Create app.py

app.py

@@ -0,0 +1,124 @@
+import gradio as gr
+import os
+import time
+import cv2
+import numpy as np
+from PIL import Image
+import tempfile
+import json
+
+# Import consolidated modules
+from ocr_module import MVM2OCREngine
+from reasoning_engine import run_agent_orchestrator
+from verification_service import calculate_symbolic_score
+from consensus_fusion import evaluate_consensus
+from report_module import generate_mvm2_report, export_to_pdf
+from image_enhancing import ImageEnhancer
+
+# Initialize Engines
+ocr_engine = MVM2OCREngine()
+enhancer = ImageEnhancer(sigma=1.2)
+
+def process_mvm2_pipeline(image, auto_enhance):
+    if image is None:
+        return 'Please upload an image.', None, None
+    
+    # 1. Preprocessing
+    if auto_enhance:
+        enhanced_img_np, meta = enhancer.enhance(image)
+        # Save temp image for OCR
+        temp_img_path = os.path.join(tempfile.gettempdir(), 'enhanced_input.png')
+        cv2.imwrite(temp_img_path, enhanced_img_np)
+    else:
+        # Save original PIL image
+        temp_img_path = os.path.join(tempfile.gettempdir(), 'original_input.png')
+        image.save(temp_img_path)
+        meta = {'metrics': {'initial_contrast': 0}}
+
+    # 2. OCR Extraction
+    ocr_results = ocr_engine.process_image(temp_img_path)
+    latex_text = ocr_results['latex_output']
+    ocr_conf = ocr_results['weighted_confidence']
+    
+    if 'No math detected' in latex_text:
+        return f'OCR Failure: {latex_text}', None, None
+
+    # 3. Multi-Agent Reasoning
+    agent_responses = run_agent_orchestrator(latex_text)
+    
+    # 4. Consensus Fusion
+    consensus_result = evaluate_consensus(agent_responses, ocr_confidence=ocr_conf)
+    
+    # 5. Report Generation
+    reports = generate_mvm2_report(consensus_result, latex_text, ocr_conf)
+    md_report = reports['markdown']
+    json_report = json.loads(reports['json'])
+    
+    # 6. Export to PDF
+    pdf_path = os.path.join(tempfile.gettempdir(), f'MVM2_Report_{reports["report_id"]}.pdf')
+    export_to_pdf(json_report, pdf_path)
+    
+    return md_report, pdf_path, latex_text
+
+# Custom CSS for Professional Educational Styling
+custom_css = """
+.gradio-container {
+    font-family: 'Inter', sans-serif;
+}
+.mvm2-header {
+    text-align: center;
+    background: linear-gradient(90deg, #4b6cb7 0%, #182848 100%);
+    color: white;
+    padding: 20px;
+    border-radius: 10px;
+    margin-bottom: 20px;
+}
+.report-area {
+    background-color: #f9f9f9;
+    padding: 15px;
+    border-radius: 8px;
+    border: 1px solid #ddd;
+}
+"""
+
+with gr.Blocks(css=custom_css, title='MVM2: Math Verification & Multi-Signal Consensus') as demo:
+    gr.Markdown(
+        """
+        <div class="mvm2-header">
+            <h1>MVM2: Neuro-Symbolic Math Verification</h1>
+            <p>Adaptive Multi-Signal Consensus for Handwritten Mathematical Equation Verification</p>
+        </div>
+        """
+    )
+    
+    with gr.Row():
+        with gr.Column(scale=1):
+            input_img = gr.Image(type='pil', label='Upload Handwritten Math (Student Notebook)')
+            enhance_toggle = gr.Checkbox(label='Auto-Enhance for Handwritten Math (CLAHE + Gaussian Blur)', value=True)
+            run_btn = gr.Button('Run Multimodal Verification', variant='primary')
+            
+        with gr.Column(scale=2):
+            with gr.Tabs():
+                with gr.TabItem('Explainable Diagnostic Report'):
+                    report_output = gr.Markdown(label='Verification Report', elem_classes='report-area')
+                    download_btn = gr.File(label='Download PDF Report')
+                with gr.TabItem('Raw OCR Extraction'):
+                   ocr_output = gr.Textbox(label='Transcribed LaTeX', interactive=False)
+
+    gr.Markdown(
+        """
+        ### Project MVM2 Capabilities:
+        - Robust OCR: Pix2Text handles complex LaTeX commands and handwritten strokes.
+        - Neuro-Symbolic Fusion: Weighted Score_j formula combines LLM logic with SymPy validation.
+        - Hallucination Detection: Automatically flags agents with low consistency scores (< 0.7).
+        """
+    )
+
+    run_btn.click(
+        fn=process_mvm2_pipeline,
+        inputs=[input_img, enhance_toggle],
+        outputs=[report_output, download_btn, ocr_output]
+    )
+
+if __name__ == "__main__":
+    demo.launch()