From 8f6096617dbf42fa6a0cc2e94df7e0b7c486491b Mon Sep 17 00:00:00 2001
From: Ian Griffin <migriffin@disroot.org>
Date: Wed, 12 Nov 2025 18:16:46 +0800
Subject: [PATCH] replaced parse_markdown with a separate library

---
 ragger.py | 67 ++-----------------------------------------------------
 1 file changed, 2 insertions(+), 65 deletions(-)

diff --git a/ragger.py b/ragger.py
index 731ef6c..d572684 100755
--- a/ragger.py
+++ b/ragger.py
@@ -8,7 +8,6 @@
 
 import os
 import mimetypes
-import re
 import readline
 
 from argparse import ArgumentParser
@@ -34,7 +33,7 @@ from langgraph.checkpoint.memory import MemorySaver
 from langgraph.graph import START, StateGraph
 from langgraph.graph.message import add_messages
 from sys import stderr
-from termcolor import colored
+from term_color_md import render as md_render
 from typing import NotRequired, Sequence
 from typing_extensions import Annotated, TypedDict
 from urllib.parse import urlparse
@@ -152,7 +151,6 @@ def main():
         ]
     )
 
-
     history_aware_retriever = create_history_aware_retriever(
         llm, retriever, contextualize_q_prompt
     )
@@ -226,7 +224,7 @@ def main():
             print()
             break
 
-        print(colored("A:", "green", attrs=["reverse"]), parse_markdown(app.invoke({"input": question},
+        print(colored("A:", "green", attrs=["reverse"]), md_render(app.invoke({"input": question},
     config=config)["answer"]), end="\n\n")
 
 # We define a dict representing the state of the application.
@@ -237,66 +235,5 @@ class State(TypedDict):
     context: NotRequired[str]
     answer: NotRequired[str]
 
-def parse_markdown(text):
-    lines = text.splitlines()
-    formatted_text = ""
-    in_code_block = False
-
-    for line in lines:
-        # Check for code blocks
-        if line.startswith("```"):
-            in_code_block = not in_code_block
-            continue  # Skip the line with ```
-        elif in_code_block:
-            formatted_text += colored(line + "\n", "green")
-            continue
-
-        # Check for headers
-        if line.startswith("# "):
-            header_text = line.lstrip("#").strip()
-            formatted_text += colored(header_text, "blue", attrs=["bold", "underline"]) + "\n"
-            continue
-
-        if line.startswith("## "):
-            header_text = line.lstrip("#").strip()
-            formatted_text += colored(header_text, "blue", attrs=["bold"]) + "\n"
-            continue
-
-        if line.startswith("### "):
-            header_text = line.lstrip("#").strip()
-            formatted_text += colored(header_text, "cyan", attrs=["bold"]) + "\n"
-            continue
-
-        # Check for blockquotes
-        if line.startswith(">"):
-            quote_text = line.lstrip(">").strip()
-            formatted_text += colored(quote_text, "yellow") + "\n"
-            continue
-
-        # Check for tables (rows separated by "|")
-        if "|" in line:
-            table_row = "\t| ".join(line.split("|")).strip()
-            formatted_text += table_row + "\n"
-            continue
-
-        # Inline formatting for bold, italic, and code (keeping the symbols)
-        # Bold (**text** or __text__)
-        line = re.sub(r"[^\*_](\*\*|__)(.+?)(\*\*|__)[^\*_]", lambda m: colored(m.group(), attrs=["bold"]), line)
-        # Italic (*text* or _text_)
-        line = re.sub(r"[^\*_](\*|_)([^\*_].+?[^\*_])(\*|_)[^\*_]", lambda m: colored(m.group(), attrs=["underline"]), line)
-        # Inline code (`code`)
-        line = re.sub(r"[^\*_](`)(.+?)`[^\*_]", lambda m: colored(m.group() + "`", "green"), line)
-
-        # List items (bullets and numbers)
-        # Bulleted list
-        line = re.sub(r"^(\s*[-*])\s", lambda m: colored(m.group(1), "cyan") + " ", line)
-        # Numbered list
-        line = re.sub(r"^(\s*\d+\.)\s", lambda m: colored(m.group(1), "cyan") + " ", line)
-
-        # Add processed line to formatted text
-        formatted_text += line + "\n"
-
-    return formatted_text
-
 if __name__ == "__main__":
     main()