#!/usr/bin/python3 # # NeoMutt helper script to create multipart/* emails with Markdown → HTML # alternative conversion, and handling of inline images, using NeoMutt's # ability to manually craft MIME trees, but automating this process. # # Configuration: # neomuttrc (needs to be a single line): # set my_mdwn_extensions="extra,admonition,codehilite,sane_lists,smarty" # macro compose B "\ # source '$my_confdir/buildmimetree.py \ # --tempdir $tempdir --extensions $my_mdwn_extensions \ # --css-file $my_confdir/htmlmail.css |'\ # sourc e \$my_mdwn_postprocess_cmd_file\ # " "Convert message into a modern MIME tree with inline images" # # (Yes, we need to call source twice, as mutt only starts to process output # from a source command when the command exits, and since we need to react # to the output, we need to be invoked again, using a $my_ variable to pass # information) # # Requirements: # - python3 # - python3-markdown # - python3-beautifulsoup4 # Optional: # - pytest # - 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 # # Copyright © 2023 martin f. krafft # Released under the GPL-2+ licence, just like Mutt itself. # import sys import os import os.path import pathlib import markdown import tempfile import argparse import re import mimetypes import bs4 from collections import namedtuple, OrderedDict from markdown.extensions import Extension from markdown.inlinepatterns import ImageInlineProcessor, IMAGE_LINK_RE from email.utils import make_msgid from urllib import request def parse_cli_args(*args, **kwargs): parser = argparse.ArgumentParser( description=( "NeoMutt helper to turn text/markdown email parts " "into full-fledged MIME trees" ) ) parser.epilog = ( "Copyright © 2023 martin f. krafft .\n" "Released under the MIT licence" ) parser.add_argument( "--extensions", metavar="EXT[,EXT[,EXT]]", type=str, default="", help="Markdown extension to use (comma-separated list)", ) if _PYNLINER: parser.add_argument( "--css-file", metavar="FILE", type=pathlib.Path, default=os.devnull, help="CSS file to merge with the final HTML", ) else: parser.set_defaults(css_file=None) parser.add_argument( "--related-to-html-only", action="store_true", help="Make related content be sibling to HTML parts only", ) def positive_integer(value): try: if int(value) > 0: return int(value) except ValueError: pass raise ValueError("Must be a positive integer") parser.add_argument( "--max-number-other-attachments", metavar="INTEGER", type=positive_integer, default=20, help="Maximum number of other attachments to expect", ) parser.add_argument( "--only-build", "--just-build", action="store_true", help="Only build, don't send the message", ) parser.add_argument( "--tempdir", metavar="DIR", type=pathlib.Path, help="Specify temporary directory to use for attachments", ) parser.add_argument( "--debug-commands", action="store_true", help="Turn on debug logging of commands generated to stderr", ) parser.add_argument( "--debug-walk", action="store_true", help="Turn on debugging to stderr of the MIME tree walk", ) subp = parser.add_subparsers(help="Sub-command parsers", dest="mode") massage_p = subp.add_parser( "massage", help="Massaging phase (internal use)" ) massage_p.add_argument( "--write-commands-to", "-o", metavar="FILE", dest="cmdpath", type=pathlib.Path, required=True, help="Temporary file path to write commands to", ) massage_p.add_argument( "MAILDRAFT", nargs="?", type=pathlib.Path, help="If provided, the script is invoked as editor on the mail draft", ) return parser.parse_args(*args, **kwargs) # [ IMAGE HANDLING ] ########################################################## InlineImageInfo = namedtuple( "InlineImageInfo", ["cid", "desc"], defaults=[None] ) class ImageRegistry: def __init__(self): self._images = OrderedDict() def register(self, path, description=None): path = os.path.expanduser(path) if path.startswith("/"): path = f"file://{path}" cid = make_msgid()[1:-1] self._images[path] = InlineImageInfo(cid, description) return cid def __iter__(self): return self._images.__iter__() def __getitem__(self, idx): return self._images.__getitem__(idx) def __len__(self): return self._images.__len__() def items(self): return self._images.items() def __repr__(self): return f"" def __str__(self): return self._images.__str__() class InlineImageExtension(Extension): class RelatedImageInlineProcessor(ImageInlineProcessor): def __init__(self, re, md, registry): super().__init__(re, md) self._registry = registry def handleMatch(self, m, data): el, start, end = super().handleMatch(m, data) if "src" in el.attrib: src = el.attrib["src"] if "://" not in src or src.startswith("file://"): # We only inline local content cid = self._registry.register( el.attrib["src"], el.attrib.get("title", el.attrib.get("alt")), ) el.attrib["src"] = f"cid:{cid}" return el, start, end def __init__(self, registry): super().__init__() self._image_registry = registry INLINE_PATTERN_NAME = "image_link" def extendMarkdown(self, md): md.registerExtension(self) inline_image_proc = self.RelatedImageInlineProcessor( IMAGE_LINK_RE, md, self._image_registry ) md.inlinePatterns.register( inline_image_proc, InlineImageExtension.INLINE_PATTERN_NAME, 150 ) def markdown_with_inline_image_support( text, *, mdwn=None, image_registry=None, extensions=None, extension_configs=None, ): registry = ( image_registry if image_registry is not None else ImageRegistry() ) inline_image_handler = InlineImageExtension(registry=registry) extensions = extensions or [] extensions.append(inline_image_handler) mdwn = markdown.Markdown( extensions=extensions, extension_configs=extension_configs ) htmltext = mdwn.convert(text) def replace_image_with_cid(matchobj): for m in (matchobj.group(1), f"file://{matchobj.group(1)}"): if m in registry: return f"(cid:{registry[m].cid}" return matchobj.group(0) text = re.sub(r"\(([^)\s]+)", replace_image_with_cid, text) return text, htmltext, registry, mdwn # [ 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 ] ######################################################## class Part( namedtuple( "Part", ["type", "subtype", "path", "desc", "cid", "orig"], defaults=[None, None, False], ) ): def __str__(self): ret = f"<{self.type}/{self.subtype}>" if self.cid: ret = f"{ret} cid:{self.cid}" if self.orig: ret = f"{ret} ORIGINAL" return ret class Multipart( namedtuple("Multipart", ["subtype", "children", "desc"], defaults=[None]) ): def __str__(self): return f" children={len(self.children)}" def __hash__(self): return hash(str(self.subtype) + "".join(str(self.children))) def filereader_fn(path, mode="r", **kwargs): with open(path, mode, **kwargs) as in_f: return in_f.read() def filewriter_fn(path, content, mode="w", **kwargs): with open(path, mode, **kwargs) as out_f: out_f.write(content) def collect_inline_images( image_registry, *, tempdir=None, filewriter_fn=filewriter_fn ): relparts = [] for path, info in image_registry.items(): if path.startswith("cid:"): continue data = request.urlopen(path) mimetype = data.headers["Content-Type"] ext = mimetypes.guess_extension(mimetype) tempfilename = tempfile.mkstemp(prefix="img", suffix=ext, dir=tempdir) path = pathlib.Path(tempfilename[1]) filewriter_fn(path, data.read(), "w+b") desc = ( f'Inline image: "{info.desc}"' if info.desc else f"Inline image {str(len(relparts)+1)}" ) relparts.append( Part(*mimetype.split("/"), path, cid=info.cid, desc=desc) ) return relparts EMAIL_SIG_SEP = "\n-- \n" HTML_SIG_MARKER = "=htmlsig " def make_html_doc(body, sig=None): ret = ( "\n" "\n" "\n" '\n' # noqa: E501 '\n' # noqa: E501 "\n" "\n" f"{body}\n" ) if sig: nl = "\n" ret = ( f'{ret}
{EMAIL_SIG_SEP.strip(nl)}\n' # noqa: E501 f"{sig}\n" "
" ) return f"{ret}\n \n" def make_text_mail(text, sig=None): return EMAIL_SIG_SEP.join((text, sig)) if sig else text def extract_signature(text, *, filereader_fn=filereader_fn): parts = text.split(EMAIL_SIG_SEP, 1) if len(parts) == 1: return text, None, None lines = parts[1].splitlines() if lines[0].startswith(HTML_SIG_MARKER): path = pathlib.Path(re.split(r" +", lines.pop(0), maxsplit=1)[1]) textsig = "\n".join(lines) sig_input = filereader_fn(path.expanduser()) soup = bs4.BeautifulSoup(sig_input, "html.parser") style = str(soup.style.extract()) if soup.style else "" for sig_selector in ( "#signature", "#signatur", "#emailsig", ".signature", ".signatur", ".emailsig", "body", "div", ): sig = soup.select_one(sig_selector) if sig: break if not sig: return parts[0], textsig, style + sig_input if sig.attrs.get("id") == "signature": sig = "".join(str(c) for c in sig.children) return parts[0], textsig, style + str(sig) return parts[0], parts[1], None def convert_markdown_to_html( origtext, draftpath, *, related_to_html_only=False, css=None, filewriter_fn=filewriter_fn, filereader_fn=filereader_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, textsig, htmlsig = extract_signature( origtext, filereader_fn=filereader_fn ) ( origtext, htmltext, image_registry, mdwn, ) = markdown_with_inline_image_support( origtext, extensions=extensions, extension_configs=extension_configs ) if htmlsig: if not textsig: # TODO: decide what to do if there is no plain-text version raise NotImplementedError("HTML signature but no text alternative") soup = bs4.BeautifulSoup(htmlsig, "html.parser") for img in soup.find_all("img"): uri = img.attrs["src"] desc = img.attrs.get("title", img.attrs.get("alt")) cid = image_registry.register(uri, desc) img.attrs["src"] = f"cid:{cid}" htmlsig = str(soup) elif textsig: ( textsig, htmlsig, image_registry, mdwn, ) = markdown_with_inline_image_support( textsig, extensions=extensions, extension_configs=extension_configs, image_registry=image_registry, mdwn=mdwn, ) origtext = make_text_mail(origtext, textsig) filewriter_fn(draftpath, origtext, encoding="utf-8") textpart = Part( "text", "plain", draftpath, "Plain-text version", orig=True ) htmltext = make_html_doc(htmltext, htmlsig) htmltext = apply_styling(htmltext, css) htmlpath = draftpath.with_suffix(".html") filewriter_fn( htmlpath, htmltext, encoding="utf-8", errors="xmlcharrefreplace" ) htmlpart = Part("text", "html", htmlpath, "HTML version") imgparts = collect_inline_images( image_registry, tempdir=tempdir, filewriter_fn=filewriter_fn ) if related_to_html_only: # If there are inline image part, they will be contained within a # multipart/related part along with the HTML part only if imgparts: # replace htmlpart with a multipart/related container of the HTML # parts and the images htmlpart = Multipart( "relative", [htmlpart] + imgparts, "Group of related content" ) return Multipart( "alternative", [textpart, htmlpart], "Group of alternative content" ) else: # If there are inline image part, they will be siblings to the # multipart/alternative tree within a multipart/related part altpart = Multipart( "alternative", [textpart, htmlpart], "Group of alternative content" ) if imgparts: return Multipart( "relative", [altpart] + imgparts, "Group of related content" ) else: return altpart class MIMETreeDFWalker: def __init__(self, *, visitor_fn=None, debug=False): 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): if len(root) > 1: root = Multipart("mixed", children=root) else: root = root[0] self._walk( root, ancestry=[], descendents=[], visitor_fn=visitor_fn or self._visitor_fn, ) def _walk(self, node, *, ancestry, descendents, visitor_fn): # Let's start by enumerating the parts at the current level. At the # 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} ancestry={[s.subtype for s in ancestry]}" ) # 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, ancestry=ancestry, descendents=descendents, visitor_fn=visitor_fn, ) assert ancestry.pop() == node sibling_descendents = descendents descendents.extend(node.children) else: self.debugprint(f"{lead}├{node}") sibling_descendents = descendents if False and ancestry: self.debugprint(lead[:-1] + " │") if visitor_fn: visitor_fn( node, ancestry, sibling_descendents, debugprint=self.debugprint ) def debugprint(self, s, **kwargs): if self._debug: print(s, file=sys.stderr, **kwargs) # [ RUN MODES ] ############################################################### class MuttCommands: """ Stupid class to interface writing out Mutt commands. This is quite a hack to deal with the fact that Mutt runs "push" commands in reverse order, so all of a sudden, things become very complicated when mixing with "real" commands. Hence we keep two sets of commands, and one set of pushes. Commands are added to the first until a push is added, after which commands are added to the second set of commands. On flush(), the first set is printed, followed by the pushes in reverse, and then the second set is printed. All 3 sets are then cleared. """ def __init__(self, out_f=sys.stdout, *, debug=False): self._cmd1, self._push, self._cmd2 = [], [], [] self._out_f = out_f self._debug = debug def cmd(self, s): self.debugprint(s) if self._push: self._cmd2.append(s) else: self._cmd1.append(s) def push(self, s): s = s.replace('"', r"\"") s = f'push "{s}"' self.debugprint(s) self._push.insert(0, s) def flush(self): print( "\n".join(self._cmd1 + self._push + self._cmd2), file=self._out_f ) self._cmd1, self._push, self._cmd2 = [], [], [] def debugprint(self, s, **kwargs): if self._debug: print(s, file=sys.stderr, **kwargs) def do_setup( *, out_f=sys.stdout, temppath=None, tempdir=None, debug_commands=False, ): temppath = temppath or pathlib.Path( tempfile.mkstemp(prefix="muttmdwn-", dir=tempdir)[1] ) cmds = MuttCommands(out_f, debug=debug_commands) editor = f"{' '.join(sys.argv)} massage --write-commands-to {temppath}" cmds.cmd('set my_editor="$editor"') cmds.cmd('set my_edit_headers="$edit_headers"') cmds.cmd(f'set editor="{editor}"') cmds.cmd("unset edit_headers") cmds.cmd(f"set my_mdwn_postprocess_cmd_file={temppath}") cmds.push("") cmds.flush() def do_massage( draft_f, draftpath, cmd_f, *, extensions=None, css_f=None, converter=convert_markdown_to_html, related_to_html_only=True, only_build=False, max_other_attachments=20, tempdir=None, debug_commands=False, debug_walk=False, ): # Here's the big picture: we're being invoked as the editor on the email # draft, and whatever commands we write to the file given as cmdpath will # be run by the second source command in the macro definition. # Let's start by cleaning up what the setup did (see above), i.e. we # restore the $editor and $edit_headers variables, and also unset the # variable used to identify the command file we're currently writing # to. cmds = MuttCommands(cmd_f, debug=debug_commands) cmds.cmd('set editor="$my_editor"') cmds.cmd('set edit_headers="$my_edit_headers"') cmds.cmd("unset my_editor") cmds.cmd("unset my_edit_headers") # let's flush those commands, as there'll be a lot of pushes from now # on, which need to be run in reverse order cmds.flush() extensions = extensions.split(",") if extensions else [] tree = converter( draft_f.read(), draftpath, css=css_f.read() if css_f else None, related_to_html_only=related_to_html_only, tempdir=tempdir, extensions=extensions, ) mimetree = MIMETreeDFWalker(debug=debug_walk) state = dict(pos=1, tags={}, parts=1) def visitor_fn(item, ancestry, descendents, *, 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" if isinstance(item, Part): # We've hit a leaf-node, i.e. an alternative or a related part # with actual content. # Let's add the part if item.orig: # 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: if item.cid: cmds.push(f"{KILL_LINE}{item.cid}") # Now for the biggest hack in this script, which is to handle # attachments, such as PDFs, that aren't related or alternatives. # The problem is that when we add an inline image, it always gets # appended to the list, i.e. inserted *after* other attachments. # Since we don't know the number of attachments, we also cannot # infer the postition of the new attachment. Therefore, we bubble # it all the way to the top, only to then move it down again: if state["pos"] > 1: # skip for the first part for i in range(max_other_attachments): # could use any number here, but has to be larger than the # number of possible attachments. The performance # difference of using a high number is negligible. # Bubble up the new part cmds.push("") # As we push the part to the right position in the list (i.e. # the last of the subset of attachments this script added), we # must handle the situation that subtrees are skipped by # NeoMutt. Hence, the actual number of positions to move down # is decremented by the number of descendents so far # encountered. for i in range(1, state["pos"] - len(descendents)): cmds.push("") elif isinstance(item, Multipart): # This node has children, but we already visited them (see # 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: n_tags = len(state["tags"][item]) 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"] -= n_tags - 1 state["parts"] += 1 else: # We should never get here 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}") 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}descendents={[d.subtype for d in descendents]}\n" f"{lead}children_positions={state['tags'][ancestry[-1]]}\n" f"{lead}pos={state['pos']}, parts={state['parts']}" ) # ----------------- # End of visitor_fn # Let's walk the tree and visit every node with our fancy visitor # function mimetree.walk(tree, visitor_fn=visitor_fn) if not only_build: cmds.push("") # Finally, cleanup. Since we're responsible for removing the temporary # file, how's this for a little hack? try: filename = cmd_f.name except AttributeError: filename = "pytest_internal_file" cmds.cmd(f"source 'rm -f {filename}|'") cmds.cmd("unset my_mdwn_postprocess_cmd_file") cmds.flush() # [ CLI ENTRY ] ############################################################### if __name__ == "__main__": args = parse_cli_args() if args.mode is None: do_setup( tempdir=args.tempdir, debug_commands=args.debug_commands, ) elif args.mode == "massage": with open(args.MAILDRAFT, "r") as draft_f, open( args.cmdpath, "w" ) as cmd_f, open(args.css_file, "r") as css_f: do_massage( draft_f, args.MAILDRAFT, cmd_f, extensions=args.extensions, css_f=css_f, related_to_html_only=args.related_to_html_only, max_other_attachments=args.max_number_other_attachments, only_build=args.only_build, tempdir=args.tempdir, debug_commands=args.debug_commands, debug_walk=args.debug_walk, ) # [ TESTS ] ################################################################### try: import pytest from io import StringIO class Tests: @pytest.fixture def const1(self): return "CONSTANT STRING 1" @pytest.fixture def const2(self): return "CONSTANT STRING 2" # NOTE: tests using the capsys fixture must specify sys.stdout to the # functions they call, else old stdout is used and not captured @pytest.mark.muttctrl def test_MuttCommands_cmd(self, const1, const2, capsys): "Assert order of commands" cmds = MuttCommands(out_f=sys.stdout) cmds.cmd(const1) cmds.cmd(const2) cmds.flush() captured = capsys.readouterr() assert captured.out == "\n".join((const1, const2, "")) @pytest.mark.muttctrl def test_MuttCommands_push(self, const1, const2, capsys): "Assert reverse order of pushes" cmds = MuttCommands(out_f=sys.stdout) cmds.push(const1) cmds.push(const2) cmds.flush() captured = capsys.readouterr() assert ( captured.out == ('"\npush "'.join(("", const2, const1, "")))[2:-6] ) @pytest.mark.muttctrl def test_MuttCommands_push_escape(self, const1, const2, capsys): cmds = MuttCommands(out_f=sys.stdout) cmds.push(f'"{const1}"') cmds.flush() captured = capsys.readouterr() assert f'"\\"{const1}\\""' in captured.out @pytest.mark.muttctrl def test_MuttCommands_cmd_push_mixed(self, const1, const2, capsys): "Assert reverse order of pushes" cmds = MuttCommands(out_f=sys.stdout) lines = ["000", "001", "010", "011", "100", "101", "110", "111"] for i in range(2): cmds.cmd(lines[4 * i + 0]) cmds.cmd(lines[4 * i + 1]) cmds.push(lines[4 * i + 2]) cmds.push(lines[4 * i + 3]) cmds.flush() captured = capsys.readouterr() lines_out = captured.out.splitlines() assert lines[0] in lines_out[0] assert lines[1] in lines_out[1] assert lines[7] in lines_out[2] assert lines[6] in lines_out[3] assert lines[3] in lines_out[4] assert lines[2] in lines_out[5] assert lines[4] in lines_out[6] assert lines[5] in lines_out[7] @pytest.fixture def mime_tree_related_to_alternative(self): return Multipart( "relative", children=[ Multipart( "alternative", children=[ Part( "text", "plain", "part.txt", desc="Plain", orig=True, ), Part("text", "html", "part.html", desc="HTML"), ], desc="Alternative", ), Part( "text", "png", "logo.png", cid="logo.png", desc="Logo" ), ], desc="Related", ) @pytest.fixture def mime_tree_related_to_html(self): return Multipart( "alternative", children=[ Part( "text", "plain", "part.txt", desc="Plain", orig=True, ), Multipart( "relative", children=[ Part("text", "html", "part.html", desc="HTML"), Part( "text", "png", "logo.png", cid="logo.png", desc="Logo", ), ], desc="Related", ), ], desc="Alternative", ) @pytest.mark.treewalk def test_MIMETreeDFWalker_depth_first_walk( self, mime_tree_related_to_alternative ): mimetree = MIMETreeDFWalker() items = [] def visitor_fn(item, ancestry, descendents, debugprint): items.append((item, len(ancestry), len(descendents))) mimetree.walk( mime_tree_related_to_alternative, visitor_fn=visitor_fn ) assert len(items) == 5 assert items[0][0].subtype == "plain" assert items[0][1] == 2 assert items[0][2] == 0 assert items[1][0].subtype == "html" assert items[1][1] == 2 assert items[1][2] == 0 assert items[2][0].subtype == "alternative" assert items[2][1] == 1 assert items[2][2] == 2 assert items[3][0].subtype == "png" assert items[3][1] == 1 assert items[3][2] == 2 assert items[4][0].subtype == "relative" assert items[4][1] == 0 assert items[4][2] == 4 @pytest.mark.treewalk def test_MIMETreeDFWalker_list_to_mixed(self, const1): mimetree = MIMETreeDFWalker() items = [] def visitor_fn(item, ancestry, descendents, debugprint): items.append(item) 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" @pytest.mark.treewalk def test_MIMETreeDFWalker_visitor_in_constructor( self, mime_tree_related_to_alternative ): items = [] def visitor_fn(item, ancestry, descendents, debugprint): items.append(item) mimetree = MIMETreeDFWalker(visitor_fn=visitor_fn) mimetree.walk(mime_tree_related_to_alternative) assert len(items) == 5 @pytest.fixture def string_io(self, const1, text=None): return StringIO(text or const1) @pytest.mark.massage def test_do_massage_basic(self, const1, string_io, capsys): def converter( drafttext, draftpath, css, related_to_html_only, extensions, tempdir, ): return Part("text", "plain", draftpath, orig=True) do_massage( draft_f=string_io, draftpath=const1, cmd_f=sys.stdout, converter=converter, ) captured = capsys.readouterr() lines = captured.out.splitlines() assert '="$my_editor"' in lines.pop(0) assert '="$my_edit_headers"' in lines.pop(0) assert "unset my_editor" == lines.pop(0) assert "unset my_edit_headers" == lines.pop(0) assert "send-message" in lines.pop(0) assert "update-encoding" in lines.pop(0) assert "first-entry" in lines.pop(0) assert "source 'rm -f " in lines.pop(0) assert "unset my_mdwn_postprocess_cmd_file" == lines.pop(0) @pytest.mark.massage def test_do_massage_fulltree( self, string_io, const1, mime_tree_related_to_alternative, capsys ): def converter( drafttext, draftpath, css, related_to_html_only, extensions, tempdir, ): return mime_tree_related_to_alternative max_attachments = 5 do_massage( draft_f=string_io, draftpath=const1, cmd_f=sys.stdout, max_other_attachments=max_attachments, converter=converter, ) captured = capsys.readouterr() lines = captured.out.splitlines()[4:-2] assert "first-entry" in lines.pop() assert "update-encoding" in lines.pop() assert "Plain" in lines.pop() assert "part.html" in lines.pop() assert "toggle-unlink" in lines.pop() for i in range(max_attachments): assert "move-up" in lines.pop() assert "move-down" in lines.pop() assert "HTML" in lines.pop() assert "jump>1" in lines.pop() assert "jump>2" in lines.pop() assert "group-alternatives" in lines.pop() assert "Alternative" in lines.pop() assert "logo.png" in lines.pop() assert "toggle-unlink" in lines.pop() assert "content-id" in lines.pop() for i in range(max_attachments): assert "move-up" in lines.pop() assert "move-down" in lines.pop() assert "Logo" in lines.pop() assert "jump>1" in lines.pop() assert "jump>4" in lines.pop() assert "group-related" in lines.pop() assert "Related" in lines.pop() assert "send-message" in lines.pop() assert len(lines) == 0 @pytest.fixture def fake_filewriter(self): class FileWriter: def __init__(self): self._writes = [] def __call__(self, path, content, mode="w", **kwargs): self._writes.append((path, content)) def pop(self, index=-1): return self._writes.pop(index) return FileWriter() @pytest.fixture def markdown_non_converter(self, const1, const2): return lambda s, text: f"{const1}{text}{const2}" @pytest.mark.converter def test_converter_tree_basic(self, const1, const2, fake_filewriter): path = pathlib.Path(const2) tree = convert_markdown_to_html( const1, path, filewriter_fn=fake_filewriter ) assert tree.subtype == "alternative" assert len(tree.children) == 2 assert tree.children[0].subtype == "plain" assert tree.children[0].path == path assert tree.children[0].orig assert tree.children[1].subtype == "html" assert tree.children[1].path == path.with_suffix(".html") def test_converter_writes( self, const1, const2, fake_filewriter, monkeypatch, markdown_non_converter, ): path = pathlib.Path(const2) with monkeypatch.context() as m: m.setattr(markdown.Markdown, "convert", markdown_non_converter) convert_markdown_to_html( const1, path, filewriter_fn=fake_filewriter ) assert (path, const1) == fake_filewriter.pop(0) written = fake_filewriter.pop(0) assert path.with_suffix(".html") == written[0] assert const1 in written[1] @pytest.mark.imgproc def test_markdown_inline_image_processor(self): imgpath1 = "file:/path/to/image.png" imgpath2 = "file:///path/to/image.png?url=params" imgpath3 = "/path/to/image.png" text = f"""![inline local image]({imgpath1}) ![image inlined with newline]({imgpath2}) ![image local path]({imgpath3})""" text, html, images, mdwn = markdown_with_inline_image_support(text) # local paths have been normalised to URLs: imgpath3 = f"file://{imgpath3}" assert 'src="cid:' in html assert "](cid:" in text assert len(images) == 3 assert imgpath1 in images assert imgpath2 in images assert imgpath3 in images assert images[imgpath1].cid != images[imgpath2].cid assert images[imgpath1].cid != images[imgpath3].cid assert images[imgpath2].cid != images[imgpath3].cid @pytest.mark.imgproc def test_markdown_inline_image_processor_title_to_desc(self, const1): imgpath = "file:///path/to/image.png" text = f'![inline local image]({imgpath} "{const1}")' text, html, images, mdwn = markdown_with_inline_image_support(text) assert images[imgpath].desc == const1 @pytest.mark.imgproc def test_markdown_inline_image_processor_alt_to_desc(self, const1): imgpath = "file:///path/to/image.png" text = f"![{const1}]({imgpath})" text, html, images, mdwn = markdown_with_inline_image_support(text) assert images[imgpath].desc == const1 @pytest.mark.imgproc def test_markdown_inline_image_processor_title_over_alt_desc( self, const1, const2 ): imgpath = "file:///path/to/image.png" text = f'![{const1}]({imgpath} "{const2}")' text, html, images, mdwn = markdown_with_inline_image_support(text) assert images[imgpath].desc == const2 @pytest.mark.imgproc def test_markdown_inline_image_not_external(self): imgpath = "https://path/to/image.png" text = f"![inline image]({imgpath})" text, html, images, mdwn = markdown_with_inline_image_support(text) assert 'src="cid:' not in html assert "](cid:" not in text assert len(images) == 0 @pytest.mark.imgproc def test_markdown_inline_image_local_file(self): imgpath = "/path/to/image.png" text = f"![inline image]({imgpath})" text, html, images, mdwn = markdown_with_inline_image_support(text) for k, v in images.items(): assert k == f"file://{imgpath}" break @pytest.mark.imgproc def test_markdown_inline_image_expanduser(self): imgpath = pathlib.Path("~/image.png") text = f"![inline image]({imgpath})" text, html, images, mdwn = markdown_with_inline_image_support(text) for k, v in images.items(): assert k == f"file://{imgpath.expanduser()}" break @pytest.fixture def test_png(self): return ( "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAE" "AAAABCAAAAAA6fptVAAAACklEQVQI12P4DwABAQEAG7buVgAA" ) @pytest.mark.imgproc def test_markdown_inline_image_processor_base64(self, test_png): text = f"![1px white inlined]({test_png})" text, html, images, mdwn = markdown_with_inline_image_support(text) assert 'src="cid:' in html assert "](cid:" in text assert len(images) == 1 assert test_png in images @pytest.mark.converter def test_converter_tree_inline_image_base64( self, test_png, const1, fake_filewriter ): text = f"![inline base64 image]({test_png})" path = pathlib.Path(const1) tree = convert_markdown_to_html( text, path, filewriter_fn=fake_filewriter, related_to_html_only=False, ) assert tree.subtype == "relative" assert tree.children[0].subtype == "alternative" assert tree.children[1].subtype == "png" written = fake_filewriter.pop() assert tree.children[1].path == written[0] assert written[1] == request.urlopen(test_png).read() @pytest.mark.converter def test_converter_tree_inline_image_base64_related_to_html( self, test_png, const1, fake_filewriter ): text = f"![inline base64 image]({test_png})" path = pathlib.Path(const1) tree = convert_markdown_to_html( text, path, filewriter_fn=fake_filewriter, related_to_html_only=True, ) assert tree.subtype == "alternative" assert tree.children[1].subtype == "relative" assert tree.children[1].children[1].subtype == "png" written = fake_filewriter.pop() assert tree.children[1].children[1].path == written[0] assert written[1] == request.urlopen(test_png).read() @pytest.mark.converter 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" @pytest.mark.imgcoll def test_inline_image_collection( self, test_png, const1, const2, fake_filewriter ): test_images = {test_png: InlineImageInfo(cid=const1, desc=const2)} relparts = collect_inline_images( test_images, filewriter_fn=fake_filewriter ) written = fake_filewriter.pop() assert b"PNG" in written[1] assert relparts[0].subtype == "png" assert relparts[0].path == written[0] assert relparts[0].cid == const1 assert relparts[0].desc.endswith(const2) if _PYNLINER: @pytest.mark.styling def test_apply_stylesheet(self): html = "

Hello, world!

" css = "p { color:red }" out = apply_styling(html, css) assert 'p style="color' in out @pytest.mark.styling def test_massage_styling_to_converter(self, string_io, const1): css = "p { color:red }" css_f = StringIO(css) out_f = StringIO() css_applied = [] def converter( drafttext, draftpath, css, related_to_html_only, extensions, tempdir, ): css_applied.append(css) return Part("text", "plain", draftpath, orig=True) do_massage( draft_f=string_io, draftpath=const1, cmd_f=out_f, css_f=css_f, converter=converter, ) assert css_applied[0] == css @pytest.mark.converter def test_converter_apply_styles( self, const1, fake_filewriter, monkeypatch ): path = pathlib.Path(const1) text = "Hello, world!" css = "p { color:red }" with monkeypatch.context() as m: m.setattr( markdown.Markdown, "convert", lambda s, t: f"

{t}

", ) convert_markdown_to_html( text, path, css=css, filewriter_fn=fake_filewriter ) assert "color: red" in fake_filewriter.pop()[1] if _PYGMENTS_CSS: @pytest.mark.styling def test_apply_stylesheet_pygments(self): html = ( f'
' "
def foo():\n    return
" ) out = apply_styling(html, _PYGMENTS_CSS) assert f'{_CODEHILITE_CLASS}" style="' in out @pytest.mark.massage def test_mime_tree_relative_within_alternative( self, string_io, const1, capsys, mime_tree_related_to_html ): def converter( drafttext, draftpath, css, related_to_html_only, extensions, tempdir, ): return mime_tree_related_to_html do_massage( draft_f=string_io, draftpath=const1, cmd_f=sys.stdout, converter=converter, ) captured = capsys.readouterr() lines = captured.out.splitlines()[4:-2] assert "first-entry" in lines.pop() assert "update-encoding" in lines.pop() assert "Plain" in lines.pop() assert "part.html" in lines.pop() assert "toggle-unlink" in lines.pop() assert "move-up" in lines.pop() while True: top = lines.pop() if "move-up" not in top: break assert "move-down" in top assert "HTML" in lines.pop() assert "logo.png" in lines.pop() assert "toggle-unlink" in lines.pop() assert "content-id" in lines.pop() assert "move-up" in lines.pop() while True: top = lines.pop() if "move-up" not in top: break assert "move-down" in top assert "move-down" in lines.pop() assert "Logo" in lines.pop() assert "jump>2" in lines.pop() assert "jump>3" in lines.pop() assert "group-related" in lines.pop() assert "Related" in lines.pop() assert "jump>1" in lines.pop() assert "jump>2" in lines.pop() assert "group-alternative" in lines.pop() assert "Alternative" in lines.pop() assert "send-message" in lines.pop() assert len(lines) == 0 @pytest.mark.massage def test_mime_tree_nested_trees_does_not_break_positioning( self, string_io, const1, capsys ): def converter( drafttext, draftpath, css, related_to_html_only, extensions, tempdir, ): return Multipart( "relative", children=[ Multipart( "alternative", children=[ Part( "text", "plain", "part.txt", desc="Plain", orig=True, ), Multipart( "alternative", children=[ Part( "text", "plain", "part.txt", desc="Nested plain", ), Part( "text", "html", "part.html", desc="Nested HTML", ), ], desc="Nested alternative", ), ], desc="Alternative", ), Part( "text", "png", "logo.png", cid="logo.png", desc="Logo", ), ], desc="Related", ) do_massage( draft_f=string_io, draftpath=const1, cmd_f=sys.stdout, converter=converter, ) captured = capsys.readouterr() lines = captured.out.splitlines() while "logo.png" not in lines.pop(): pass lines.pop() assert "content-id" in lines.pop() assert "move-up" in lines.pop() while True: top = lines.pop() if "move-up" not in top: break assert "move-down" in top # Due to the nested trees, the number of descendents of the sibling # actually needs to be considered, not just the nieces. So to move # from position 1 to position 6, it only needs one # because that jumps over the entire sibling tree. Thus what # follows next must not be another assert "Logo" in lines.pop() @pytest.mark.sig def test_signature_extraction_no_signature(self, const1): assert (const1, None, None) == extract_signature(const1) @pytest.mark.sig def test_signature_extraction_just_text(self, const1, const2): origtext, textsig, htmlsig = extract_signature( f"{const1}{EMAIL_SIG_SEP}{const2}" ) assert origtext == const1 assert textsig == const2 assert htmlsig is None @pytest.mark.sig def test_signature_extraction_html(self, const1, const2): path = pathlib.Path("somepath") sigconst = "HTML signature from {path} but as a string" def filereader_fn(path): return ( f'
{sigconst.format(path=path)}
' ) origtext, textsig, htmlsig = extract_signature( f"{const1}{EMAIL_SIG_SEP}{HTML_SIG_MARKER} {path}\n{const2}", filereader_fn=filereader_fn, ) assert origtext == const1 assert textsig == const2 assert htmlsig == sigconst.format(path=path) @pytest.mark.sig def test_signature_extraction_file_not_found(self, const1): path = pathlib.Path("/does/not/exist") with pytest.raises(FileNotFoundError): origtext, textsig, htmlsig = extract_signature( f"{const1}{EMAIL_SIG_SEP}{HTML_SIG_MARKER}{path}\n{const1}" ) @pytest.mark.imgproc def test_image_registry(self, const1): reg = ImageRegistry() cid = reg.register(const1) assert "@" in cid assert not cid.startswith("<") assert not cid.endswith(">") assert const1 in reg @pytest.mark.imgproc def test_image_registry_file_uri(self, const1): reg = ImageRegistry() reg.register("/some/path") for path in reg: assert path.startswith("file://") break @pytest.mark.converter @pytest.mark.sig def test_converter_signature_handling( self, const1, fake_filewriter, monkeypatch ): path = pathlib.Path(const1) mailparts = ( "This is the mail body\n", f"{EMAIL_SIG_SEP}", "This is a plain-text signature only", ) def filereader_fn(path): return "" with monkeypatch.context() as m: m.setattr(markdown.Markdown, "convert", lambda s, t: t) convert_markdown_to_html( "".join(mailparts), path, filewriter_fn=fake_filewriter, filereader_fn=filereader_fn, ) soup = bs4.BeautifulSoup(fake_filewriter.pop()[1], "html.parser") body = soup.body.contents assert mailparts[0] in body.pop(0) sig = soup.select_one("#signature") assert sig == body.pop(0) sep = sig.select_one("span.sig_separator") assert sep == sig.contents[0] assert f"\n{sep.text}\n" == EMAIL_SIG_SEP assert mailparts[2] in sig.contents[1] @pytest.mark.converter @pytest.mark.sig def test_converter_signature_handling_htmlsig( self, const1, fake_filewriter, monkeypatch ): path = pathlib.Path(const1) mailparts = ( "This is the mail body", f"{EMAIL_SIG_SEP}", f"{HTML_SIG_MARKER}{path}\n", "This is the plain-text version", ) htmlsig = "HTML Signature from {path}" def filereader_fn(path): return f'
{htmlsig.format(path=path)}
' def mdwn_fn(t): return t.upper() with monkeypatch.context() as m: m.setattr( markdown.Markdown, "convert", lambda s, t: mdwn_fn(t) ) convert_markdown_to_html( "".join(mailparts), path, filewriter_fn=fake_filewriter, filereader_fn=filereader_fn, ) soup = bs4.BeautifulSoup(fake_filewriter.pop()[1], "html.parser") sig = soup.select_one("#signature") sig.span.extract() assert HTML_SIG_MARKER not in sig.text assert htmlsig.format(path=path) == sig.text.strip() plaintext = fake_filewriter.pop()[1] assert plaintext.endswith(EMAIL_SIG_SEP + mailparts[-1]) @pytest.mark.converter @pytest.mark.sig def test_converter_signature_handling_htmlsig_with_image( self, const1, fake_filewriter, monkeypatch, test_png ): path = pathlib.Path(const1) mailparts = ( "This is the mail body", f"{EMAIL_SIG_SEP}", f"{HTML_SIG_MARKER}{path}\n", "This is the plain-text version", ) htmlsig = ( "HTML Signature from {path} with image\n" f'\n' ) def filereader_fn(path): return f'
{htmlsig.format(path=path)}
' def mdwn_fn(t): return t.upper() with monkeypatch.context() as m: m.setattr( markdown.Markdown, "convert", lambda s, t: mdwn_fn(t) ) convert_markdown_to_html( "".join(mailparts), path, filewriter_fn=fake_filewriter, filereader_fn=filereader_fn, ) assert fake_filewriter.pop()[0].suffix == ".png" soup = bs4.BeautifulSoup(fake_filewriter.pop()[1], "html.parser") assert soup.img.attrs["src"].startswith("cid:") @pytest.mark.converter @pytest.mark.sig def test_converter_signature_handling_textsig_with_image( self, const1, fake_filewriter, test_png ): mailparts = ( "This is the mail body", f"{EMAIL_SIG_SEP}", "This is the plain-text version with image\n", f"![Inline]({test_png})", ) tree = convert_markdown_to_html "".join(mailparts), pathlib.Path(const1), filewriter_fn=fake_filewriter, ) assert tree.subtype == "relative" assert tree.children[0].subtype == "alternative" assert tree.children[1].subtype == "png" written = fake_filewriter.pop() assert tree.children[1].path == written[0] assert written[1] == request.urlopen(test_png).read() def test_converter_attribution_to_admonition(self, fake_filewriter): except ImportError: pass