doc_processer/test_word_mathml.py

"""Test Word-compatible MathML generation."""

from app.services.converter import Converter


def test_mathml_word_compatibility():
    """Test that generated MathML is Word-compatible."""
    
    converter = Converter()
    
    print("=" * 80)
    print("Testing Word-Compatible MathML Generation")
    print("=" * 80)
    
    # Test case: Matrix with determinant (the problematic example)
    latex = r"""\left| \begin{array}{cccc} a_{11} & a_{12} & \dots & a_{1n} \\ \vdots & \vdots & & \vdots \\ a_{i1} & 0 & \dots & 0 \\ \vdots & \vdots & & \vdots \\ a_{n1} & a_{n2} & \dots & a_{nn} \end{array} \right|"""
    
    print(f"\nLaTeX: {latex[:80]}...")
    print("\n" + "-" * 80)
    
    # Convert to formats
    result = converter.convert_to_formats(f"$${latex}$$")
    
    if not result.mathml:
        print("✗ No MathML generated")
        return False
    
    mathml = result.mathml
    
    print("Checking Word compatibility features:")
    print("-" * 80)
    
    # Check 1: Display attribute
    if 'display="block"' in mathml:
        print("✓ Has display='block' attribute")
    else:
        print("✗ Missing or wrong display attribute")
        print(f"  Found: {mathml[:100]}...")
    
    # Check 2: No Unicode entities for common symbols
    unicode_issues = []
    problematic_entities = ['+', '…', '⋮', '=', '|']
    for entity in problematic_entities:
        if entity in mathml:
            unicode_issues.append(entity)
    
    if unicode_issues:
        print(f"✗ Contains Unicode entities: {unicode_issues}")
    else:
        print("✓ No problematic Unicode entities")
    
    # Check 3: Uses mfenced for brackets (Word-friendly)
    if '<mfenced' in mathml or '<mo fence="true"' in mathml or 'stretchy="true"' in mathml:
        print("✓ Uses fence elements")
    else:
        print("? No fence elements found (might be OK)")
    
    # Check 4: Has proper namespace
    if 'xmlns="http://www.w3.org/1998/Math/MathML"' in mathml:
        print("✓ Has MathML namespace")
    else:
        print("✗ Missing MathML namespace")
    
    # Show preview
    print("\n" + "-" * 80)
    print("MathML Preview (first 500 chars):")
    print("-" * 80)
    print(mathml[:500])
    if len(mathml) > 500:
        print("...")
    
    print("\n" + "-" * 80)
    print(f"Total length: {len(mathml)} characters")
    
    # Check if this looks like Pandoc-generated MathML
    if 'mfenced' in mathml or 'columnalign' in mathml:
        print("✓ Appears to be Pandoc-generated (good for Word)")
    elif 'stretchy' in mathml and 'fence' in mathml:
        print("✓ Uses standard fence attributes")
    else:
        print("? MathML structure unclear")
    
    return True


def test_simple_formulas():
    """Test simple formulas for Word compatibility."""
    
    converter = Converter()
    
    print("\n" + "=" * 80)
    print("Testing Simple Formulas")
    print("=" * 80)
    
    test_cases = [
        ("Fraction", r"\frac{a}{b}"),
        ("Square root", r"\sqrt{x^2 + y^2}"),
        ("Summation", r"\sum_{i=1}^{n} i"),
        ("Equation", r"E = mc^2"),
        ("Matrix", r"\begin{pmatrix} a & b \\ c & d \end{pmatrix}"),
    ]
    
    all_passed = True
    
    for name, latex in test_cases:
        print(f"\n{name}: ${latex}$")
        
        try:
            result = converter.convert_to_formats(f"${latex}$")
            mathml = result.mathml
            
            # Quick checks
            checks = [
                ('display="block"' in mathml, "display=block"),
                ('&#x0002B;' not in mathml, "no +entity"),
                ('&#x0003D;' not in mathml, "no =entity"),
                ('xmlns=' in mathml, "namespace"),
            ]
            
            status = "✓" if all(check[0] for check in checks) else "✗"
            failed_checks = [check[1] for check in checks if not check[0]]
            
            print(f"  {status} Length: {len(mathml)} chars", end="")
            if failed_checks:
                print(f" | Issues: {', '.join(failed_checks)}")
                all_passed = False
            else:
                print(" | All checks passed")
                
        except Exception as e:
            print(f"  ✗ Error: {e}")
            all_passed = False
    
    return all_passed


def compare_with_reference():
    """Compare our MathML with reference Word-compatible MathML."""
    
    print("\n" + "=" * 80)
    print("Comparison with Reference MathML")
    print("=" * 80)
    
    converter = Converter()
    
    # Simple matrix example
    latex = r"\left| \begin{array}{cc} a & b \\ c & d \end{array} \right|"
    
    result = converter.convert_to_formats(f"$${latex}$$")
    our_mathml = result.mathml
    
    print("\nOur MathML structure:")
    print("-" * 80)
    
    # Analyze structure
    features = {
        "mfenced": "<mfenced" in our_mathml,
        "mo fence": '<mo fence="' in our_mathml or '<mo stretchy="true"' in our_mathml,
        "mtable": "<mtable" in our_mathml,
        "display block": 'display="block"' in our_mathml,
        "unicode entities": any(f"&#x{x};" in our_mathml for x in ["0002B", "0003D", "0007C"]),
    }
    
    print("Features:")
    for feature, present in features.items():
        status = "✓" if present != (feature == "unicode entities") else "✗"
        print(f"  {status} {feature}: {present}")
    
    print(f"\nLength: {len(our_mathml)} characters")
    print(f"Preview:\n{our_mathml[:300]}...")
    
    return not features["unicode entities"]


if __name__ == "__main__":
    print("Word-Compatible MathML Test Suite\n")
    
    try:
        test1 = test_mathml_word_compatibility()
        test2 = test_simple_formulas()
        test3 = compare_with_reference()
        
        print("\n" + "=" * 80)
        print("SUMMARY")
        print("=" * 80)
        
        if test1 and test2 and test3:
            print("✓✓✓ ALL TESTS PASSED ✓✓✓")
            print("\nMathML should be Word-compatible!")
            print("Try copying the mathml output and pasting into Word.")
        else:
            print("✗✗✗ SOME TESTS FAILED ✗✗✗")
            print("\nMathML may not be fully Word-compatible.")
        
        print("=" * 80)
        
    except KeyboardInterrupt:
        print("\n\nTests interrupted")
    except Exception as e:
        print(f"\n\nTest error: {e}")
        import traceback
        traceback.print_exc()