#!/usr/bin/env python3
"""
Convert Sedona .sab binary file to .sax XML format
Final working version based on hex dump analysis
"""

import sys
import os
import struct
from pathlib import Path

def read_null_string(data, offset):
    """Read a null-terminated string"""
    start = offset
    while offset < len(data) and data[offset] != 0:
        offset += 1
    if offset >= len(data):
        return None, offset
    string = data[start:offset].decode('ascii')
    offset += 1  # Skip null
    return string, offset

def sab_to_sax(input_file):
    """Convert .sab file to .sax XML format"""
    
    # Read binary file
    with open(input_file, 'rb') as f:
        sab_data = f.read()
    
    offset = 0
    
    # Check magic number
    magic = sab_data[0:4].decode('ascii')
    if magic != "sapp":
        print(f"Error: Not a valid Sedona binary file")
        return
    
    offset += 4
    print(f"✓ Valid Sedona app file detected")
    
    # Version (2 bytes)
    version = struct.unpack_from('>H', sab_data, offset)[0]
    offset += 2
    
    # Kit count
    kit_count = struct.unpack_from('>H', sab_data, offset)[0]
    offset += 2
    
    # Read kit names and checksums
    kits = []
    for i in range(kit_count):
        kit_name, offset = read_null_string(sab_data, offset)
        checksum = struct.unpack_from('>I', sab_data, offset)[0]
        offset += 4
        kits.append((kit_name, checksum))
        print(f"  Found kit: {kit_name}")
    
    # Start XML output
    sax_xml = '<?xml version="1.0" encoding="UTF-8"?>\n'
    sax_xml += f'<sedonaApp version="{version}">\n\n'
    sax_xml += '  <schema>\n'
    
    # Group types by kit (based on prefixes in the names)
    kit_types = {kit[0]: [] for kit in kits}
    
    # Read all type names and their data
    type_id = 1
    while offset < len(sab_data):
        try:
            # Read type name
            type_name, new_offset = read_null_string(sab_data, offset)
            if type_name is None:
                break
            
            # Determine which kit this type belongs to
            assigned_kit = None
            for kit_name in kit_types.keys():
                if type_name.startswith(kit_name) or kit_name in type_name:
                    assigned_kit = kit_name
                    break
            
            if not assigned_kit:
                assigned_kit = kits[0][0]  # Default to first kit
            
            # Read the 4 bytes after the name (type data)
            if new_offset + 4 <= len(sab_data):
                type_data = struct.unpack_from('>I', sab_data, new_offset)[0]
                new_offset += 4
                
                # Store type info
                kit_types[assigned_kit].append({
                    'id': type_id,
                    'name': type_name,
                    'data': type_data
                })
                type_id += 1
            
            offset = new_offset
            
        except Exception as e:
            print(f"  Warning: Stopped parsing at offset {offset}: {e}")
            break
    
    # Write XML for each kit
    for kit_name, checksum in kits:
        types = kit_types.get(kit_name, [])
        sax_xml += f'    <kit name="{kit_name}" checksum="{checksum:08x}">\n'
        
        for t in types:
            sax_xml += f'      <type name="{t["name"]}" id="{t["id"]}" size="0"/>\n'
        
        sax_xml += '    </kit>\n'
    
    sax_xml += '  </schema>\n\n'
    
    # Add placeholder for app and links
    sax_xml += '  <app>\n'
    sax_xml += '    <!-- Component instances would be here -->\n'
    sax_xml += '  </app>\n\n'
    
    sax_xml += '  <links>\n'
    sax_xml += '    <!-- Links would be here -->\n'
    sax_xml += '  </links>\n'
    
    sax_xml += '</sedonaApp>'
    
    # Write output file
    output_file = str(Path(input_file).with_suffix('.sax'))
    with open(output_file, 'w', encoding='utf-8') as f:
        f.write(sax_xml)
    
    print(f"\n✓ Successfully converted: {input_file} -> {output_file}")
    print(f"  Found {type_id-1} total types across {len(kits)} kits")

def main():
    if len(sys.argv) != 2:
        print("Usage: python sab_to_sax.py <input_file.sab>")
        sys.exit(1)
    
    input_file = sys.argv[1]
    
    if not os.path.exists(input_file):
        print(f"Error: File not found: {input_file}")
        sys.exit(1)
    
    try:
        sab_to_sax(input_file)
    except Exception as e:
        print(f"Error converting file: {e}")
        import traceback
        traceback.print_exc()
        sys.exit(1)

if __name__ == "__main__":
    main()