Update app/services/orchestrator/build_question_set.py

app/services/orchestrator/build_question_set.py

@@ -4,13 +4,15 @@ import json
 import random
 import os
 
-from app.services.model_a_question_generator import generate_mcqs
-from app.services.model_b_answer_predictor import predict_answer
-from app.services.model_c_explanation_generator import generate_short_explanation
-from app.services.model_d_solution_generator import generate_solution  # FIXED
+# No AI models — local mode
+from app.services.model_d_solution_generator import generate_solution
+
 
 # Correct path to local_questions.json
-DATA_FILE = os.path.join(os.path.dirname(__file__), "..", "..", "data", "local_questions.json")
+DATA_FILE = os.path.join(
+    os.path.dirname(__file__),
+    "..", "..", "data", "local_questions.json"
+)
 DATA_FILE = os.path.abspath(DATA_FILE)
 
 with open(DATA_FILE, "r", encoding="utf-8") as f:
@@ -19,41 +21,36 @@ with open(DATA_FILE, "r", encoding="utf-8") as f:
 
 async def build_question_set(topic: str, num: int = 10):
     """
-    Returns a question set with:
-    - question
-    - options
-    - correct answer (from local or A/B/C/D)
-    - short explanation (optional)
-    - detailed solution (local only)
+    Offline-only question builder.
+    Uses:
+    - local_questions.json
+    - local_solutions.json
     """
 
-    # STEP 1 — Get questions (local fallback or generated)
+    # STEP 1 — Select topic or fallback to empty list
     if topic in LOCAL_DB["topics"]:
         question_pool = LOCAL_DB["topics"][topic]
-        selected = random.sample(question_pool, min(num, len(question_pool)))
     else:
-        # fallback (HF models)
-        raw = await generate_mcqs(topic=topic, num=num)
-        selected = raw
+        return []
+
+    # Random selection
+    selected = random.sample(question_pool, min(num, len(question_pool)))
 
     question_set = []
 
-    # STEP 2 — Build each question item
+    # STEP 2 — Build objects
     for i, q in enumerate(selected):
 
-        question_text = q["question"]
-        options = q["options"]
-        correct = q.get("answer", "")  # might be empty for generated questions
+        # q is a STRING
+        question_text = q
 
-        # STEP 3 — Predict answer if needed (uses model_b or empty)
-        if not correct:
-            predicted = await predict_answer(question_text, options)
-            correct = predicted
+        # Because PDF options were not extracted
+        options = ["A", "B", "C", "D"]
 
-        # STEP 4 — Short explanation (local or empty)
-        short_exp = await generate_short_explanation(question_text, correct)
+        # No answer key available
+        correct = None
 
-        # STEP 5 — Detailed solution (LOCAL ONLY) — FIXED
+        # Load detailed solution from local_solutions.json
         detailed = await generate_solution(i + 1)
 
         question_set.append({
@@ -61,7 +58,7 @@ async def build_question_set(topic: str, num: int = 10):
             "question": question_text,
             "options": options,
             "answer": correct,
-            "short_explanation": short_exp,
+            "short_explanation": "",
             "detailed_solution": detailed
         })