Upload folder using huggingface_hub

api/flashcards.py

@@ -54,7 +54,7 @@ async def generate_flashcards(
         )
 
         if not cards_data:
-            raise HTTPException(status_code=500, detail="Failed to generate flashcards")
+            raise HTTPException(status_code=500, detail="AI returned an empty response")
 
         # 2. Save Flashcard Set to DB
         title = request.topic if request.topic else f"Flashcards {len(cards_data)}"
@@ -80,7 +80,15 @@ async def generate_flashcards(
         db.commit()
         db.refresh(db_set)
 
-        return db_set
+        return {
+            "id": db_set.id,
+            "title": db_set.title,
+            "difficulty": db_set.difficulty,
+            "created_at": db_set.created_at,
+            "parent_file_id": db_set.source_id,
+            "parent_file_key": source.s3_key if source else None,
+            "flashcards": db_set.flashcards
+        }
 
     except HTTPException:
         raise
@@ -100,7 +108,18 @@ async def list_flashcard_sets(
         sets = db.query(db_models.FlashcardSet).filter(
             db_models.FlashcardSet.user_id == current_user.id
         ).order_by(db_models.FlashcardSet.created_at.desc()).all()
-        return sets
+        return [
+            {
+                "id": s.id,
+                "title": s.title,
+                "difficulty": s.difficulty,
+                "created_at": s.created_at,
+                "parent_file_id": s.source_id,
+                "parent_file_key": s.source.s3_key if s.source else None,
+                "flashcards": s.flashcards
+            }
+            for s in sets
+        ]
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))
 
@@ -121,7 +140,15 @@ async def get_flashcard_set(
     if not db_set:
         raise HTTPException(status_code=404, detail="Flashcard set not found")
     
-    return db_set
+    return {
+        "id": db_set.id,
+        "title": db_set.title,
+        "difficulty": db_set.difficulty,
+        "created_at": db_set.created_at,
+        "parent_file_id": db_set.source_id,
+        "parent_file_key": db_set.source.s3_key if db_set.source else None,
+        "flashcards": db_set.flashcards
+    }
 
 @router.post("/explain")
 async def explain_flashcard(

api/mindmaps.py

@@ -55,8 +55,12 @@ async def generate_mindmap(
         db.refresh(db_mindmap)
 
         return MindMapResponse(
+            id=db_mindmap.id,
             title=db_mindmap.title,
             mermaid_code=db_mindmap.mermaid_code,
+            parent_file_id=db_mindmap.source_id,
+            parent_file_key=source.s3_key if source else None,
+            created_at=db_mindmap.created_at,
             message="Mind map generated successfully"
         )
 
@@ -81,8 +85,12 @@ async def list_mindmaps(
         
         return [
             MindMapResponse(
+                id=m.id,
                 title=m.title,
                 mermaid_code=m.mermaid_code,
+                parent_file_id=m.source_id,
+                parent_file_key=m.source.s3_key if m.source else None,
+                created_at=m.created_at,
                 message="Retrieved successfully"
             ) for m in mindmaps
         ]

api/podcast.py

@@ -75,6 +75,9 @@ async def generate_podcast(
             ).first()
             if not source:
                 raise HTTPException(status_code=403, detail="Not authorized to access this file")
+            source_id = source.id
+        else:
+            source_id = None
 
         # 2. Generate Script
         script = await podcast_service.generate_script(
@@ -126,7 +129,8 @@ async def generate_podcast(
             s3_key=s3_key,
             s3_url=public_url,
             script=script,
-            user_id=current_user.id
+            user_id=current_user.id,
+            source_id=source_id
         )
         db.add(db_podcast)
         db.commit()
@@ -170,6 +174,8 @@ async def list_podcasts(
                 "public_url": p.s3_url,
                 "private_url": s3_service.get_presigned_url(p.s3_key),
                 "script_preview": (p.script[:200] + "...") if p.script else "",
+                "parent_file_id": p.source_id,
+                "parent_file_key": p.source.s3_key if p.source else None,
                 "created_at": p.created_at
             }
             for p in podcasts

api/quizzes.py

@@ -82,7 +82,15 @@ async def generate_quiz(
         db.commit()
         db.refresh(db_set)
 
-        return db_set
+        return {
+            "id": db_set.id,
+            "title": db_set.title,
+            "difficulty": db_set.difficulty,
+            "created_at": db_set.created_at,
+            "parent_file_id": db_set.source_id,
+            "parent_file_key": source.s3_key if source else None,
+            "questions": db_set.questions
+        }
 
     except HTTPException:
         raise
@@ -102,7 +110,18 @@ async def list_quiz_sets(
         sets = db.query(db_models.QuizSet).filter(
             db_models.QuizSet.user_id == current_user.id
         ).order_by(db_models.QuizSet.created_at.desc()).all()
-        return sets
+        return [
+            {
+                "id": s.id,
+                "title": s.title,
+                "difficulty": s.difficulty,
+                "created_at": s.created_at,
+                "parent_file_id": s.source_id,
+                "parent_file_key": s.source.s3_key if s.source else None,
+                "questions": s.questions
+            }
+            for s in sets
+        ]
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))
 
@@ -123,7 +142,15 @@ async def get_quiz_set(
     if not db_set:
         raise HTTPException(status_code=404, detail="Quiz set not found")
     
-    return db_set
+    return {
+        "id": db_set.id,
+        "title": db_set.title,
+        "difficulty": db_set.difficulty,
+        "created_at": db_set.created_at,
+        "parent_file_id": db_set.source_id,
+        "parent_file_key": db_set.source.s3_key if db_set.source else None,
+        "questions": db_set.questions
+    }
 
 @router.delete("/set/{set_id}")
 async def delete_quiz_set(

api/reports.py

@@ -87,7 +87,11 @@ async def generate_report(
         db.commit()
         db.refresh(db_report)
 
-        return db_report
+        return {
+            **db_report.__dict__,
+            "parent_file_id": db_report.source_id,
+            "parent_file_key": source.s3_key if source else None
+        }
 
     except HTTPException:
         raise
@@ -107,7 +111,18 @@ async def list_reports(
         reports = db.query(db_models.Report).filter(
             db_models.Report.user_id == current_user.id
         ).order_by(db_models.Report.created_at.desc()).all()
-        return reports
+        return [
+            {
+                "id": r.id,
+                "title": r.title,
+                "content": r.content,
+                "format_key": r.format_key,
+                "parent_file_id": r.source_id,
+                "parent_file_key": r.source.s3_key if r.source else None,
+                "created_at": r.created_at
+            }
+            for r in reports
+        ]
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))
 
@@ -128,7 +143,15 @@ async def get_report(
     if not report:
         raise HTTPException(status_code=404, detail="Report not found")
     
-    return report
+    return {
+        "id": report.id,
+        "title": report.title,
+        "content": report.content,
+        "format_key": report.format_key,
+        "parent_file_id": report.source_id,
+        "parent_file_key": report.source.s3_key if report.source else None,
+        "created_at": report.created_at
+    }
 
 @router.delete("/{report_id}")
 async def delete_report(

api/video_generator.py

@@ -67,6 +67,8 @@ async def generate_video_summary(
             "s3_key": db_summary.s3_key,
             "public_url": db_summary.s3_url,
             "private_url": s3_service.get_presigned_url(db_summary.s3_key),
+            "parent_file_id": db_summary.source_id,
+            "parent_file_key": db_summary.source.s3_key if db_summary.source else None,
             "created_at": db_summary.created_at
         }
 
@@ -94,6 +96,8 @@ async def list_video_summaries(
                 "s3_key": s.s3_key,
                 "public_url": s.s3_url,
                 "private_url": s3_service.get_presigned_url(s.s3_key),
+                "parent_file_id": s.source_id,
+                "parent_file_key": s.source.s3_key if s.source else None,
                 "created_at": s.created_at
             }
             for s in summaries

models/db_models.py

@@ -34,6 +34,13 @@ class Source(Base):
     created_at = Column(DateTime(timezone=True), server_default=func.now())
 
     owner = relationship("User", back_populates="sources")
+    podcasts = relationship("Podcast", back_populates="source")
+    flashcard_sets = relationship("FlashcardSet", back_populates="source")
+    mind_maps = relationship("MindMap", back_populates="source")
+    quiz_sets = relationship("QuizSet", back_populates="source")
+    reports = relationship("Report", back_populates="source")
+    video_summaries = relationship("VideoSummary", back_populates="source")
+    rag_documents = relationship("RAGDocument", back_populates="source")
 
 class Podcast(Base):
     __tablename__ = "podcasts"
@@ -44,9 +51,11 @@ class Podcast(Base):
     s3_url = Column(String(1024), nullable=False)
     script = Column(UnicodeText)
     user_id = Column(Integer, ForeignKey("users.id"))
+    source_id = Column(Integer, ForeignKey("sources.id"), nullable=True)
     created_at = Column(DateTime(timezone=True), server_default=func.now())
 
     owner = relationship("User", back_populates="podcasts")
+    source = relationship("Source", back_populates="podcasts")
 
 class FlashcardSet(Base):
     __tablename__ = "flashcard_sets"
@@ -59,6 +68,7 @@ class FlashcardSet(Base):
     created_at = Column(DateTime(timezone=True), server_default=func.now())
 
     owner = relationship("User", back_populates="flashcard_sets")
+    source = relationship("Source", back_populates="flashcard_sets")
     flashcards = relationship("Flashcard", back_populates="flashcard_set", cascade="all, delete-orphan")
 
 class MindMap(Base):
@@ -72,6 +82,7 @@ class MindMap(Base):
     created_at = Column(DateTime(timezone=True), server_default=func.now())
 
     owner = relationship("User", back_populates="mind_maps")
+    source = relationship("Source", back_populates="mind_maps")
 
 class QuizSet(Base):
     __tablename__ = "quiz_sets"
@@ -84,6 +95,7 @@ class QuizSet(Base):
     created_at = Column(DateTime(timezone=True), server_default=func.now())
 
     owner = relationship("User", back_populates="quiz_sets")
+    source = relationship("Source", back_populates="quiz_sets")
     questions = relationship("QuizQuestion", back_populates="quiz_set", cascade="all, delete-orphan")
 
 class QuizQuestion(Base):
@@ -111,6 +123,7 @@ class Report(Base):
     created_at = Column(DateTime(timezone=True), server_default=func.now())
 
     owner = relationship("User", back_populates="reports")
+    source = relationship("Source", back_populates="reports")
 
 class VideoSummary(Base):
     __tablename__ = "video_summaries"
@@ -124,6 +137,7 @@ class VideoSummary(Base):
     created_at = Column(DateTime(timezone=True), server_default=func.now())
 
     owner = relationship("User", back_populates="video_summaries")
+    source = relationship("Source", back_populates="video_summaries")
 
 class Flashcard(Base):
     __tablename__ = "flashcards"
@@ -147,6 +161,7 @@ class RAGDocument(Base):
     created_at = Column(DateTime(timezone=True), server_default=func.now())
 
     owner = relationship("User", back_populates="rag_documents")
+    source = relationship("Source", back_populates="rag_documents")
 
 class ChatMessage(Base):
     __tablename__ = "chat_messages"

models/schemas.py

@@ -84,6 +84,8 @@ class FlashcardSetResponse(BaseModel):
     title: Optional[str]
     difficulty: str
     created_at: datetime
+    parent_file_id: Optional[int] = None
+    parent_file_key: Optional[str] = None
     flashcards: List[FlashcardResponse]
 
     class Config:
@@ -96,8 +98,12 @@ class MindMapGenerateRequest(BaseModel):
     title: Optional[str] = None
 
 class MindMapResponse(BaseModel):
+    id: Optional[int] = None
     title: str
     mermaid_code: str
+    parent_file_id: Optional[int] = None
+    parent_file_key: Optional[str] = None
+    created_at: Optional[datetime] = None
     message: str
 
 # Quiz Schemas
@@ -122,6 +128,8 @@ class QuizSetResponse(BaseModel):
     title: Optional[str]
     difficulty: str
     created_at: datetime
+    parent_file_id: Optional[int] = None
+    parent_file_key: Optional[str] = None
     questions: List[QuizQuestionResponse]
 
     class Config:
@@ -148,6 +156,8 @@ class ReportResponse(BaseModel):
     title: str
     content: str
     format_key: str
+    parent_file_id: Optional[int] = None
+    parent_file_key: Optional[str] = None
     created_at: datetime
 
     class Config:
@@ -167,6 +177,8 @@ class VideoSummaryResponse(BaseModel):
     s3_key: str
     public_url: str
     private_url: Optional[str] = None
+    parent_file_id: Optional[int] = None
+    parent_file_key: Optional[str] = None
     created_at: datetime
 
     class Config:

scripts/migrate_all.py

@@ -0,0 +1,44 @@
+import sys
+import os
+
+# Add the project root to the python path
+sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from sqlalchemy import text
+from core.database import engine
+
+def migrate():
+    print("Starting migration: Ensuring source_id exists in all output tables...")
+    
+    tables = [
+        "podcasts",
+        "video_summaries",
+        "mind_maps",
+        "flashcard_sets",
+        "quiz_sets",
+        "reports"
+    ]
+    
+    try:
+        with engine.connect() as connection:
+            for table in tables:
+                print(f"Processing table: {table}...")
+                
+                # Check and add column/constraint
+                sql = f"""
+                IF NOT EXISTS (SELECT * FROM sys.columns WHERE object_id = OBJECT_ID(N'[dbo].[{table}]') AND name = 'source_id')
+                BEGIN
+                    ALTER TABLE [dbo].[{table}] ADD [source_id] INT NULL;
+                    ALTER TABLE [dbo].[{table}] ADD CONSTRAINT [FK_{table}_sources] 
+                    FOREIGN KEY ([source_id]) REFERENCES [dbo].[sources] ([id]);
+                END
+                """
+                connection.execute(text(sql))
+            
+            connection.commit()
+            print("Migration completed successfully!")
+    except Exception as e:
+        print(f"Error during migration: {e}")
+
+if __name__ == "__main__":
+    migrate()

services/flashcard_service.py

@@ -26,20 +26,18 @@ class FlashcardService:
         language: str = "English"
     ) -> List[Dict[str, str]]:
         """
-        Generates flashcards from either an S3 PDF or direct text input.
+        Generates flashcards from either an S3 PDF or direct text input (Original File-ID Method).
         """
         try:
             system_prompt = get_flashcard_system_prompt(difficulty, quantity, language)
             if topic:
                 system_prompt += get_flashcard_topic_prompt(topic)
 
-            content_to_analyze = ""
-            
             if file_key:
                 # Download PDF from S3
                 tmp = tempfile.NamedTemporaryFile(delete=False, suffix=".pdf")
                 tmp_path = tmp.name
-                tmp.close() # Close handle so other processes can access on Windows
+                tmp.close()
 
                 try:
                     s3_service.s3_client.download_file(
@@ -68,14 +66,13 @@ class FlashcardService:
                     ]
                     
                     response = self.openai_client.chat.completions.create(
-                        model="gpt-4o-mini", # Using 4o-mini for efficiency
+                        model="gpt-4o-mini",
                         messages=messages,
                         temperature=0.7
                     )
                     
                     # Clean up OpenAI file
                     self.openai_client.files.delete(uploaded_file.id)
-                    
                     raw_content = response.choices[0].message.content
                     
                 finally:
@@ -98,18 +95,16 @@ class FlashcardService:
                 raise ValueError("Either file_key or text_input must be provided")
 
             # Parse JSON
-            # Remove markdown code blocks if present
             if "```json" in raw_content:
                 raw_content = raw_content.split("```json")[1].split("```")[0].strip()
             elif "```" in raw_content:
                 raw_content = raw_content.split("```")[1].split("```")[0].strip()
             
-            flashcards = json.loads(raw_content)
-            return flashcards
+            return json.loads(raw_content)
 
         except Exception as e:
-            logger.error(f"Flashcard generation failed: {e}")
-            raise
+            logger.error(f"Flashcard generation failed: {str(e)}")
+            raise e
 
     async def generate_explanation(self, question: str, file_key: Optional[str] = None, language: str = "English") -> str:
         """
@@ -119,7 +114,6 @@ class FlashcardService:
             explanation_prompt = get_flashcard_explanation_prompt(question, language)
             
             if file_key:
-                # Similar logic to generation if PDF context is needed
                 tmp = tempfile.NamedTemporaryFile(delete=False, suffix=".pdf")
                 tmp_path = tmp.name
                 tmp.close()
@@ -158,7 +152,7 @@ class FlashcardService:
                 return response.choices[0].message.content
 
         except Exception as e:
-            logger.error(f"Explanation generation failed: {e}")
-            raise
+            logger.error(f"Explanation generation failed: {str(e)}")
+            raise e
 
 flashcard_service = FlashcardService()