import os
import re
import json

lanes_dir = "/home/donghyeon/Documents/LLM Wiki/docs/superpowers/specs/2026-05-27-branch-notes-audit/lanes"
lane_files = sorted(os.listdir(lanes_dir))

all_extracted = []

for lf in lane_files:
    if not lf.endswith(".md"):
        continue
    
    path = os.path.join(lanes_dir, lf)
    with open(path, "r", encoding="utf-8") as f:
        content = f.read()
        
    print(f"Parsing {lf}...")
    
    # We want to find each Finding section in the file.
    # A finding section typically starts with a header like:
    # "### L1-F01: ..." or "#### Finding 4.1.1: ..."
    # Let's search for headers or blocks that contain "Severity:" or "심각도:"
    # and "Source file:", "Source quote:", etc.
    
    # Let's split the file by headers to isolate findings.
    sections = re.split(r"\n(###+ [^\n]+)\n", content)
    
    current_header = None
    for i, part in enumerate(sections):
        if i == 0:
            continue
        if i % 2 == 1:
            current_header = part
        else:
            section_content = part
            if "Severity" in section_content or "심각도" in section_content:
                # This is a finding!
                severity = "Medium"
                for sev in ["Critical", "High", "Medium", "Low"]:
                    if re.search(r"(?:Severity|심각도)\s*:\s*" + sev, section_content, re.IGNORECASE):
                        severity = sev
                        break
                        
                source_file = "Unknown"
                sf_match = re.search(r"(?:Source file|Source|소스 파일)\s*:\s*`?([^`\n\r]+)`?", section_content)
                if sf_match:
                    source_file = sf_match.group(1).strip()
                    
                finding_id = "Unknown"
                id_match = re.search(r"(L\d+-F\d+|Finding\s+\d+\.\d+\.\d+|F-\d+)", current_header)
                if id_match:
                    finding_id = id_match.group(1).strip()
                else:
                    id_match_in_body = re.search(r"(L\d+-F\d+)", section_content)
                    if id_match_in_body:
                        finding_id = id_match_in_body.group(1).strip()
                
                title = current_header.replace("#", "").strip()
                # Clean up title
                title = re.sub(r"^(L\d+-F\d+|Finding\s+\d+\.\d+\.\d+):\s*", "", title)
                
                all_extracted.append({
                    "lane_file": lf,
                    "finding_id": finding_id,
                    "title": title,
                    "severity": severity,
                    "source_file": source_file,
                    "header": current_header
                })

print(f"\nExtracted {len(all_extracted)} findings in detail!")
with open("extracted_findings.json", "w", encoding="utf-8") as out_j:
    json.dump(all_extracted, out_j, indent=2, ensure_ascii=False)
print("Saved to extracted_findings.json")