_archive/temp/decode_manifest.py

"""
Decode manifest.bin protobuf from onemodel — extract detector config parameters.
This is a raw protobuf that we decode manually (wire format).
"""
import struct

def decode_varint(data, pos):
    result = 0
    shift = 0
    while pos < len(data):
        b = data[pos]
        result |= (b & 0x7f) << shift
        pos += 1
        if not (b & 0x80):
            break
        shift += 7
    return result, pos

def decode_protobuf(data, depth=0, prefix=""):
    pos = 0
    indent = "  " * depth
    while pos < len(data):
        if pos >= len(data):
            break
        
        tag_wire, new_pos = decode_varint(data, pos)
        if new_pos == pos:
            break
        
        field_num = tag_wire >> 3
        wire_type = tag_wire & 0x7
        pos = new_pos
        
        if wire_type == 0:  # Varint
            val, pos = decode_varint(data, pos)
            print(f"{indent}Field {field_num} (varint): {val}")
        
        elif wire_type == 1:  # 64-bit
            if pos + 8 > len(data):
                break
            val_d = struct.unpack('<d', data[pos:pos+8])[0]
            print(f"{indent}Field {field_num} (64-bit/double): {val_d}")
            pos += 8
        
        elif wire_type == 2:  # Length-delimited
            length, pos = decode_varint(data, pos)
            if pos + length > len(data):
                break
            payload = data[pos:pos+length]
            pos += length
            
            # Try to interpret as string
            try:
                s = payload.decode('utf-8')
                if all(32 <= ord(c) < 127 or c in '\n\r\t' for c in s):
                    if len(s) < 300:
                        print(f"{indent}Field {field_num} (string): \"{s}\"")
                    else:
                        print(f"{indent}Field {field_num} (string, {len(s)} chars): \"{s[:100]}...\"")
                    continue
            except:
                pass
            
            # Try as sub-message
            if len(payload) > 2:
                # Heuristic: if first byte looks like a valid tag
                first_tag = payload[0]
                if 0 < first_tag < 128 and (first_tag & 0x7) <= 5:
                    print(f"{indent}Field {field_num} (message, {len(payload)} bytes):")
                    try:
                        decode_protobuf(payload, depth + 1, prefix + f".{field_num}")
                    except:
                        # Not a valid protobuf, dump hex
                        hex_str = payload[:50].hex()
                        print(f"{indent}  [raw hex: {hex_str}...]")
                else:
                    hex_str = payload[:50].hex()
                    print(f"{indent}Field {field_num} (bytes, {len(payload)}): {hex_str}")
            else:
                hex_str = payload.hex()
                print(f"{indent}Field {field_num} (bytes, {len(payload)}): {hex_str}")
        
        elif wire_type == 5:  # 32-bit
            if pos + 4 > len(data):
                break
            val_f = struct.unpack('<f', data[pos:pos+4])[0]
            val_i = struct.unpack('<I', data[pos:pos+4])[0]
            print(f"{indent}Field {field_num} (32-bit/float): {val_f}  (int: {val_i})")
            pos += 4
        
        else:
            print(f"{indent}Field {field_num} (wire_type={wire_type}): UNKNOWN")
            break

with open('oneocr_extracted/config_data/manifest.bin', 'rb') as f:
    data = f.read()

print(f"Manifest size: {len(data)} bytes")
print("="*60)
print("DECODED PROTOBUF")
print("="*60)
decode_protobuf(data)