Update app.py

app.py

@@ -1,4 +1,5 @@
 import os
+import re
 import gradio as gr
 import google.generativeai as genai
 from markdown_pdf import MarkdownPdf, Section
@@ -173,7 +174,58 @@ def create_model():
     except Exception:
         return genai.GenerativeModel("gemini-2.5-flash", generation_config={"temperature": 0})
 
-# ---------- PIPELINE: ALIGN + GRADE ----------
+# ---------- NEW: Pretty math conversion ----------
+_SUPER_MAP = {
+    "0": "⁰", "1": "¹", "2": "²", "3": "³", "4": "⁴",
+    "5": "⁵", "6": "⁶", "7": "⁷", "8": "⁸", "9": "⁹",
+    "+": "⁺", "-": "⁻", "(": "⁽", ")": "⁾",
+    "n": "ⁿ", "i": "ⁱ", "a": "ᵃ", "b": "ᵇ", "c": "ᶜ", "d": "ᵈ", "e": "ᵉ",
+    "o": "ᵒ", "r": "ʳ", "t": "ᵗ", "u": "ᵘ", "v": "ᵛ", "w": "ʷ", "x": "ˣ", "y": "ʸ"
+}
+
+def _to_superscript(s: str) -> str:
+    """Map characters to available Unicode superscripts; fallback to original char if unavailable."""
+    out = []
+    for ch in s:
+        out.append(_SUPER_MAP.get(ch, _SUPER_MAP.get(ch.lower(), ch)))
+    return "".join(out)
+
+def pretty_math(text: str) -> str:
+    """
+    Convert caret-notation exponents to Unicode superscripts.
+
+    Handles:
+      x^2, x^{12}, x^(12), 10^4, (3x10^4)^3, and similar patterns.
+    Only characters with known superscripts are converted (digits, +-(), some letters).
+    """
+
+    if not text:
+        return text
+
+    new = text
+
+    # Convert instances like ^{...}
+    new = re.sub(r'\^\{\s*([^}]+)\s*\}', lambda m: _to_superscript(m.group(1)), new)
+
+    # Convert instances like ^(...)
+    new = re.sub(r'\^\(\s*([^\)]+)\s*\)', lambda m: _to_superscript(m.group(1)), new)
+
+    # Convert caret followed by a simple integer (e.g., ^12)
+    new = re.sub(r'\^([+-]?\d+)', lambda m: _to_superscript(m.group(1)), new)
+
+    # Convert caret followed by single non-space token (e.g., x^n)
+    new = re.sub(r'\^([A-Za-z0-9\+\-\(\)]+)', lambda m: _to_superscript(m.group(1)), new)
+
+    # Replace common scientific notation '3x10^4' -> '3×10⁴' when 'x' is used as multiplication
+    # Only apply when the 'x' sits between digits and '10' (heuristic)
+    new = re.sub(r'(\d)\s*[xX]\s*(10)', r'\1×\2', new)
+
+    # Also compact spaced forms: '3 x 10^4' -> '3×10⁴' (keeps previously superscripted exponent)
+    new = re.sub(r'(\d)\s*×\s*(10)', r'\1×\2', new)
+
+    return new
+
+# -------------------- PIPELINE: ALIGN + GRADE --------------------
 def align_and_grade(qp_file, ms_file, ans_file):
     try:
         # Step 0: Compress if needed
@@ -188,40 +240,44 @@ def align_and_grade(qp_file, ms_file, ans_file):
 
         model = create_model()
 
-        # Step 2: Alignment
+        # Step 2: Alignment (raw)
         resp = model.generate_content([
             PROMPTS["ALIGNMENT_PROMPT"]["content"],
             qp_uploaded,
             ms_uploaded,
             ans_uploaded
         ])
-        aligned_text = getattr(resp, "text", None)
-        if not aligned_text and resp.candidates:
-            aligned_text = resp.candidates[0].content.parts[0].text
+        aligned_text_raw = getattr(resp, "text", None)
+        if not aligned_text_raw and getattr(resp, "candidates", None):
+            aligned_text_raw = resp.candidates[0].content.parts[0].text
 
-        aligned_pdf_path = save_as_pdf(aligned_text, "aligned_qp_ms_as.pdf")
+        # Pretty version for display/PDF (does NOT affect the raw text used for grading)
+        aligned_text_pretty = pretty_math(aligned_text_raw) if aligned_text_raw else aligned_text_raw
+        aligned_pdf_path = save_as_pdf(aligned_text_pretty or "[No aligned text produced]", "aligned_qp_ms_as.pdf")
 
-        # Step 3: Grading
+        # Step 3: Grading (use raw aligned text as input to grader)
         response = model.generate_content([
             PROMPTS["GRADING_PROMPT"]["content"],
-            aligned_text
+            aligned_text_raw or "[No aligned text produced]"
         ])
-        grading = getattr(response, "text", None)
-        if not grading and response.candidates:
-            grading = response.candidates[0].content.parts[0].text
+        grading_raw = getattr(response, "text", None)
+        if not grading_raw and getattr(response, "candidates", None):
+            grading_raw = response.candidates[0].content.parts[0].text
 
-        # Save grading report with student's answer filename
+        # Pretty version of grading output for display/PDF
+        grading_pretty = pretty_math(grading_raw) if grading_raw else grading_raw
         base_name = os.path.splitext(os.path.basename(ans_file))[0]
-        grading_pdf_path = save_as_pdf(grading, f"{base_name}_graded.pdf")
+        grading_pdf_path = save_as_pdf(grading_pretty or "[No grading produced]", f"{base_name}_graded.pdf")
 
-        return aligned_text, aligned_pdf_path, grading, grading_pdf_path
+        # Return pretty/display versions (raws remain unmodified for internal processing)
+        return aligned_text_pretty or "", aligned_pdf_path, grading_pretty or "", grading_pdf_path
 
     except Exception as e:
         return f"❌ Error: {e}", None, None, None
 
 # ---------- GRADIO APP ----------
 with gr.Blocks(title="LeadIB AI Grading (Alignment + Auto-Grading)") as demo:
-    gr.Markdown("## LeadIB AI Grading\nUpload Question Paper, Markscheme, and Student Answer Sheet.\nThe system will align and grade automatically.")
+    gr.Markdown("## LeadIB AI Grading\nUpload Question Paper, Markscheme, and Student Answer Sheet.\nThe system will align and grade automatically. Mathematical exponents (e.g. `x^2`) will be converted to Unicode superscripts (e.g. `x²`) in the displayed text and PDFs.")
 
     with gr.Row():
         qp_file = gr.File(label="Upload Question Paper (PDF)", type="filepath")