# - 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
"--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(
+ "--related-to-html-only",
+ action="store_true",
+ help="Make related content be sibling to HTML parts only"
)
parser.add_argument(
)
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",
)
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()
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 ] ########################################################
def __str__(self):
return f"<multipart/{self.subtype}> 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:
):
relparts = []
for path, info in images.items():
+ if path.startswith("cid:"):
+ continue
+
data = request.urlopen(path)
mimetype = data.headers["Content-Type"]
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
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
+ origtext, extensions=extensions, extension_configs=extension_configs
)
filewriter_fn(draftpath, origtext, encoding="utf-8")
"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"
)
htmlpart = Part("text", "html", htmlpath, "HTML version")
- altpart = Multipart(
- "alternative", [textpart, htmlpart], "Group of alternative content"
- )
-
imgparts = collect_inline_images(
images, tempdir=tempdir, filewriter_fn=filewriter_fn
)
- if imgparts:
+
+ 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(
- "relative", [altpart] + imgparts, "Group of related content"
+ "alternative", [textpart, htmlpart], "Group of alternative content"
)
+
else:
- return altpart
+ # 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=[],
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:
cmd_f,
*,
extensions=None,
+ cssfile=None,
converter=convert_markdown_to_html,
+ related_to_html_only=True,
only_build=False,
tempdir=None,
debug_commands=False,
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,
+ 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, *, 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
# 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("<first-entry>")
cmds.push("<update-encoding>")
+
+ # 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"<attach-file>{item.path}<enter>")
cmds.push("<toggle-unlink><toggle-disposition>")
+ # 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:
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"<jump>{tag}<enter><tag-entry>")
+
if item.subtype == "alternative":
cmds.push("<group-alternatives>")
elif item.subtype in ("relative", "related"):
cmds.push("<group-related>")
elif item.subtype == "multilingual":
cmds.push("<group-multilingual>")
+ 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"<edit-description>{KILL_LINE}{item.desc}<enter>")
- # 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("<tag-entry>")
+ 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
pathlib.Path(args.MAILDRAFT),
cmd_f,
extensions=args.extensions,
+ cssfile=args.css_file,
+ related_to_html_only=args.related_to_html_only,
only_build=args.only_build,
tempdir=args.tempdir,
debug_commands=args.debug_commands,
assert lines[5] in lines_out[7]
@pytest.fixture
- def basic_mime_tree(self):
+ def mime_tree_related_to_alternative(self):
return Multipart(
"relative",
children=[
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, debugprint):
+ items.append((item, len(ancestry)))
- 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[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(
- self, basic_mime_tree
+ self, mime_tree_related_to_alternative
):
items = []
- def visitor_fn(item, stack, debugprint):
+ def visitor_fn(item, ancestry, 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
@pytest.fixture
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,
+ related_to_html_only,
+ extensions,
+ tempdir,
+ ):
return Part("text", "plain", draftpath, orig=True)
do_massage(
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, basic_mime_tree, capsys
+ self, string_io, const1, mime_tree_related_to_alternative, capsys
):
- def converter(drafttext, draftpath, extensions, tempdir):
- return basic_mime_tree
+ def converter(
+ drafttext,
+ draftpath,
+ cssfile,
+ related_to_html_only,
+ extensions,
+ tempdir,
+ ):
+ return mime_tree_related_to_alternative
do_massage(
draft_f=string_io,
)
captured = capsys.readouterr()
- lines = captured.out.splitlines()[4:]
- assert "send-message" in lines.pop(0)
- assert "Related" in lines.pop(0)
- assert "group-related" in lines.pop(0)
- assert "tag-entry" in lines.pop(0)
- assert "Logo" in lines.pop(0)
- assert "content-id" in lines.pop(0)
- assert "toggle-unlink" in lines.pop(0)
- assert "logo.png" in lines.pop(0)
- assert "tag-entry" in lines.pop(0)
- assert "Alternative" in lines.pop(0)
- assert "group-alternatives" in lines.pop(0)
- assert "tag-entry" in lines.pop(0)
- assert "HTML" in lines.pop(0)
- assert "toggle-unlink" in lines.pop(0)
- assert "part.html" in lines.pop(0)
- assert "tag-entry" in lines.pop(0)
- assert "Plain" in lines.pop(0)
- assert "update-encoding" in lines.pop(0)
- assert len(lines) == 2
+ 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 "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()
+ 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):
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, markdown_non_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
text = f"![inline base64 image]({test_png})"
path = pathlib.Path(const1)
tree = convert_markdown_to_html(
- text, path, filewriter_fn=fake_filewriter
+ 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
):
assert relparts[0].cid == const1
assert relparts[0].desc.endswith(const2)
+ def test_apply_stylesheet(self):
+ if _PYNLINER:
+ html = "<p>Hello, world!</p>"
+ 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'<div class="{_CODEHILITE_CLASS}">'
+ "<pre>def foo():\n return</pre></div>"
+ )
+ 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 "HTML" in lines.pop()
+ assert "logo.png" in lines.pop()
+ assert "toggle-unlink" in lines.pop()
+ assert "content-id" 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
+
except ImportError:
pass