Update app.py

app.py

@@ -1,6 +1,5 @@
 import gradio as gr
 from PIL import Image
-import numpy as np
 from pix2tex.cli import LatexOCR
 import sympy as sp
 from sympy.parsing.latex import parse_latex
@@ -11,7 +10,7 @@ model = LatexOCR()
 def preprocess_handwritten_image(pil_img):
     return pil_img.convert('RGB')
 
-# Tab 1 cleaning (unchanged)
+# Tab 1 cleaning (KEEP UNCHANGED)
 def clean_latex(latex):
     latex = latex.replace('\\ ', '')
     latex = latex.replace('\\\\', '\\')
@@ -52,42 +51,53 @@ def clean_latex(latex):
     latex = latex.replace('E', '2.7183')
     return latex
 
+# ✅ Tab 2 specific cleaner
 def clean_latex2(latex):
     latex = latex.replace('\\ ', '')
     latex = latex.replace('\\\\', '\\')
+    latex = re.sub(r'\\[ \t\n\r\f\v]*', '', latex)
+    latex = latex.replace(r'\times', '*')
+    latex = latex.replace(r'\cdot', '*')
     latex = latex.replace('−', '-').replace('–', '-')
-    latex = re.sub(r'\\times', '*', latex)
-    latex = re.sub(r'\\cdot', '*', latex)
 
-    # Strip LaTeX wrappers
+    # Remove LaTeX wrappers like \mathcal{}, \mathbf{}, etc.
     latex = re.sub(r'\\(text|mathbf|mathrm|mathit|textbf|mathcal|cal)\s*\{([^{}]+)\}', r'\2', latex)
-    latex = re.sub(r'\\[a-zA-Z]+\s*', '', latex)  # leftover LaTeX commands
-    latex = latex.replace('{', '').replace('}', '').replace('&', '')
 
-    # Normalize spacing and equation delimiters
+    # Remove remaining LaTeX commands (like \!\!\! and \,)
+    latex = re.sub(r'\\[a-zA-Z]+', '', latex)
+
+    # Remove nested or empty braces
+    latex = re.sub(r'\{+\}+', '', latex)
+    latex = re.sub(r'\{([^\{\}]*)\}', r'\1', latex)
+
+    # Remove extra commas, semicolons, weird formatting characters
     latex = re.sub(r'[;,]', '\n', latex)
-    latex = re.sub(r'=+', '=', latex)
-    latex = re.sub(r'\s+', ' ', latex).strip()
+    latex = re.sub(r'[^\w\s=+\-*/().]', '', latex)
 
-    # Fix OCR misreads
-    latex = re.sub(r'\bchi\b|\bxi\b|\balpha\b', 'x', latex)
-    latex = re.sub(r'\bbeta\b', 'b', latex)
-    latex = re.sub(r'\bgamma\b', 'y', latex)
+    # Replace multiple spaces with single
+    latex = re.sub(r'\s+', ' ', latex)
 
-    return latex
+    # Normalize equals
+    latex = re.sub(r'(?<![=<>])=(?![=<>])', ' = ', latex)
+
+    # Replace ambiguous variable artifacts
+    replacements = {
+        'chi': 'x', 'xi': 'x', 'alpha': 'x', 'beta': 'b',
+        'gamma': 'y', 'vartheta': '3', 'mathcal': '',
+        'cal': '', 'mathrm': ''
+    }
+    for wrong, right in replacements.items():
+        latex = re.sub(wrong, right, latex)
+
+    return latex.strip()
 
 def solve_polynomial(image):
     try:
-        if isinstance(image, np.ndarray):
-            image = Image.fromarray(image)
         img = preprocess_handwritten_image(image)
         latex_result = model(img)
-
-        if not latex_result or len(latex_result.strip()) < 2 or "ndarray" in latex_result:
+        if not latex_result or len(latex_result.strip()) < 2:
             return "❌ Could not extract valid LaTeX from image.", "", ""
-
         cleaned_latex = clean_latex(latex_result)
-
         try:
             expr = parse_latex(cleaned_latex)
             expr = expr.subs(sp.pi, sp.Float(3.1416)).subs(sp.E, sp.Float(2.7183))
@@ -122,17 +132,18 @@ def solve_polynomial(image):
         return f"❌ Error: {str(e)}", "", ""
 
 def wrapped_solver(img):
-    return solve_polynomial(img)
+    result, cleaned, prompt = solve_polynomial(img)
+    return result, cleaned, prompt
 
+# ✅ FIXED: moved outside of gr.Blocks()
 def solve_system_of_equations(image):
     try:
         img = preprocess_handwritten_image(image)
         raw_latex = model(img)
-
         if not raw_latex or len(raw_latex.strip()) < 3:
             return "❌ Could not extract valid LaTeX."
 
-        # Handle \array environment
+        # Handle array environment
         if r"\begin{array" in raw_latex:
             lines = re.findall(r'\\begin{array}{[^}]+}(.+?)\\end{array}', raw_latex, re.DOTALL)
             if lines:
@@ -146,12 +157,13 @@ def solve_system_of_equations(image):
         cleaned_equations = []
         for eq in equations:
             cleaned = clean_latex2(eq)
+            cleaned = re.sub(r'(?<![=<>])=(?![=<>])', ' = ', cleaned)
             if '=' in cleaned:
                 try:
                     parsed = parse_latex(cleaned)
                     if isinstance(parsed, sp.Equality):
                         cleaned_equations.append(cleaned)
-                except:
+                except Exception:
                     continue
 
         sympy_eqs = []
@@ -179,8 +191,6 @@ def solve_system_of_equations(image):
     except Exception as e:
         return f"❌ Error: {str(e)}"
 
-
-        
 # === UI ===
 with gr.Blocks() as demo:
     with gr.Tab("🖼️ Parse from Image"):