Feat: Complete CBT PDF upload, AI parsing, and MinIO storage integration

backend/worker.py

@@ -4,9 +4,12 @@ from celery import Celery
 from celery.schedules import crontab
 from sqlmodel import Session, select
 import json
+import asyncio
 
 from database import engine
 import models
+import storage
+from pdf_parser import parse_pdf_to_questions
 
 REDIS_URL = os.getenv("REDIS_URL", "redis://redis:6379/0")
 
@@ -75,6 +78,76 @@ def generate_daily_study_plans(self):
         print(f"Database error during study plan generation: {exc}")
         raise self.retry(exc=exc)
 
+@celery_app.task(name="process_pdf_test", bind=True, max_retries=1)
+def process_pdf_test(self, test_id_str: str, object_name: str, admin_id_str: str, exam_pattern: str):
+    """
+    Background task to download PDF from MinIO, extract text, 
+    use LLM to generate questions, and attach them to a Test record.
+    """
+    try:
+        print(f"Starting PDF processing for test {test_id_str}")
+        
+        # 1. Download PDF from MinIO
+        pdf_bytes = storage.get_file(object_name)
+        if not pdf_bytes:
+            print("Failed to download PDF from MinIO")
+            return
+            
+        # 2. Extract Questions using LLM (Async)
+        questions_data = asyncio.run(parse_pdf_to_questions(pdf_bytes, exam_pattern))
+        if not questions_data:
+            print("LLM parsing returned empty data")
+            return
+            
+        import uuid
+        test_id = uuid.UUID(test_id_str)
+        admin_id = uuid.UUID(admin_id_str)
+        
+        with Session(engine) as session:
+            test_obj = session.get(models.Test, test_id)
+            if not test_obj:
+                print("Test record not found in DB")
+                return
+
+            question_ids = []
+            for q_data in questions_data:
+                # Create a Question DB object
+                new_q = models.Question(
+                    subject=q_data.get("subject", "Physics"),
+                    topic=q_data.get("topic", "General"),
+                    subtopic=q_data.get("subtopic", "General"),
+                    difficulty=q_data.get("difficulty", "Medium"),
+                    type=q_data.get("type", models.QuestionTypeEnum.MCQ_SINGLE),
+                    question_text=q_data.get("question_text", ""),
+                    options=q_data.get("options", {}),
+                    correct_answer=q_data.get("correct_answer", {}),
+                    solution_text=q_data.get("solution_text", ""),
+                    language=models.MediumEnum.ENGLISH,
+                    exam=models.ExamEnum(exam_pattern),
+                    created_by=admin_id
+                )
+                session.add(new_q)
+                session.flush() # To get the ID
+                question_ids.append(str(new_q.id))
+                
+            # Update Test record with the new sections
+            test_obj.sections = {
+                "Section 1": {
+                    "questions": question_ids,
+                    "time_limit": test_obj.total_time
+                }
+            }
+            # Automatically activate it when done processing
+            test_obj.is_active = True
+            
+            session.commit()
+            print(f"Successfully processed PDF and created {len(question_ids)} questions for test {test_id_str}")
+            return f"Processed {len(question_ids)} questions"
+    except Exception as exc:
+        print(f"Error processing PDF test: {exc}")
+        # Mark test as failed if we had a status field, for now just log it
+        raise self.retry(exc=exc)
+
 # Schedule the tasks using Celery Beat
 celery_app.conf.beat_schedule = {
     'generate-study-plans-every-morning': {