buildmimetree.py: implement inline image handling

diff --git a/.config/neomutt/buildmimetree.py b/.config/neomutt/buildmimetree.py
index d033bda2f9a1e7f2ae7ede1e4406514de62f5771..c16e519db3ab3e137e04a915c16be9e0ac96df61 100755 (executable)
--- a/.config/neomutt/buildmimetree.py
+++ b/.config/neomutt/buildmimetree.py
@@ -36,7 +36,13 @@ import pathlib
 import markdown
 import tempfile
 import argparse
 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):
 
 
 def parse_cli_args(*args, **kwargs):


@@ -114,6 +120,74 @@ def parse_cli_args(*args, **kwargs):
     return parser.parse_args(*args, **kwargs)
 
 
     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):
+    inline_image_handler = InlineImageExtension()
+    extensions = extensions or []
+    extensions.append(inline_image_handler)
+    mdwn = markdown.Markdown(extensions=extensions)
+    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
+
+

 # [ PARTS GENERATION ] ########################################################
 
 
 # [ PARTS GENERATION ] ########################################################
 
 


@@ -140,51 +214,65 @@ class Multipart(
         return f"<multipart/{self.subtype}> children={len(self.children)}"
 
 
         return f"<multipart/{self.subtype}> children={len(self.children)}"
 
 

+def filewriter_fn(path, content, mode="w", **kwargs):
+    with open(path, mode, **kwargs) as out_f:
+        out_f.write(content)

 
 

-def convert_markdown_to_html(
-    origtext, draftpath, *, filewriter_fn=None, extensions=None
+
+def collect_inline_images(
+    images, *, tempdir=None, filewriter_fn=filewriter_fn

 ):
 ):

-    mdwn = markdown.Markdown(extensions=extensions)
+    relparts = []
+    for path, info in images.items():
+        data = request.urlopen(path)
+
+        mimetype = data.headers["Content-Type"]
+        ext = mimetypes.guess_extension(mimetype)
+        tempfilename = tempfile.mkstemp(prefix="img", suffix=ext, dir=tempdir)[
+            1
+        ]
+        path = pathlib.Path(tempfilename)
+
+        filewriter_fn(path, data.read(), "w+b")
+
+        relparts.append(
+            Part(
+                *mimetype.split("/"), path, cid=info.cid, desc=info.desc
+            )
+        )
+
+    return relparts

 
 

-    if not filewriter_fn:

 
 

-        def filewriter_fn(path, content, mode="w", **kwargs):
-            with open(path, mode, **kwargs) as out_f:
-                out_f.write(content)
+def convert_markdown_to_html(
+    origtext, draftpath, *, filewriter_fn=filewriter_fn, extensions=None
+):
+    origtext, htmltext, images = markdown_with_inline_image_support(
+        origtext, extensions=extensions
+    )

 
     filewriter_fn(draftpath, origtext, encoding="utf-8")
     textpart = Part(
         "text", "plain", draftpath, "Plain-text version", orig=True
     )
 
 
     filewriter_fn(draftpath, origtext, encoding="utf-8")
     textpart = Part(
         "text", "plain", draftpath, "Plain-text version", orig=True
     )
 

-    htmltext = mdwn.convert(origtext)
-

     htmlpath = draftpath.with_suffix(".html")
     filewriter_fn(
         htmlpath, htmltext, encoding="utf-8", errors="xmlcharrefreplace"
     )
     htmlpart = Part("text", "html", htmlpath, "HTML version")
 
     htmlpath = draftpath.with_suffix(".html")
     filewriter_fn(
         htmlpath, htmltext, encoding="utf-8", errors="xmlcharrefreplace"
     )
     htmlpart = Part("text", "html", htmlpath, "HTML version")
 

-    logopart = Part(
-        "image",
-        "png",
-        "/usr/share/doc/neomutt/logo/neomutt-256.png",
-        "Logo",
-        "neomutt-256.png",
+    altpart = Multipart(
+        "alternative", [textpart, htmlpart], "Group of alternative content"

     )
 
     )
 

-    return Multipart(
-        "relative",
-        [
-            Multipart(
-                "alternative",
-                [textpart, htmlpart],
-                "Group of alternative content",
-            ),
-            logopart,
-        ],
-        "Group of related content",
-    )
+    imgparts = collect_inline_images(images, filewriter_fn=filewriter_fn)
+    if imgparts:
+        return Multipart(
+            "relative", [altpart] + imgparts, "Group of related content"
+        )
+    else:
+        return altpart

 
 
 class MIMETreeDFWalker:
 
 
 class MIMETreeDFWalker:


@@ -636,23 +724,174 @@ try:
             )
 
             captured = capsys.readouterr()
             )
 
             captured = capsys.readouterr()

-            lines = captured.out.splitlines()[4:][::-1]
-            assert "Related" in lines.pop()
-            assert "group-related" in lines.pop()
-            assert "tag-entry" in lines.pop()
-            assert "Logo" in lines.pop()
-            assert "content-id" in lines.pop()
-            assert "toggle-unlink" in lines.pop()
-            assert "logo.png" in lines.pop()
-            assert "tag-entry" in lines.pop()
-            assert "Alternative" in lines.pop()
-            assert "group-alternatives" in lines.pop()
-            assert "tag-entry" in lines.pop()
-            assert "HTML" in lines.pop()
-            assert "toggle-unlink" in lines.pop()
-            assert "part.html" in lines.pop()
-            assert "tag-entry" in lines.pop()
-            assert "Plain" in lines.pop()
+            lines = captured.out.splitlines()[4:]
+            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
+
+        @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, markdown_non_converter
+        ):
+            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
+
+        def test_markdown_inline_image_processor_base64(self):
+            img = (
+                "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAE"
+                "AAAABCAAAAAA6fptVAAAACklEQVQI12P4DwABAQEAG7buVgAA"
+            )
+            text = f"![1px white inlined]({img})"
+            text, html, images = markdown_with_inline_image_support(text)
+
+            assert 'src="cid:' in html
+            assert "](cid:" in text
+            assert len(images) == 1
+            assert img in images
+
+        def test_converter_tree_inline_image_base64(
+            self, const1, fake_filewriter
+        ):
+            img = (
+                "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAE"
+                "AAAABCAAAAAA6fptVAAAACklEQVQI12P4DwABAQEAG7buVgAA"
+            )
+            text = f"![inline base64 image]({img})"
+            path = pathlib.Path(const1)
+            tree = convert_markdown_to_html(
+                text, path, filewriter_fn=fake_filewriter
+            )
+
+            assert tree.subtype == "relative"
+            assert tree.children[1].subtype == "png"
+            written = fake_filewriter.pop()
+            assert tree.children[1].path == written[0]
+            assert written[1] == request.urlopen(img).read()

 
 except ImportError:
     pass
 
 except ImportError:
     pass