underlines + cleanup

md2mu.py

@@ -1,49 +1,146 @@
 import mistune
+from mistune import Markdown
+from mistune.core import BlockState
+from mistune.util import strip_end
+from mistune.renderers._list import render_list
+from mistune.renderers.markdown import MarkdownRenderer
+from typing import Dict, Any
+from textwrap import indent
 import argparse
+import re
 
-class Markdown2Micron(mistune.HTMLRenderer):
-    def text(self, text) -> str:
-        return text
-    def emphasis(self, text: str) -> str:
-        return '`*' + text + '`*'
-    def strong(self, text: str) -> str:
-        return '`!' + text + '`!'
-    def codespan(self, text: str) -> str:
-        return '`=' + text + '`='
-    def heading(self, text, level, **attrs) -> str:
+UNDERLINED = r'\b_{1,3}(?=[^\s_])'
+UNDERLINED_END_RE = {
+    '_': re.compile(r'(?:(?<!\\)(?:\\\\)*\\_|[^\s_])_(?!_)\b'),
+    '__': re.compile(r'(?:(?<!\\)(?:\\\\)*\\_|[^\s_])__(?!_)\b'),
+    '___': re.compile(r'(?:(?<!\\)(?:\\\\)*\\_|[^\s_])___(?!_)\b'),
+}
+
+def parse_underlined(inline, m, state):
+    print(inline, m, state)
+    text = m.group(2)
+    print('rrr')
+    state.append_token({'type': 'underlined', 'raw': text})
+    return m.end()
+
+def parse_underlined(self, m, state) -> int:
+        print(state)
+        pos = m.end()
+        marker = m.group(0)
+        mlen = len(marker)
+        _end_re = UNDERLINED_END_RE[marker]
+        m1 = _end_re.search(state.src, pos)
+        if not m1:
+            state.append_token({'type': 'text', 'raw': marker})
+            return pos
+        end_pos = m1.end()
+        text = state.src[pos:end_pos-mlen]
+        prec_pos = self.precedence_scan(m, state, end_pos)
+        if prec_pos:
+            return prec_pos
+        new_state = state.copy()
+        new_state.src = text
+        new_state.in_underlined = True
+        state.append_token({
+            'type': 'underlined',
+            'children': self.render(new_state),
+        })
+        return end_pos
+
+def render_underlined(self, token, state) -> str:
+        return '`_' + self.render_children(token, state) + '`_'
+
+class Markdown2Micron(MarkdownRenderer):
+    NAME = 'micron'
+
+    def __call__(self, tokens, state: BlockState):
+        out = self.render_tokens(tokens, state)
+        # special handle for line breaks
+        out += '\n\n'.join(self.render_referrences(state)) + '\n'
+        return strip_end(out)
+    
+    def render_children(self, token, state: BlockState):
+        children = token['children']
+        return self.render_tokens(children, state)
+
+    def text(self, token: Dict[str, Any], state: BlockState) -> str:
+        return token['raw']
+    
+    def emphasis(self, token: Dict[str, Any], state: BlockState) -> str:
+        return '`*' + self.render_children(token, state) + '`*'
+
+    def strong(self, token: Dict[str, Any], state: BlockState) -> str:
+        return '`!' + self.render_children(token, state) + '`!'
+    
+    def link(self, token: Dict[str, Any], state: BlockState) -> str:
+        label = token.get('label')
+        text = self.render_children(token, state)
+        out = '`[' + text + '`'
+        if label:
+            return out + '`[' + label + '`'
+        attrs = token['attrs']
+        url = attrs['url']
+        if text == url:
+            return '`[' + text + '`'
+        elif 'mailto:' + text == url:
+            return '`[' + text + '`'
+        out += url
+        return out + ']'
+    
+    def image(self, token: Dict[str, Any], state: BlockState) -> str:
+        return self.link(token, state)
+
+    def codespan(self, token: Dict[str, Any], state: BlockState) -> str:
+        return '`=' + token['raw'] + '`='
+
+    def linebreak(self, token: Dict[str, Any], state: BlockState) -> str:
+        return '  \n'
+    
+    def softbreak(self, token: Dict[str, Any], state: BlockState) -> str:
+        return '\n'
+    
+    def blank_line(self, token: Dict[str, Any], state: BlockState) -> str:
+        return ''
+    
+    def inline_html(self, token: Dict[str, Any], state: BlockState) -> str:
+        return ''
+    
+    def paragraph(self, token: Dict[str, Any], state: BlockState) -> str:
+        text = self.render_children(token, state)
+        return text + '\n\n'
+
+    def heading(self, token: Dict[str, Any], state: BlockState) -> str:
+        level = token['attrs']['level']
         if level > 3:
             level = 3
-        return '>' * level + text + '\n'
-    def link(self, text: str, url: str, title=None) -> str:
-        s = self.safe_url(url)
-        if text:
-            s = text + '`' + s
-        return f'`[{s}]'
-    def image(self, text: str, url: str, title=None) -> str:
-        s = self.safe_url(url)
-        if title:
-            s = title + '`' + s
-        return f'`[{s}]'
-    def blank_line(self) -> str:
-        return '' + '\n'
-    def linebreak(self) -> str:
-        return '\n'
-    def softbreak(self) -> str:
-        return '\n'
-    def inline_html(self, html: str) -> str:
-        return '`=\n' + html + '`=\n'
-    def thematic_break(self):
-        return '-' + '\n'
-    def paragraph(self, text):
-        return text + '\n'
-    def block_quote(self, text: str) -> str:
-        return '>>>>' + text
-    def list(self, text: str, ordered: bool, **attrs) -> str:
-        return text + '\n'
-    def list_item(self, text: str) -> str:
-        return '+ ' + text + '\n'
-    def block_code(self, code: str, info=None) -> str:
-        return '`=\n' + code + '`=\n'
+        marker = '>' * level
+        text = self.render_children(token, state)
+        return marker + ' ' + text + '\n\n'
+    def thematic_break(self, token: Dict[str, Any], state: BlockState) -> str:
+        return '-\n\n'
+    
+    def block_text(self, token: Dict[str, Any], state: BlockState) -> str:
+        return self.render_children(token, state) + '\n'
+    
+    def block_code(self, token: Dict[str, Any], state: BlockState) -> str:
+        code = token['raw']
+        if code and code[-1] != '\n':
+            code += '\n'
+        marker = '`='
+        return marker + '\n' + code + marker + '\n\n'
+    
+    def block_quote(self, token: Dict[str, Any], state: BlockState) -> str:
+        text = indent(self.render_children(token, state), '>>>>')
+        return text + '\n\n'
+    
+    def block_html(self, token: Dict[str, Any], state: BlockState) -> str:
+        return ''
+    
+    def block_error(self, token: Dict[str, Any], state: BlockState) -> str:
+        return ''
+    
+    def list(self, token: Dict[str, Any], state: BlockState) -> str:
+        return render_list(self, token, state)
 
 def main():
 
@@ -55,11 +152,13 @@ def main():
 
     with open(args.md_file, 'r') as mdf:
         md_str = mdf.read()
-    m2μ = Markdown2Micron()
-    md2micron = mistune.create_markdown(renderer=m2μ)
+    m2μr = Markdown2Micron()
+    m2μ = Markdown(renderer=m2μr)
+    m2μ.inline.register('underlined', UNDERLINED, parse_underlined, before='emphasis')
+    m2μ.renderer.register('underlined', render_underlined)
 
     with open(args.mu_file, 'w') as muf:
-        md_str = muf.write(md2micron(md_str))
+        md_str = muf.write(m2μ(md_str))
 
 if __name__ == "__main__":
 	main()

micron.mu

@@ -1,11 +1,16 @@
->h1 Heading 8-)
->>h2 Heading
->>>h3 Heading
->>>h4 Heading
->>>h5 Heading
->>>h6 Heading
+> h1 Heading 8-)
 
->>Horizontal Rules
+>> h2 Heading
+
+>>> h3 Heading
+
+>>> h4 Heading
+
+>>> h5 Heading
+
+>>> h6 Heading
+
+>> Horizontal Rules
 
 -
 
@@ -13,43 +18,51 @@
 
 -
 
->>Emphasis
+>> Emphasis
 
 `!This is bold text`!
-`!This is bold text`!
-`*This is italic text`*
+`_This is bold text`_
 `*This is italic text`*
+`_This is italic text`_
 ~~Strikethrough~~
 
->>Blockquotes
+>> Blockquotes
 
 >>>>Blockquotes can also be nested...
->>>>by using additional greater-than signs right next to each other...
->>>>or with spaces between arrows.
 
->>Lists
+>>>>>>>>by using additional greater-than signs right next to each other...
 
->>>Unordered
+>>>>>>>>>>>>or with spaces between arrows.
+
+
+
+
+
+
+
+>> Lists
+
+>>> Unordered
 
 + Create a list by starting a line with `=+`=, `=-`=, or `=*`=
 
-+ Marker character change forces new list start:
+- Marker character change forces new list start:
 
-+ Ac tristique libero volutpat at
+* Ac tristique libero volutpat at
 
 + Facilisis in pretium nisl aliquet
 
-+ Nulla volutpat aliquam velit
+- Nulla volutpat aliquam velit
 
 + Very easy
 
->>>Ordered
+>>> Ordered
 
-+ Lorem ipsum dolor sit amet
-+ Consectetur adipiscing elit
-+ Integer molestie lorem at massa
+1. Lorem ipsum dolor sit amet
+2. Consectetur adipiscing elit
+3. Integer molestie lorem at massa
 
->>Code
+>> Code
 
 Inline `=code`=
 
@@ -59,7 +72,9 @@ Indented code
 // Some comments
 line 1 of code
 line 2 of code
-line 3 of code`=
+line 3 of code
+`=
+
 Block code "fences"
 
 `=
@@ -76,13 +91,13 @@ var foo = function (bar) {
 console.log(foo(5));
 `=
 
->>Links
+>> Links
 
 `[link text`http://dev.nodeca.com]
 
 `[link with title`http://nodeca.github.io/pica/demo/]
 
->>Images
+>> Images
 
-`[https://octodex.github.com/images/minion.png]
-`[The Stormtroopocat`https://octodex.github.com/images/stormtroopocat.jpg]
+`[Minion`https://octodex.github.com/images/minion.png]
+`[Stormtroopocat`https://octodex.github.com/images/stormtroopocat.jpg]