X-Git-Url: https://git.madduck.net/etc/neomutt.git/blobdiff_plain/52c7c861405ec54f1ebcdd1d887abb0e4c7b589b..2fa1e4c1aec3affdf63c08a2e1d7005829cab88f:/.config/neomutt/buildmimetree.py?ds=inline diff --git a/.config/neomutt/buildmimetree.py b/.config/neomutt/buildmimetree.py index c0d6364..0052632 100755 --- a/.config/neomutt/buildmimetree.py +++ b/.config/neomutt/buildmimetree.py @@ -6,8 +6,10 @@ # # Configuration: # neomuttrc (needs to be a single line): +# set my_mdwn_extensions="extra,admonition,codehilite,sane_lists,smarty" # macro compose B "\ -# source '$my_confdir/buildmimetree.py setup|'\ +# source '$my_confdir/buildmimetree.py \ +# --tempdir $tempdir --extensions $my_mdwn_extensions|'\ # sourc e \$my_mdwn_postprocess_cmd_file\ # " "Convert message into a modern MIME tree with inline images" # @@ -21,8 +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 @@ -36,7 +38,13 @@ import pathlib import markdown import tempfile import argparse -from collections import namedtuple +import re +import mimetypes +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): @@ -47,67 +55,201 @@ def parse_cli_args(*args, **kwargs): ) ) parser.epilog = ( - "Copyright © 2022 martin f. krafft .\n" + "Copyright © 2023 martin f. krafft .\n" "Released under the MIT licence" ) - subp = parser.add_subparsers(help="Sub-command parsers", dest="mode") - parser_setup = subp.add_parser("setup", help="Setup phase") - parser_massage = subp.add_parser("massage", help="Massaging phase") + parser.add_argument( + "--extensions", + type=str, + default="", + help="Markdown extension to use (comma-separated list)", + ) - parser_setup.add_argument( - "--debug-commands", + if _PYNLINER: + parser.add_argument( + "--css-file", + type=pathlib.Path, + 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="Turn on debug logging of commands generated to stderr", + help="Make related content be sibling to HTML parts only", ) - parser_setup.add_argument( - "--extension", - "-x", - metavar="EXTENSION", - dest="extensions", - nargs="?", - default=[], - action="append", - help="Markdown extension to add to the list of extensions use", + def positive_integer(value): + try: + if int(value) > 0: + return int(value) + + except ValueError: + pass + + raise ValueError(f"Must be a positive integer") + + parser.add_argument( + "--max-number-other-attachments", + type=positive_integer, + help="Make related content be sibling to HTML parts only", ) - parser_massage.add_argument( - "--debug-commands", + parser.add_argument( + "--only-build", action="store_true", - help="Turn on debug logging of commands generated to stderr", + help="Only build, don't send the message", ) - parser_massage.add_argument( - "--debug-walk", + parser.add_argument( + "--tempdir", + type=pathlib.Path, + help="Specify temporary directory to use for attachments", + ) + + parser.add_argument( + "--debug-commands", action="store_true", - help="Turn on debugging to stderr of the MIME tree walk", + help="Turn on debug logging of commands generated to stderr", ) - parser_massage.add_argument( - "--extensions", - metavar="EXTENSIONS", - type=str, - default="", - help="Markdown extension to use (comma-separated list)", + subp = parser.add_subparsers(help="Sub-command parsers", dest="mode") + massage_p = subp.add_parser( + "massage", help="Massaging phase (internal use)" ) - parser_massage.add_argument( + massage_p.add_argument( "--write-commands-to", + "-o", metavar="PATH", dest="cmdpath", + type=pathlib.Path, + required=True, help="Temporary file path to write commands to", ) - parser_massage.add_argument( + massage_p.add_argument( + "--debug-walk", + action="store_true", + help="Turn on debugging to stderr of the MIME tree walk", + ) + + 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) +# [ MARKDOWN WRAPPING ] ####################################################### + + +InlineImageInfo = namedtuple( + "InlineImageInfo", ["cid", "desc"], defaults=[None] +) + + +class InlineImageExtension(Extension): + class RelatedImageInlineProcessor(ImageInlineProcessor): + def __init__(self, re, md, ext): + super().__init__(re, md) + self._ext = ext + + 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._ext.get_cid_for_image(el.attrib) + el.attrib["src"] = f"cid:{cid}" + return el, start, end + + def __init__(self): + super().__init__() + self._images = OrderedDict() + + def extendMarkdown(self, md): + md.registerExtension(self) + inline_image_proc = self.RelatedImageInlineProcessor( + IMAGE_LINK_RE, md, self + ) + md.inlinePatterns.register(inline_image_proc, "image_link", 150) + + def get_cid_for_image(self, attrib): + msgid = make_msgid()[1:-1] + path = attrib["src"] + if path.startswith("/"): + path = f"file://{path}" + self._images[path] = InlineImageInfo( + msgid, attrib.get("title", attrib.get("alt")) + ) + return msgid + + def get_images(self): + return self._images + + +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, extension_configs=extension_configs + ) + htmltext = mdwn.convert(text) + + images = inline_image_handler.get_images() + + def replace_image_with_cid(matchobj): + for m in (matchobj.group(1), f"file://{matchobj.group(1)}"): + if m in images: + return f"(cid:{images[m].cid}" + return matchobj.group(0) + + text = re.sub(r"\(([^)\s]+)", replace_image_with_cid, text) + 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 ] ######################################################## @@ -133,101 +275,176 @@ class Multipart( def __str__(self): return f" children={len(self.children)}" + def __hash__(self): + return hash(str(self.subtype) + "".join(str(self.children))) + -def convert_markdown_to_html(maildraft, *, extensions=None): - draftpath = pathlib.Path(maildraft) +def filewriter_fn(path, content, mode="w", **kwargs): + with open(path, mode, **kwargs) as out_f: + out_f.write(content) + + +def collect_inline_images( + images, *, tempdir=None, filewriter_fn=filewriter_fn +): + relparts = [] + for path, info in images.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") + + relparts.append( + Part( + *mimetype.split("/"), + path, + cid=info.cid, + desc=f"Image: {info.desc}", + ) + ) + + return relparts + + +def convert_markdown_to_html( + origtext, + draftpath, + *, + related_to_html_only=False, + 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, extension_configs=extension_configs + ) + + filewriter_fn(draftpath, origtext, encoding="utf-8") textpart = Part( "text", "plain", draftpath, "Plain-text version", orig=True ) - with open(draftpath, "r", encoding="utf-8") as textmarkdown: - text = textmarkdown.read() - - mdwn = markdown.Markdown(extensions=extensions) - html = mdwn.convert(text) + htmltext = apply_styling(htmltext, cssfile) htmlpath = draftpath.with_suffix(".html") + filewriter_fn( + htmlpath, htmltext, encoding="utf-8", errors="xmlcharrefreplace" + ) htmlpart = Part("text", "html", htmlpath, "HTML version") - with open( - htmlpath, "w", encoding="utf-8", errors="xmlcharrefreplace" - ) as texthtml: - texthtml.write(html) - - logopart = Part( - "image", - "png", - "/usr/share/doc/neomutt/logo/neomutt-256.png", - "Logo", - "neomutt-256.png", + imgparts = collect_inline_images( + images, tempdir=tempdir, filewriter_fn=filewriter_fn ) - return Multipart( - "relative", - [ - Multipart( - "alternative", - [textpart, htmlpart], - "Group of alternative content", - ), - logopart, - ], - "Group of related content", - ) + 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 + 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=[], + descendents=[], visitor_fn=visitor_fn or self._visitor_fn, ) - def _walk(self, node, *, stack, 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, 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, + descendents=descendents, visitor_fn=visitor_fn, ) - self.debugprint("| " * len(stack)) - assert stack.pop() == node + assert ancestry.pop() == node + sibling_descendents = descendents + descendents.extend(node.children) else: - self.debugprint(f"{lead}{node}") + self.debugprint(f"{lead}├{node}") + sibling_descendents = descendents + + if False and ancestry: + self.debugprint(lead[:-1] + " │") if visitor_fn: - visitor_fn(node, stack, debugprint=self.debugprint) + visitor_fn( + node, ancestry, sibling_descendents, debugprint=self.debugprint + ) def debugprint(self, s, **kwargs): if self._debug: @@ -282,17 +499,18 @@ class MuttCommands: def do_setup( - extensions=None, *, out_f=sys.stdout, temppath=None, debug_commands=False + *, + out_f=sys.stdout, + temppath=None, + tempdir=None, + debug_commands=False, ): - extensions = extensions or [] temppath = temppath or pathlib.Path( - tempfile.mkstemp(prefix="muttmdwn-")[1] + tempfile.mkstemp(prefix="muttmdwn-", dir=tempdir)[1] ) cmds = MuttCommands(out_f, debug=debug_commands) - editor = f"{sys.argv[0]} massage --write-commands-to {temppath}" - if extensions: - editor = f'{editor} --extensions {",".join(extensions)}' + editor = f"{' '.join(sys.argv)} massage --write-commands-to {temppath}" cmds.cmd('set my_editor="$editor"') cmds.cmd('set my_edit_headers="$edit_headers"') @@ -304,11 +522,17 @@ def do_setup( def do_massage( - maildraft, + draft_f, + draftpath, cmd_f, *, extensions=None, + cssfile=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, ): @@ -331,59 +555,131 @@ def do_massage( cmds.flush() extensions = extensions.split(",") if extensions else [] - tree = converter(maildraft, extensions=extensions) + tree = converter( + draft_f.read(), + draftpath, + cssfile=cssfile, + related_to_html_only=related_to_html_only, + 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, 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. - # If the part is not an original part, i.e. doesn't already - # exist, we must first add it. - if not item.orig: + # 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("") - if item.cid: - cmds.push(f"\\Ca\\Ck{item.cid}") - # If the item (including the original) comes with a - # description, then we might just as well update the NeoMutt + # 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.desc: - cmds.push(f"\\Ca\\Ck{item.desc}") + 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(f"") + + # 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(f"") 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: + 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 == "relative": + 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" + ) - # Again, if there is a description, we might just as well: - if item.desc: - cmds.push(f"\\Ca\\Ck{item.desc}") + state["pos"] -= n_tags - 1 + state["parts"] += 1 else: # We should never get here - assert not "is valid part" - - # 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("") + 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 @@ -392,6 +688,9 @@ def do_massage( # 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: @@ -408,15 +707,26 @@ def do_massage( if __name__ == "__main__": args = parse_cli_args() - if args.mode == "setup": - do_setup(args.extensions, debug_commands=args.debug_commands) + if args.mode is None: + do_setup( + tempdir=args.tempdir, + debug_commands=args.debug_commands, + ) elif args.mode == "massage": - with open(args.cmdpath, "w") as cmd_f: + with open(args.MAILDRAFT, "r") as draft_f, open( + args.cmdpath, "w" + ) as cmd_f: do_massage( + draft_f, args.MAILDRAFT, cmd_f, extensions=args.extensions, + cssfile=args.css_file, + 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, ) @@ -426,6 +736,7 @@ if __name__ == "__main__": try: import pytest + from io import StringIO class Tests: @pytest.fixture @@ -483,14 +794,20 @@ try: assert lines[5] in lines_out[7] @pytest.fixture - def basic_mime_tree(self): + def mime_tree_related_to_alternative(self): return Multipart( - "related", + "relative", children=[ Multipart( "alternative", children=[ - Part("text", "plain", "part.txt", desc="Plain"), + Part( + "text", + "plain", + "part.txt", + desc="Plain", + orig=True, + ), Part("text", "html", "part.html", desc="HTML"), ], desc="Alternative", @@ -502,70 +819,529 @@ try: desc="Related", ) - def test_MIMETreeDFWalker_depth_first_walk(self, basic_mime_tree): + @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", + ) + + def test_MIMETreeDFWalker_depth_first_walk( + self, mime_tree_related_to_alternative + ): mimetree = MIMETreeDFWalker() items = [] - def visitor_fn(item, stack, debugprint): - items.append((item, len(stack))) + def visitor_fn(item, ancestry, descendents, debugprint): + items.append((item, len(ancestry), len(descendents))) - mimetree.walk(basic_mime_tree, visitor_fn=visitor_fn) + 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[4][0].subtype == "related" + assert items[3][2] == 2 + assert items[4][0].subtype == "relative" assert items[4][1] == 0 + assert items[4][2] == 4 - 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, descendents, 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( - self, basic_mime_tree + self, mime_tree_related_to_alternative ): items = [] - def visitor_fn(item, stack, debugprint): + def visitor_fn(item, ancestry, descendents, debugprint): items.append(item) mimetree = MIMETreeDFWalker(visitor_fn=visitor_fn) - mimetree.walk(basic_mime_tree) + mimetree.walk(mime_tree_related_to_alternative) assert len(items) == 5 - def test_do_setup_no_extensions(self, const1, capsys): - "Assert basics about the setup command output" - do_setup(temppath=const1, out_f=sys.stdout) - captout = capsys.readouterr() - lines = captout.out.splitlines() - assert lines[2].endswith(f'{const1}"') - assert lines[4].endswith(const1) - assert "first-entry" in lines[-1] - assert "edit-file" in lines[-1] - - def test_do_setup_extensions(self, const1, const2, capsys): - "Assert that extensions are passed to editor" - do_setup( - temppath=const1, extensions=[const2, const1], out_f=sys.stdout + @pytest.fixture + def string_io(self, const1, text=None): + return StringIO(text or const1) + + def test_do_massage_basic(self, const1, string_io, capsys): + def converter( + drafttext, + draftpath, + cssfile, + 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) + + def test_do_massage_fulltree( + self, string_io, const1, mime_tree_related_to_alternative, capsys + ): + def converter( + drafttext, + draftpath, + cssfile, + 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, ) - captout = capsys.readouterr() - lines = captout.out.splitlines() - # assert comma-separated list of extensions passed - assert lines[2].endswith(f'{const2},{const1}"') - assert lines[4].endswith(const1) + + 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}" + + 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) + assert ( + path.with_suffix(".html"), + markdown_non_converter(None, const1), + ) == fake_filewriter.pop(0) + + 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 = 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 + + 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 = markdown_with_inline_image_support(text) + assert images[imgpath].desc == const1 + + def test_markdown_inline_image_processor_alt_to_desc(self, const1): + imgpath = "file:///path/to/image.png" + text = f"![{const1}]({imgpath})" + text, html, images = markdown_with_inline_image_support(text) + assert images[imgpath].desc == const1 + + 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 = markdown_with_inline_image_support(text) + assert images[imgpath].desc == const2 + + def test_markdown_inline_image_not_external(self): + imgpath = "https://path/to/image.png" + text = f"![inline image]({imgpath})" + text, html, images = markdown_with_inline_image_support(text) + + assert 'src="cid:' not in html + assert "](cid:" not in text + assert len(images) == 0 + + def test_markdown_inline_image_local_file(self): + imgpath = "/path/to/image.png" + text = f"![inline image]({imgpath})" + text, html, images = markdown_with_inline_image_support(text) + + for k, v in images.items(): + assert k == f"file://{imgpath}" + break + + @pytest.fixture + def test_png(self): + return ( + "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAE" + "AAAABCAAAAAA6fptVAAAACklEQVQI12P4DwABAQEAG7buVgAA" + ) + + def test_markdown_inline_image_processor_base64(self, test_png): + text = f"![1px white inlined]({test_png})" + text, html, images = markdown_with_inline_image_support(text) + + assert 'src="cid:' in html + assert "](cid:" in text + assert len(images) == 1 + assert test_png in images + + 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() + + 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() + + 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 + ): + 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) + + 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 + + def test_mime_tree_relative_within_alternative( + self, string_io, const1, capsys, mime_tree_related_to_html + ): + def converter( + drafttext, + draftpath, + cssfile, + 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 + + def test_mime_tree_nested_trees_does_not_break_positioning( + self, string_io, const1, capsys + ): + def converter( + drafttext, + draftpath, + cssfile, + 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 not "logo.png" 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() except ImportError: pass