,7 @@ try: ): items = [] - def visitor_fn(item, stack, debugprint): + def visitor_fn(item, ancestry, debugprint): items.append(item) mimetree = MIMETreeDFWalker(visitor_fn=visitor_fn) @@ -669,7 +792,7 @@ try: return StringIO(text or const1) def test_do_massage_basic(self, const1, string_io, capsys): - def converter(drafttext, draftpath, extensions, tempdir): + def converter(drafttext, draftpath, cssfile, extensions, tempdir): return Part("text", "plain", draftpath, orig=True) do_massage( @@ -693,7 +816,7 @@ try: def test_do_massage_fulltree( self, string_io, const1, basic_mime_tree, capsys ): - def converter(drafttext, draftpath, extensions, tempdir): + def converter(drafttext, draftpath, cssfile, extensions, tempdir): return basic_mime_tree do_massage( @@ -873,6 +996,23 @@ try: assert tree.children[1].path == written[0] assert written[1] == request.urlopen(test_png).read() + def test_converter_tree_inline_image_cid( + self, const1, fake_filewriter + ): + text = f"![inline base64 image](cid:{const1})" + path = pathlib.Path(const1) + tree = convert_markdown_to_html( + text, + path, + filewriter_fn=fake_filewriter, + related_to_html_only=False, + ) + assert len(tree.children) == 2 + assert tree.children[0].cid != const1 + assert tree.children[0].type != "image" + assert tree.children[1].cid != const1 + assert tree.children[1].type != "image" + def test_inline_image_collection( self, test_png, const1, const2, fake_filewriter ): @@ -889,5 +1029,22 @@ try: assert relparts[0].cid == const1 assert relparts[0].desc.endswith(const2) + def test_apply_stylesheet(self): + if _PYNLINER: + html = "

Hello, world!

" + css = "p { color:red }" + out = apply_styling(html, css) + assert 'p style="color' in out + + def test_apply_stylesheet_pygments(self): + if _PYGMENTS_CSS: + html = ( + f'
' + "
def foo():\n    return
" + ) + out = apply_styling(html, _PYGMENTS_CSS) + assert f'{_CODEHILITE_CLASS}" style="' in out + + except ImportError: pass