Method process_pdf

backend/simple_pdf_processor.py:66–112

Process a PDF file and store in database

(self, pdf_bytes: bytes, filename: str, session_id: str)

            return ""
    
    def process_pdf(self, pdf_bytes: bytes, filename: str, session_id: str) -> Dict[str, Any]:
        """Process a PDF file and store in database"""
        print(f"📄 Processing PDF: {filename}")
        
        # Extract text
        text = self.extract_text_from_pdf(pdf_bytes)
        if not text:
            return {
                "success": False,
                "error": "Could not extract text from PDF",
                "filename": filename
            }
        
        print(f"📝 Extracted {len(text)} characters from {filename}")
        
        # Store in database
        document_id = str(uuid.uuid4())
        now = datetime.now().isoformat()
        
        try:
            conn = sqlite3.connect(self.db_path)
            
            # Store document
            conn.execute('''
                INSERT INTO pdf_documents (id, session_id, filename, content, created_at)
                VALUES (?, ?, ?, ?, ?)
            ''', (document_id, session_id, filename, text, now))
            
            conn.commit()
            conn.close()
            
            print(f"💾 Stored document {filename} in database")
            
            return {
                "success": True,
                "filename": filename,
                "file_id": document_id,
                "text_length": len(text)
            }
            
        except Exception as e:
            print(f"❌ Error storing in database: {str(e)}")
            return {
                "success": False,
                "error": f"Database storage failed: {str(e)}",
                "filename": filename
            }
    
    def get_session_documents(self, session_id: str) -> List[Dict[str, Any]]:
        """Get all documents for a session"""