Add conversion for HTML to markdown

apprise/conversion.py

@@ -50,8 +50,7 @@ def convert_between(from_format, to_format, content):
         (NotifyFormat.MARKDOWN, NotifyFormat.HTML): markdown_to_html,
         (NotifyFormat.TEXT, NotifyFormat.HTML): text_to_html,
         (NotifyFormat.HTML, NotifyFormat.TEXT): html_to_text,
-        # For now; use same converter for Markdown support
-        (NotifyFormat.HTML, NotifyFormat.MARKDOWN): html_to_text,
+        (NotifyFormat.HTML, NotifyFormat.MARKDOWN): html_to_markdown,
     }
 
     convert = converters.get((from_format, to_format))
@@ -86,6 +85,17 @@ def html_to_text(content):
     return parser.converted
 
 
+def html_to_markdown(content):
+    """
+    Converts a content from HTML to markdown.
+    """
+
+    parser = HTMLMarkDownConverter()
+    parser.feed(content)
+    parser.close()
+    return parser.converted
+
+
 class HTMLConverter(HTMLParser, object):
     """An HTML to plain text converter tuned for email messages."""
 
@@ -200,3 +210,112 @@ def handle_endtag(self, tag):
 
         if tag in self.BLOCK_TAGS:
             self._result.append(self.BLOCK_END)
+
+
+class HTMLMarkDownConverter(HTMLConverter):
+    """An HTML to markdown converter tuned for email messages."""
+
+    # Escape markdown characters
+    MARKDOWN_ESCAPE = re.compile(r'([\\`*_{}[\]<>()#+\-.!|])',
+                                 re.DOTALL | re.MULTILINE)
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+
+        # Store href value
+        self._link = ""
+
+    def handle_data(self, data, *args, **kwargs):
+        """
+        Store our data if it is not on the ignore list
+        """
+
+        # initialize our previous flag
+        if self._do_store:
+
+            # Tidy our whitespace
+            content = self.WS_TRIM.sub(' ', data)
+            content = self.MARKDOWN_ESCAPE.sub(r'\\\1', content)
+
+            # Add hyperlink
+            if self._link == "":
+                self._result.append(content)
+            else:
+                self._result.append("[" + content + "]" + self._link)
+
+    def handle_starttag(self, tag, attrs):
+        """
+        Process our starting HTML Tag
+        """
+        # Toggle initial states
+        self._do_store = tag not in self.IGNORE_TAGS
+        self._link = ""
+
+        if tag in self.BLOCK_TAGS:
+            self._result.append(self.BLOCK_END)
+
+        if tag == 'li':
+            self._result.append('- ')
+
+        elif tag == 'br':
+            self._result.append('\n')
+
+        elif tag == 'hr':
+            if self._result:
+                self._result[-1] = self._result[-1].rstrip(' ')
+
+            self._result.append('\n---\n')
+
+        elif tag == 'blockquote':
+            self._result.append('> ')
+
+        elif tag == 'h1':
+            self._result.append('# ')
+
+        elif tag == 'h2':
+            self._result.append('## ')
+
+        elif tag == 'h3':
+            self._result.append('### ')
+
+        elif tag == 'h4':
+            self._result.append('#### ')
+
+        elif tag == 'h5':
+            self._result.append('##### ')
+
+        elif tag == 'h6':
+            self._result.append('###### ')
+
+        elif tag in ['strong', 'b']:
+            self._result.append('**')
+
+        elif tag in ['em', 'i']:
+            self._result.append('*')
+
+        elif tag == 'code':
+            self._result.append('`')
+
+        elif tag == 'a':
+            for name, link in attrs:
+                if name == 'href':
+                    self._link = '(' + link + ')'
+
+    def handle_endtag(self, tag):
+        """
+        Edge case handling of open/close tags
+        """
+        self._do_store = True
+        self._link = ""
+
+        if tag in self.BLOCK_TAGS:
+            self._result.append(self.BLOCK_END)
+
+        if tag in ['strong', 'b']:
+            self._result.append('**')
+
+        elif tag in ['em', 'i']:
+            self._result.append('*')
+
+        elif tag == 'code':
+            self._result.append('`')

test/test_conversion.py

@@ -143,6 +143,110 @@ def to_html(body):
         assert to_html(object)
 
 
+def test_conversion_html_to_markdown():
+    """conversion: Test HTML to plain text
+    """
+
+    def to_markdown(body):
+        """
+        A function to simply html conversion tests
+        """
+        return convert_between(NotifyFormat.HTML, NotifyFormat.MARKDOWN, body)
+
+    assert to_markdown("No HTML code here.") == "No HTML code here\."
+
+    clist = to_markdown("<ul><li>Lots and lots</li><li>of lists.</li></ul>")
+    assert "- Lots and lots" in clist
+    assert "- of lists\." in clist
+
+    assert "> To be or not to be\." == to_markdown(
+        "<blockquote>To be or not to be.</blockquote>")
+
+    cspace = to_markdown(
+        "<h2>Fancy heading</h2>"
+        "<p>And a paragraph too.<br>Plus line break.</p>")
+    assert "# Fancy heading" in cspace
+    assert "And a paragraph too\.\nPlus line break\." in cspace
+
+    assert to_markdown(
+        "<style>body { font: 200%; }</style>"
+        "<p>Some obnoxious text here.</p>") == "Some obnoxious text here\."
+
+    assert to_markdown(
+        "<p>line 1</p>"
+        "<p>line 2</p>"
+        "<p>line 3</p>") == "line 1\nline 2\nline 3"
+
+    # Case sensitivity
+    assert to_markdown(
+        "<p>line 1</P>"
+        "<P>line 2</P>"
+        "<P>line 3</P>") == "line 1\nline 2\nline 3"
+
+    # double new lines (testing <br> and </br>)
+    assert to_markdown(
+        "some information<br/><br>and more information") == \
+        "some information\n\nand more information"
+
+    #
+    # Test bad tags
+    #
+
+    # first 2 entries are okay, but last will do as best as it can
+    assert to_markdown(
+        "<p>line 1</>"
+        "<p>line 2</gar>"
+        "<p>line 3>") == "line 1\nline 2\nline 3\>"
+
+    # Make sure we ignore fields that aren't important to us
+    assert to_markdown(
+        "<script>ignore this</script>"
+        "<p>line 1</p>"
+        "Another line without being enclosed") == \
+        "line 1\nAnother line without being enclosed"
+
+    # Test cases when there are no new lines (we're dealing with just inline
+    # entries); an empty entry as well
+    assert to_markdown("<span></span<<span>test</span> "
+                       "<a href='#'>my link</a>") == \
+           "test [my link](#)"
+
+    # </p> missing
+    assert to_markdown("<body><div>line 1 <b>bold</b></div>  "
+                       " <a href='/link'>my link</a>"
+                       "<p>3rd line</body>") == \
+           "line 1 **bold**\n[my link](/link)\n3rd line"
+
+    # <hr/> on it's own
+    assert to_markdown("<hr/>") == "---"
+    assert to_markdown("<hr>") == "---"
+
+    # We need to handle HTML Encodings
+    assert to_markdown("""
+        <html>
+            <title>ignore this entry</title>
+        <body>
+          Let&apos;s handle&nbsp;special html encoding
+          <hr/>
+        </body>
+        """) == "Let's handle special html encoding\n---"
+
+    # If you give nothing, you get nothing in return
+    assert to_markdown("") == ""
+
+    with pytest.raises(TypeError):
+        # Invalid input
+        assert to_markdown(None)
+
+    with pytest.raises(TypeError):
+        # Invalid input
+        assert to_markdown(42)
+
+    with pytest.raises(TypeError):
+        # Invalid input
+        assert to_markdown(object)
+
+
 def test_conversion_text_to():
     """conversion: Test Text to all types
     """