feat: complete pragativeer v2 frontend and backend, docker setup, and cbt engine

backend/predictor.py

@@ -0,0 +1,76 @@
+from sqlmodel import Session, select
+from uuid import UUID
+import models
+
+def calculate_readiness_score(user_id: UUID, session: Session) -> dict:
+    """
+    Calculates the Exam Readiness Score (0-100) based on PRD A-05 parameters.
+    Weights: Accuracy (40%), Speed (20%), Syllabus Coverage (20%), 
+    Consistency (10%), Mistake Recovery (10%).
+    """
+    user = session.get(models.User, user_id)
+    if not user:
+        return {}
+
+    attempts = session.exec(
+        select(models.Attempt).where(models.Attempt.user_id == user_id)
+    ).all()
+    
+    if not attempts:
+        return {"score": 0, "status": "Not enough data"}
+
+    # 1. Accuracy (40% Weight)
+    avg_accuracy = sum(a.accuracy for a in attempts) / len(attempts)
+    w_accuracy = (avg_accuracy / 100) * 40
+
+    # 2. Speed (20% Weight) - Assuming benchmark is 120s per question
+    time_per_q = [ (a.time_taken / a.total_questions) for a in attempts if a.total_questions > 0 and a.time_taken ]
+    avg_time = sum(time_per_q) / len(time_per_q) if time_per_q else 120
+    speed_factor = min(1.0, 120 / max(avg_time, 1)) # Cap at 1.0
+    w_speed = speed_factor * 20
+
+    # 3. Consistency (10% Weight)
+    streak = min(user.streak_count, 30) # Max consistency score hit at 30 days
+    w_consistency = (streak / 30) * 10
+
+    # 4. Mistake Recovery (10% Weight)
+    mistakes = session.exec(select(models.Mistake).where(models.Mistake.user_id == user_id)).all()
+    mastered_count = sum(1 for m in mistakes if m.is_mastered)
+    total_mistakes = max(len(mistakes), 1)
+    w_recovery = (mastered_count / total_mistakes) * 10
+
+    # 5. Syllabus Coverage (20% Weight) 
+    # For local sandbox, we will estimate based on total questions attempted
+    total_q_attempted = sum(a.total_questions for a in attempts)
+    coverage_factor = min(total_q_attempted / 500, 1.0) # Assume 500 questions = decent baseline coverage
+    w_coverage = coverage_factor * 20
+
+    # Total Calculation
+    total_score = round(w_accuracy + w_speed + w_consistency + w_recovery + w_coverage)
+
+    # Calculate College Prediction based on PRD A-07
+    prediction = "Keep Practicing"
+    target_type = "State Level"
+    
+    if total_score >= 85:
+        prediction = "IIT Bombay / IIT Delhi within reach!"
+        target_type = "Reach"
+    elif total_score >= 70:
+        prediction = "NIT Trichy / NIT Surathkal on track."
+        target_type = "Target"
+    elif total_score >= 50:
+        prediction = "Top GFTIs / State Engineering Colleges."
+        target_type = "Safety"
+
+    return {
+        "score": total_score,
+        "breakdown": {
+            "accuracy": round(w_accuracy, 1),
+            "speed": round(w_speed, 1),
+            "consistency": round(w_consistency, 1),
+            "recovery": round(w_recovery, 1),
+            "coverage": round(w_coverage, 1)
+        },
+        "college_prediction": prediction,
+        "prediction_category": target_type
+    }