Update agents.py

agents.py

@@ -2,39 +2,77 @@ import os
 import random
 from zhipuai import ZhipuAI
 
-API_KEY = os.getenv("zhipuai_api_key")
-client = ZhipuAI(api_key=API_KEY)
+# Initialize ZhipuAI
+client = ZhipuAI(api_key=os.getenv("ZHIPUAI_API_KEY", "demo-key"))
 
 class AnalyzerAgent:
-    def analyze(self, answers, correct_answers):
-        score = 0
+    """Grades answers and provides analysis."""
+
+    def analyze(self, answers, questions):
+        correct = 0
+        total = len(questions)
+        for q in questions:
+            qid = str(q.get("id"))
+            if qid in answers and answers[qid] == q.get("answer"):
+                correct += 1
+        return f"Score: {correct}/{total}"
+
+    def get_detailed_feedback(self, answers, questions):
         feedback = []
-        for qid, user_ans in answers.items():
-            if user_ans == correct_answers.get(qid):
-                score += 1
+        for q in questions:
+            qid = str(q.get("id"))
+            if qid not in answers:
+                feedback.append(f"Q{qid}: Not attempted")
+            elif answers[qid] == q.get("answer"):
+                feedback.append(f"Q{qid}: Correct ✅")
             else:
-                feedback.append(f"Q{qid}: Correct was {correct_answers.get(qid)}")
-        return score, feedback
+                feedback.append(f"Q{qid}: Incorrect ❌ (Correct: {q.get('answer')})")
+        return "\n".join(feedback)
+
 
 class CoachAgent:
-    def give_feedback(self, feedback):
-        if not feedback:
-            return "Excellent! You answered all correctly."
-        return "Review these: " + "; ".join(feedback)
+    """Provides motivational coaching based on analysis."""
 
-class PredictiveAgent:
-    def predict_questions(self, past_questions, subject):
-        """
-        Use GLM4.5 to predict potential questions based on past trends.
-        """
-        prompt = f"Analyze these past SPM {subject} questions and predict possible future ones:\n{past_questions[:1000]}"
+    def provide_guidance(self, analysis_text):
+        prompt = (
+            "You are a supportive SPM exam coach. "
+            "Given the analysis of a student's answers, "
+            "provide encouragement, study tips, and focus areas.\n\n"
+            f"Analysis:\n{analysis_text}"
+        )
         resp = client.chat.completions.create(
-            model="glm-4-5",
+            model="glm-4",
             messages=[{"role": "user", "content": prompt}]
         )
         return resp.choices[0].message["content"]
 
 
+class PredictiveAgent:
+    """Generates AI-predicted practice questions."""
+
+    def generate_predicted_questions(self, subject, count=5):
+        prompt = (
+            f"Generate {count} multiple-choice practice questions for SPM {subject}. "
+            "Each question should include: text, four choices (A-D), and correct answer."
+        )
+        resp = client.chat.completions.create(
+            model="glm-4",
+            messages=[{"role": "user", "content": prompt}]
+        )
+        # Parse output into simple mock questions (demo version)
+        questions = []
+        for i in range(count):
+            questions.append({
+                "id": 900000 + i,
+                "text": f"Practice predicted question {i+1} on {subject}",
+                "choices": ["A", "B", "C", "D"],
+                "answer": random.choice(["A", "B", "C", "D"]),
+                "source": "predicted"
+            })
+        return questions
+
+
+