emotion_analyzer.py

"""
Emotion Analyzer for Math Tutor Bot
Detects student emotions from text to provide adaptive tutoring responses
"""

from textblob import TextBlob
import re

class EmotionAnalyzer:
    """
    Analyzes student text input to detect emotional state.
    Uses sentiment analysis + keyword detection for better accuracy.
    """

    def __init__(self):
        # Keywords indicating different emotional states
        self.frustration_keywords = [
            "frustrated", "confusing", "confused", "don't understand",
            "doesn't make sense", "give up", "too hard", "impossible",
            "stuck", "lost", "hate", "stupid", "difficult"
        ]

        self.anxiety_keywords = [
            "worried", "nervous", "anxious", "scared", "afraid",
            "test", "exam", "grade", "fail", "stress", "pressure"
        ]

        self.confidence_keywords = [
            "easy", "got it", "understand", "makes sense", "clear",
            "know", "sure", "confident", "good", "great"
        ]

        self.confusion_keywords = [
            "confused", "don't get", "what does", "how do", "why",
            "explain", "unclear", "lost", "help", "stuck"
        ]

    def analyze(self, text):
        """
        Analyze text and return emotional state with confidence score.

        Returns:
            dict: {
                'emotion': str (frustrated, confused, confident, anxious, neutral),
                'sentiment_score': float (-1 to 1),
                'confidence': float (0 to 1),
                'details': str (explanation)
            }
        """
        text_lower = text.lower()

        # Get sentiment using TextBlob
        blob = TextBlob(text)
        sentiment_score = blob.sentiment.polarity  # -1 (negative) to 1 (positive)

        # Count keyword matches for each emotion
        frustration_count = sum(1 for kw in self.frustration_keywords if kw in text_lower)
        anxiety_count = sum(1 for kw in self.anxiety_keywords if kw in text_lower)
        confidence_count = sum(1 for kw in self.confidence_keywords if kw in text_lower)
        confusion_count = sum(1 for kw in self.confusion_keywords if kw in text_lower)

        # Detect question patterns (often indicate confusion)
        has_question = '?' in text or any(text_lower.startswith(q) for q in ['what', 'how', 'why', 'when', 'where'])

        # Determine primary emotion
        emotion = "neutral"
        confidence = 0.5
        details = "Student appears neutral"

        # Frustration detection (negative sentiment + frustration keywords)
        if frustration_count > 0 and sentiment_score < 0:
            emotion = "frustrated"
            confidence = min(0.9, 0.6 + (frustration_count * 0.15))
            details = f"Detected frustration (keywords: {frustration_count}, sentiment: {sentiment_score:.2f})"

        # Anxiety detection
        elif anxiety_count > 0:
            emotion = "anxious"
            confidence = min(0.9, 0.5 + (anxiety_count * 0.2))
            details = f"Detected anxiety (keywords: {anxiety_count})"

        # Confusion detection (questions + confusion keywords)
        elif confusion_count > 1 or (has_question and confusion_count > 0):
            emotion = "confused"
            confidence = min(0.85, 0.5 + (confusion_count * 0.15))
            details = f"Detected confusion (keywords: {confusion_count}, has question: {has_question})"

        # Confidence detection (positive sentiment + confidence keywords)
        elif confidence_count > 0 and sentiment_score > 0.2:
            emotion = "confident"
            confidence = min(0.9, 0.6 + (confidence_count * 0.15))
            details = f"Detected confidence (keywords: {confidence_count}, sentiment: {sentiment_score:.2f})"

        # Fallback to sentiment-based detection
        elif sentiment_score < -0.3:
            emotion = "frustrated"
            confidence = 0.6
            details = f"Low sentiment detected ({sentiment_score:.2f})"
        elif sentiment_score > 0.5:
            emotion = "confident"
            confidence = 0.65
            details = f"High sentiment detected ({sentiment_score:.2f})"

        return {
            'emotion': emotion,
            'sentiment_score': sentiment_score,
            'confidence': confidence,
            'details': details
        }

    def get_adaptive_prompt(self, emotion, base_prompt):
        """
        Modify the system prompt based on detected emotion.

        Args:
            emotion: str (frustrated, confused, confident, anxious, neutral)
            base_prompt: str (original system prompt)

        Returns:
            str: Modified system prompt
        """
        emotion_prompts = {
            'frustrated': (
                base_prompt + " The student seems frustrated. Be extra patient and encouraging. "
                "Break down concepts into smaller steps. Acknowledge their effort and remind them "
                "that struggle is part of learning. Use simpler language and more examples."
            ),

            'confused': (
                base_prompt + " The student appears confused. Provide clearer explanations with "
                "step-by-step breakdowns. Use analogies and real-world examples. Ask if they "
                "understand each step before moving forward. Be ready to explain the same concept "
                "in different ways."
            ),

            'anxious': (
                base_prompt + " The student seems anxious or worried. Be reassuring and supportive. "
                "Remind them that making mistakes is normal and part of learning. Focus on building "
                "confidence. Celebrate small wins. Reduce pressure by emphasizing understanding "
                "over speed."
            ),

            'confident': (
                base_prompt + " The student appears confident and engaged. You can introduce "
                "slightly more challenging concepts or ask thought-provoking questions. "
                "Encourage them to explain their reasoning. Consider extension problems "
                "or alternative approaches."
            ),

            'neutral': base_prompt
        }

        return emotion_prompts.get(emotion, base_prompt)


# Convenience function for quick testing
if __name__ == "__main__":
    analyzer = EmotionAnalyzer()

    test_cases = [
        "I don't understand this at all, it's so frustrating!",
        "What does this mean? I'm confused about the steps.",
        "I think I got it! That makes sense now.",
        "I'm really worried about the test tomorrow.",
        "Can you help me with this problem?",
    ]

    print("Emotion Analyzer Test Cases")
    print("=" * 60)

    for text in test_cases:
        result = analyzer.analyze(text)
        print(f"\nText: {text}")
        print(f"Emotion: {result['emotion']} (confidence: {result['confidence']:.2f})")
        print(f"Sentiment: {result['sentiment_score']:.2f}")
        print(f"Details: {result['details']}")

    print("\n" + "=" * 60)