#!/usr/bin/env python3
"""Test script to verify parse_scan fix for encoding issues."""

from typing import Tuple, Optional
import re

SEP_RE = re.compile(r'[\x1D\x1E]')

def parse_scan(raw: str) -> Tuple[Optional[str], Optional[int]]:
    """
    Parse scanned barcode to extract part code and quantity.

    Args:
        raw: Raw barcode string

    Returns:
        Tuple of (part_code, quantity) or (None, None) if parsing fails
    """
    def clean_part_code(code: str) -> str:
        """Clean part code by removing non-ASCII and invalid characters."""
        # Keep only ASCII printable characters (excluding control chars)
        # This removes characters like ¡­ and other encoding artifacts
        cleaned = ''.join(char for char in code if 32 <= ord(char) <= 126)
        return cleaned.strip()

    # Handle JSON-like format
    if raw.startswith('{') and '}' in raw:
        content = raw.strip()[1:-1]
        part = None
        qty = None

        for kv in content.split(','):
            if ':' not in kv:
                continue
            k, v = kv.split(':', 1)
            key = k.strip().upper()
            val = v.strip()

            if key == 'PM':
                part = clean_part_code(val)
            elif key == 'QTY':
                try:
                    qty = int(val)
                except ValueError:
                    pass
        return part, qty

    # Handle separator-based format
    part = None
    qty = None
    fields = SEP_RE.split(raw)

    for f in fields:
        if not f:
            continue
        if f.startswith('30P'):
            part = clean_part_code(f[3:])
        elif f.lower().startswith('1p'):
            part = clean_part_code(f[2:])
        elif f.lower().startswith('q') and f[1:].isdigit():
            qty = int(f[1:])

    return part, qty


# Test with the provided scan data
test_scan = "{pbn:PICK251017100019,on:WM2510170196,pc:C18548292,pm:STHW4-DU-HS24041¡­,qty:150,mc:,cc:1,pdi:180368458,hp:null,wc:JS}"

print("Testing parse_scan with problematic barcode data:")
print(f"Input: {test_scan}")
print()

part, qty = parse_scan(test_scan)

print(f"Parsed part code: '{part}'")
print(f"Parsed quantity: {qty}")
print()

# Verify the fix
expected_part = "STHW4-DU-HS24041"
expected_qty = 150

if part == expected_part and qty == expected_qty:
    print("✅ SUCCESS! Part code is correctly parsed.")
    print(f"   Expected: '{expected_part}'")
    print(f"   Got:      '{part}'")
else:
    print("❌ FAILURE! Parse result doesn't match expected values.")
    print(f"   Expected part: '{expected_part}', Got: '{part}'")
    print(f"   Expected qty: {expected_qty}, Got: {qty}")

# Show character codes for verification
print("\nCharacter analysis of parsed part code:")
for i, char in enumerate(part or ""):
    print(f"  [{i}] '{char}' (ASCII {ord(char)})")