From 323816262704dbee1fc1b2aa08204d55d4386b16 Mon Sep 17 00:00:00 2001
From: David Allemang <david.allemang@kitware.com>
Date: Wed, 13 Nov 2024 16:49:01 -0500
Subject: [PATCH] wip python prototype to feel out the parser

---
 notes.md         |   2 +-
 pyparse.py       | 112 +++++++++++++++++++++++++++++++++++++++++++++++
 requirements.txt |   2 +
 3 files changed, 115 insertions(+), 1 deletion(-)
 create mode 100644 pyparse.py
 create mode 100644 requirements.txt

diff --git a/notes.md b/notes.md
index 20bc895..a4de5b6 100644
--- a/notes.md
+++ b/notes.md
@@ -34,4 +34,4 @@ So the first task in the parser should be to parse the block structure.
   - HTML
   - Text
 
-<!--vim: ts=2 sw=2 et linebreak :-->
+<!--vim: ts=2 sw=2 et linebreak :-->
\ No newline at end of file
diff --git a/pyparse.py b/pyparse.py
new file mode 100644
index 0000000..27db082
--- /dev/null
+++ b/pyparse.py
@@ -0,0 +1,112 @@
+import sys
+from enum import Enum, auto
+from idlelib.configdialog import is_int
+from pathlib import Path
+from pprint import pprint
+from typing import Optional
+
+
+class Block:
+    def __init__(self, *parts):
+        self.tag = type(self).__name__
+        self.data = list(parts)
+
+    def extend(self, *parts):
+        self.data.extend(parts)
+
+    def __repr__(self):
+        return f'{self.tag}:: {''.join(self.data)!r}'
+
+
+class Break(Block): pass
+
+
+class ATXHeading(Block): pass
+
+
+class SetextHeading(Block): pass
+
+
+class IndentedChunk(Block): pass
+
+
+class Fence(Block):
+    def __init__(self, meta, *data):
+        super().__init__(*data)
+        self.meta = meta
+        self.complete = False
+
+    def __repr__(self):
+        return f'{self.tag}:{self.meta}:: {''.join(self.data)!r}'
+
+
+class HTML(Block): pass
+
+
+class Definition(Block): pass
+
+
+class Paragraph(Block): pass
+
+
+class Blank(Block): pass
+
+
+def convert(md: str):
+    blocks: list[Block] = []
+
+    cur_fence: Optional[Fence] = None
+
+    def get(idx):
+        try:
+            return blocks[idx]
+        except IndexError:
+            return None
+
+    for line in md.splitlines(keepends=True):
+        if cur_fence:
+            if line.lstrip(' ').startswith('```'):
+                blocks.append(cur_fence)
+                cur_fence = None
+            else:
+                cur_fence.extend(line)
+        else:
+            if line.isspace():
+                if len(blocks) >= 1 and isinstance(blocks[-1], Blank):
+                    blocks[-1].extend(line)
+                else:
+                    blocks.append(Blank(line))
+
+            elif line.startswith('    ') or line.startswith('\t'):
+                if len(blocks) >= 1 and isinstance(blocks[-1], IndentedChunk):
+                    blocks[-1].extend(line)
+                elif len(blocks) >= 2 and isinstance(blocks[-1], Blank) and isinstance(blocks[-2], IndentedChunk):
+                    blocks[-2].extend(*blocks[-1].data, line)
+                    blocks.pop(-1)
+                else:
+                    blocks.append(IndentedChunk(line))
+
+            elif line.lstrip(' ').startswith('```'):
+                meta = line.strip().removeprefix('```')
+                cur_fence = Fence(meta)
+            else:
+                if len(blocks) >= 1 and isinstance(blocks[-1], Paragraph):
+                    blocks[-1].extend(line)
+                else:
+                    blocks.append(Paragraph(line))
+
+    pprint(blocks)
+
+
+def main():
+    for arg in sys.argv[1:]:
+        md = Path(arg).read_text()
+        html = convert(md)
+
+        print('=' * 80)
+        print(html)
+        print('=' * 80)
+
+
+if __name__ == '__main__':
+    main()
diff --git a/requirements.txt b/requirements.txt
new file mode 100644
index 0000000..e7a2eec
--- /dev/null
+++ b/requirements.txt
@@ -0,0 +1,2 @@
+httpx
+parsel