From e31017cfe7b7c24e597a7a8ff26ba9cd8bdf31ad Mon Sep 17 00:00:00 2001
From: liuyuanchuang <yuanchuang_liu@qingsongchou.com>
Date: Wed, 4 Feb 2026 12:45:34 +0800
Subject: [PATCH] fix: add preprocess

---
 app/services/converter.py  |  35 +++++-
 test_array_fix.py          | 102 +++++++++++++++++
 test_omml_preprocessing.py | 218 +++++++++++++++++++++++++++++++++++++
 3 files changed, 354 insertions(+), 1 deletion(-)
 create mode 100644 test_array_fix.py
 create mode 100644 test_omml_preprocessing.py

diff --git a/app/services/converter.py b/app/services/converter.py
index b5ff2ba..04f3d9d 100644
--- a/app/services/converter.py
+++ b/app/services/converter.py
@@ -217,6 +217,9 @@ class Converter:
         This is a separate method due to the performance overhead of OMML conversion,
         which requires creating a temporary DOCX file.
 
+        The formula is preprocessed using the same logic as export_to_file to ensure
+        proper conversion.
+
         Args:
             latex_formula: Pure LaTeX formula (without delimiters like $ or $$).
 
@@ -230,7 +233,37 @@ class Converter:
         if not latex_formula or not latex_formula.strip():
             raise ValueError("LaTeX formula cannot be empty")
 
-        return self._latex_to_omml(latex_formula.strip())
+        # Preprocess formula using the same preprocessing as export
+        preprocessed = self._preprocess_formula_for_omml(latex_formula.strip())
+        
+        return self._latex_to_omml(preprocessed)
+
+    def _preprocess_formula_for_omml(self, latex_formula: str) -> str:
+        """Preprocess LaTeX formula for OMML conversion.
+
+        Applies the same preprocessing steps as preprocess_for_export to ensure
+        consistency. This fixes common issues that cause Pandoc OMML conversion to fail.
+
+        Args:
+            latex_formula: Pure LaTeX formula.
+
+        Returns:
+            Preprocessed LaTeX formula.
+        """
+        # Use the same preprocessing methods as export
+        # 1. Convert matrix environments
+        latex_formula = self._convert_matrix_environments(latex_formula)
+        
+        # 2. Fix array column specifiers (remove spaces)
+        latex_formula = self._fix_array_column_specifiers(latex_formula)
+        
+        # 3. Fix brace spacing
+        latex_formula = self._fix_brace_spacing(latex_formula)
+        
+        # 4. Convert special environments (cases, aligned)
+        latex_formula = self._convert_special_environments(latex_formula)
+        
+        return latex_formula
 
     def _extract_latex_formula(self, text: str) -> str:
         """Extract LaTeX formula from text by removing delimiters.
diff --git a/test_array_fix.py b/test_array_fix.py
new file mode 100644
index 0000000..324239e
--- /dev/null
+++ b/test_array_fix.py
@@ -0,0 +1,102 @@
+"""Test script for array column specifier fix."""
+
+from app.services.converter import Converter
+
+
+def test_array_specifier_fix():
+    """Test that array column specifiers with spaces are fixed."""
+    
+    converter = Converter()
+    
+    # The problematic LaTeX from the error
+    latex_formula = r"""\begin{array}{l} D = \left| \begin{array}{c c c c} a _ {1 1} & a _ {1 2} & \dots & a _ {1 n} \\ \vdots & \vdots & & \vdots \\ a _ {i 1} + 0 + \dots + 0 & 0 + a _ {i 2} + \dots + 0 & \dots & 0 + \dots + 0 + a _ {i n} \\ \vdots & \vdots & & \vdots \\ a _ {n 1} & a _ {n 2} & \dots & a _ {n n} \end{array} \right| \\ = \left| \begin{array}{c c c c} a _ {1 1} & a _ {1 2} & \dots & a _ {1 n} \\ \vdots & \vdots & & \vdots \\ a _ {i 1} & 0 & \dots & 0 \\ \vdots & \vdots & & \vdots \\ a _ {n 1} & a _ {n 2} & \dots & a _ {n n} \end{array} \right| + \left| \begin{array}{c c c c} a _ {1 1} & a _ {1 2} & \dots & a _ {1 n} \\ \vdots & \vdots & & \vdots \\ 0 & a _ {i 2} & \dots & 0 \\ \vdots & \vdots & & \vdots \\ a _ {n 1} & a _ {n 2} & \dots & a _ {n n} \end{array} \right| \\ + \dots + \left| \begin{array}{c c c c} a _ {1 1} & a _ {1 2} & \dots & a _ {1 n} \\ \vdots & \vdots & & \vdots \\ 0 & 0 & \dots & a _ {i n} \\ \vdots & \vdots & & \vdots \\ a _ {n 1} & a _ {n 2} & \dots & a _ {n n} \end{array} \right|, \\ \end{array}"""
+    
+    print("Testing array column specifier fix")
+    print("=" * 80)
+    print(f"\nOriginal LaTeX (first 200 chars):\n{latex_formula[:200]}...")
+    
+    # Test preprocessing
+    print("\n" + "-" * 80)
+    print("Step 1: Preprocessing")
+    preprocessed = converter._preprocess_formula_for_omml(latex_formula)
+    
+    # Check if spaces were removed from array specifiers
+    if "{c c c c}" in preprocessed:
+        print("✗ FAILED: Spaces not removed from array specifiers")
+        print(f"Found: {preprocessed[preprocessed.find('{c c c c}'):preprocessed.find('{c c c c}')+10]}")
+    elif "{cccc}" in preprocessed:
+        print("✓ SUCCESS: Spaces removed from array specifiers")
+        print(f"Changed '{{{\"c c c c\"}}}' → '{{cccc}}'")
+    else:
+        print("? Could not find array specifier in preprocessed output")
+    
+    # Test OMML conversion
+    print("\n" + "-" * 80)
+    print("Step 2: OMML Conversion")
+    try:
+        omml = converter.convert_to_omml(latex_formula)
+        print(f"✓ SUCCESS: OMML conversion completed")
+        print(f"OMML length: {len(omml)} characters")
+        print(f"OMML preview (first 300 chars):\n{omml[:300]}...")
+        
+        # Check if it contains oMath element
+        if "oMath" in omml:
+            print("\n✓ Valid OMML: Contains oMath element")
+        else:
+            print("\n✗ WARNING: OMML might be incomplete (no oMath element found)")
+            
+    except Exception as e:
+        print(f"✗ FAILED: OMML conversion error")
+        print(f"Error: {e}")
+        return False
+    
+    print("\n" + "=" * 80)
+    print("✓ All tests passed!")
+    return True
+
+
+def test_simple_array():
+    """Test with a simpler array example."""
+    
+    converter = Converter()
+    
+    print("\nTesting simple array")
+    print("=" * 80)
+    
+    # Simple array with spaces in column specifier
+    latex_formula = r"\begin{array}{c c c} a & b & c \\ d & e & f \end{array}"
+    
+    print(f"LaTeX: {latex_formula}")
+    
+    try:
+        omml = converter.convert_to_omml(latex_formula)
+        print(f"✓ SUCCESS: Converted to OMML ({len(omml)} chars)")
+        print(f"Preview: {omml[:200]}...")
+        return True
+    except Exception as e:
+        print(f"✗ FAILED: {e}")
+        return False
+
+
+if __name__ == "__main__":
+    print("Array Column Specifier Fix Test Suite\n")
+    
+    try:
+        test1 = test_simple_array()
+        test2 = test_array_specifier_fix()
+        
+        if test1 and test2:
+            print("\n" + "=" * 80)
+            print("✓✓✓ ALL TESTS PASSED ✓✓✓")
+            print("=" * 80)
+        else:
+            print("\n" + "=" * 80)
+            print("✗✗✗ SOME TESTS FAILED ✗✗✗")
+            print("=" * 80)
+            
+    except KeyboardInterrupt:
+        print("\n\nTests interrupted by user")
+    except Exception as e:
+        print(f"\n\nTest suite error: {e}")
+        import traceback
+        traceback.print_exc()
diff --git a/test_omml_preprocessing.py b/test_omml_preprocessing.py
new file mode 100644
index 0000000..b36616c
--- /dev/null
+++ b/test_omml_preprocessing.py
@@ -0,0 +1,218 @@
+"""Comprehensive test for OMML conversion with preprocessing."""
+
+from app.services.converter import Converter
+
+
+def test_case_1_array_with_spaces():
+    """Test: Array with spaces in column specifier (the original issue)."""
+    print("\n" + "=" * 80)
+    print("Test 1: Array with spaces in column specifier")
+    print("=" * 80)
+    
+    converter = Converter()
+    
+    # The problematic LaTeX from the error
+    latex = r"""\begin{array}{l} D = \left| \begin{array}{c c c c} a _ {1 1} & a _ {1 2} & \dots & a _ {1 n} \\ \vdots & \vdots & & \vdots \\ a _ {i 1} + 0 + \dots + 0 & 0 + a _ {i 2} + \dots + 0 & \dots & 0 + \dots + 0 + a _ {i n} \\ \vdots & \vdots & & \vdots \\ a _ {n 1} & a _ {n 2} & \dots & a _ {n n} \end{array} \right| \\ = \left| \begin{array}{c c c c} a _ {1 1} & a _ {1 2} & \dots & a _ {1 n} \\ \vdots & \vdots & & \vdots \\ a _ {i 1} & 0 & \dots & 0 \\ \vdots & \vdots & & \vdots \\ a _ {n 1} & a _ {n 2} & \dots & a _ {n n} \end{array} \right| + \left| \begin{array}{c c c c} a _ {1 1} & a _ {1 2} & \dots & a _ {1 n} \\ \vdots & \vdots & & \vdots \\ 0 & a _ {i 2} & \dots & 0 \\ \vdots & \vdots & & \vdots \\ a _ {n 1} & a _ {n 2} & \dots & a _ {n n} \end{array} \right| \\ + \dots + \left| \begin{array}{c c c c} a _ {1 1} & a _ {1 2} & \dots & a _ {1 n} \\ \vdots & \vdots & & \vdots \\ 0 & 0 & \dots & a _ {i n} \\ \vdots & \vdots & & \vdots \\ a _ {n 1} & a _ {n 2} & \dots & a _ {n n} \end{array} \right|, \\ \end{array}"""
+    
+    print(f"LaTeX length: {len(latex)} chars")
+    print(f"Preview: {latex[:100]}...")
+    
+    try:
+        omml = converter.convert_to_omml(latex)
+        print(f"\n✓ SUCCESS: Converted to OMML")
+        print(f"OMML length: {len(omml)} chars")
+        
+        if "oMath" in omml:
+            print("✓ Valid OMML structure detected")
+        
+        # Check preprocessing worked
+        preprocessed = converter._preprocess_formula_for_omml(latex)
+        if "{c c c c}" not in preprocessed and "{cccc}" in preprocessed:
+            print("✓ Array column specifiers fixed: '{c c c c}' → '{cccc}'")
+        
+        return True
+        
+    except Exception as e:
+        print(f"\n✗ FAILED: {e}")
+        return False
+
+
+def test_case_2_vmatrix():
+    """Test: vmatrix environment conversion."""
+    print("\n" + "=" * 80)
+    print("Test 2: vmatrix environment")
+    print("=" * 80)
+    
+    converter = Converter()
+    
+    latex = r"\begin{vmatrix} a & b \\ c & d \end{vmatrix}"
+    print(f"LaTeX: {latex}")
+    
+    try:
+        omml = converter.convert_to_omml(latex)
+        print(f"✓ SUCCESS: Converted to OMML ({len(omml)} chars)")
+        
+        # Check if vmatrix was converted
+        preprocessed = converter._preprocess_formula_for_omml(latex)
+        if "vmatrix" not in preprocessed and r"\left|" in preprocessed:
+            print("✓ vmatrix converted to \\left| ... \\right|")
+        
+        return True
+        
+    except Exception as e:
+        print(f"✗ FAILED: {e}")
+        return False
+
+
+def test_case_3_cases_environment():
+    """Test: cases environment conversion."""
+    print("\n" + "=" * 80)
+    print("Test 3: cases environment")
+    print("=" * 80)
+    
+    converter = Converter()
+    
+    latex = r"f(x) = \begin{cases} x^2 & x \geq 0 \\ -x & x < 0 \end{cases}"
+    print(f"LaTeX: {latex}")
+    
+    try:
+        omml = converter.convert_to_omml(latex)
+        print(f"✓ SUCCESS: Converted to OMML ({len(omml)} chars)")
+        
+        # Check if cases was converted to array
+        preprocessed = converter._preprocess_formula_for_omml(latex)
+        if "cases" not in preprocessed and "array" in preprocessed:
+            print("✓ cases converted to array environment")
+        
+        return True
+        
+    except Exception as e:
+        print(f"✗ FAILED: {e}")
+        return False
+
+
+def test_case_4_aligned_environment():
+    """Test: aligned environment conversion."""
+    print("\n" + "=" * 80)
+    print("Test 4: aligned environment")
+    print("=" * 80)
+    
+    converter = Converter()
+    
+    latex = r"\begin{aligned} x + y &= 5 \\ 2x - y &= 1 \end{aligned}"
+    print(f"LaTeX: {latex}")
+    
+    try:
+        omml = converter.convert_to_omml(latex)
+        print(f"✓ SUCCESS: Converted to OMML ({len(omml)} chars)")
+        
+        # Check if aligned was converted
+        preprocessed = converter._preprocess_formula_for_omml(latex)
+        if "aligned" not in preprocessed and "array" in preprocessed:
+            print("✓ aligned converted to array environment")
+        if "&" not in preprocessed or preprocessed.count("&") < latex.count("&"):
+            print("✓ Alignment markers removed")
+        
+        return True
+        
+    except Exception as e:
+        print(f"✗ FAILED: {e}")
+        return False
+
+
+def test_case_5_simple_formula():
+    """Test: Simple formula (should work without preprocessing)."""
+    print("\n" + "=" * 80)
+    print("Test 5: Simple formula")
+    print("=" * 80)
+    
+    converter = Converter()
+    
+    latex = r"x = \frac{-b \pm \sqrt{b^2 - 4ac}}{2a}"
+    print(f"LaTeX: {latex}")
+    
+    try:
+        omml = converter.convert_to_omml(latex)
+        print(f"✓ SUCCESS: Converted to OMML ({len(omml)} chars)")
+        return True
+        
+    except Exception as e:
+        print(f"✗ FAILED: {e}")
+        return False
+
+
+def test_case_6_nested_structures():
+    """Test: Nested structures with multiple issues."""
+    print("\n" + "=" * 80)
+    print("Test 6: Nested structures")
+    print("=" * 80)
+    
+    converter = Converter()
+    
+    latex = r"\left\{ \begin{array}{l c} \begin{vmatrix} a & b \\ c & d \end{vmatrix} & = ad - bc \\ f(x) = \begin{cases} 1 & x > 0 \\ 0 & x \leq 0 \end{cases} & \text{step function} \end{array} \right."
+    print(f"LaTeX: {latex}")
+    
+    try:
+        omml = converter.convert_to_omml(latex)
+        print(f"✓ SUCCESS: Converted to OMML ({len(omml)} chars)")
+        
+        preprocessed = converter._preprocess_formula_for_omml(latex)
+        print("\nPreprocessing applied:")
+        if "vmatrix" not in preprocessed:
+            print("  ✓ vmatrix converted")
+        if "cases" not in preprocessed:
+            print("  ✓ cases converted")
+        if "{l c}" not in preprocessed and "{lc}" in preprocessed:
+            print("  ✓ Array specifiers fixed")
+        
+        return True
+        
+    except Exception as e:
+        print(f"✗ FAILED: {e}")
+        return False
+
+
+if __name__ == "__main__":
+    print("=" * 80)
+    print("OMML CONVERSION TEST SUITE")
+    print("Testing preprocessing and conversion")
+    print("=" * 80)
+    
+    results = []
+    
+    try:
+        results.append(("Simple formula", test_case_5_simple_formula()))
+        results.append(("Array with spaces", test_case_1_array_with_spaces()))
+        results.append(("vmatrix", test_case_2_vmatrix()))
+        results.append(("cases", test_case_3_cases_environment()))
+        results.append(("aligned", test_case_4_aligned_environment()))
+        results.append(("Nested structures", test_case_6_nested_structures()))
+        
+        # Summary
+        print("\n" + "=" * 80)
+        print("TEST SUMMARY")
+        print("=" * 80)
+        
+        passed = sum(1 for _, result in results if result)
+        total = len(results)
+        
+        for name, result in results:
+            status = "✓ PASS" if result else "✗ FAIL"
+            print(f"{status}: {name}")
+        
+        print("\n" + "-" * 80)
+        print(f"Total: {passed}/{total} tests passed")
+        
+        if passed == total:
+            print("\n✓✓✓ ALL TESTS PASSED ✓✓✓")
+        else:
+            print(f"\n✗✗✗ {total - passed} TESTS FAILED ✗✗✗")
+        
+        print("=" * 80)
+        
+    except KeyboardInterrupt:
+        print("\n\nTests interrupted by user")
+    except Exception as e:
+        print(f"\n\nTest suite error: {e}")
+        import traceback
+        traceback.print_exc()