import re
import json
import uuid

def make_span(text, marks=None):
    return {
        "_key": uuid.uuid4().hex[:8],
        "_type": "span",
        "marks": marks or [],
        "text": text
    }

def make_block(text, style="normal", list_item=None, level=None):
    # Parse bold and italic marks if present
    spans = []
    # simple span for now
    spans.append(make_span(text))
    
    block = {
        "_key": uuid.uuid4().hex[:8],
        "_type": "block",
        "style": style,
        "markDefs": [],
        "children": spans
    }
    if list_item:
        block["listItem"] = list_item
        block["level"] = level or 1
    return block

def parse_markdown(md_path):
    with open(md_path, "r", encoding="utf-8") as f:
        content = f.read()

    lines = content.split("\n")
    blocks = []
    in_schema = False
    in_table = False
    
    for line in lines:
        stripped = line.strip()
        
        # Skip empty lines
        if not stripped:
            continue
            
        # Skip metadata header if at top
        if stripped.startswith("Tiêu đề SEO:") or stripped.startswith("Mô tả meta:") or stripped.startswith("Từ khóa chính:"):
            continue
            
        # Detect schema blocks
        if "<script type=\"application/ld+json\">" in stripped:
            in_schema = True
            continue
        if "</script>" in stripped:
            in_schema = False
            continue
        if in_schema:
            continue
            
        # Detect markdown images
        img_match = re.match(r'!\[(.*?)\]\((.*?)\)', stripped)
        if img_match:
            alt = img_match.group(1)
            src = img_match.group(2)
            blocks.append({
                "_type": "image_placeholder",
                "alt": alt,
                "src": src
            })
            continue

        # Skip comment metadata or caption lines right after image
        if stripped.startswith("<!-- Metadata:") or (stripped.startswith("*") and stripped.endswith("*") and len(stripped) < 120):
            continue

        # Headings
        if stripped.startswith("# "):
            blocks.append(make_block(stripped[2:].strip(), "h1"))
        elif stripped.startswith("## "):
            blocks.append(make_block(stripped[3:].strip(), "h2"))
        elif stripped.startswith("### "):
            blocks.append(make_block(stripped[4:].strip(), "h3"))
        elif stripped.startswith("#### "):
            blocks.append(make_block(stripped[5:].strip(), "h4"))
        elif stripped.startswith("* ") or stripped.startswith("- "):
            blocks.append(make_block(stripped[2:].strip(), "normal", list_item="bullet"))
        elif re.match(r'^\d+\.\s', stripped):
            item_text = re.sub(r'^\d+\.\s', '', stripped)
            blocks.append(make_block(item_text, "normal", list_item="number"))
        elif stripped.startswith("> "):
            blocks.append(make_block(stripped[2:].strip(), "blockquote"))
        elif stripped.startswith("|"):
            # Table row -> convert to clean text block
            cells = [c.strip() for c in stripped.split("|")[1:-1]]
            if not all(c.startswith("-") or not c for c in cells): # skip separator
                blocks.append(make_block(" | ".join(cells), "normal"))
        else:
            blocks.append(make_block(stripped, "normal"))

    return blocks

blocks = parse_markdown("/opt/ai-os/products/ceo/projects/blog/vwg-content/tai-sao-chon-viet-nam/draft_vi_v1.md")
print(f"Parsed {len(blocks)} blocks successfully!")