Dynamically created a schema - MongoDB

app.py

@@ -115,102 +115,7 @@ def extract_invoice_data(file_data, content_type):
             ],
             response_format={
                 "type": "json_schema",
-                "json_schema": {
-                    "name": "invoice",
-                    "strict": True,
-                    "schema": {
-                        "type": "object",
-                        "title": "Invoice Information Extractor",
-                        "$schema": "http://json-schema.org/draft-07/schema#",
-                        "properties": {
-                            "LineItems": {
-                                "type": "array",
-                                "items": {
-                                    "type": "object",
-                                    "required": [
-                                        "ProductCode",
-                                        "Description",
-                                        "Amount"
-                                    ],
-                                    "properties": {
-                                        "ProductCode": {
-                                            "type": "string",
-                                            "title": "Product Code",
-                                            "description": "The code of the product"
-                                        },
-                                        "Description": {
-                                            "type": "string",
-                                            "title": "Description",
-                                            "description": "Description of the product"
-                                        },
-                                        "Amount": {
-                                            "type": "number",
-                                            "title": "Amount",
-                                            "description": "The amount of the product"
-                                        }
-                                    },
-                                    "additionalProperties": False
-                                },
-                                "title": "Line Items",
-                                "description": "List of line items on the invoice"
-                            },
-                            "TaxAmount": {
-                                "type": "number",
-                                "title": "Tax Amount",
-                                "description": "The tax amount on the invoice"
-                            },
-                            "VendorGST": {
-                                "type": "string",
-                                "title": "Vendor GST",
-                                "description": "The GST number of the vendor"
-                            },
-                            "VendorName": {
-                                "type": "string",
-                                "title": "Vendor Name",
-                                "description": "The name of the vendor"
-                            },
-                            "InvoiceDate": {
-                                "type": "string",
-                                "title": "Invoice Date",
-                                "description": "The date of the invoice (format: dd-MMM-yyyy)"
-                            },
-                            "TotalAmount": {
-                                "type": "number",
-                                "title": "Total Amount",
-                                "description": "The total amount on the invoice"
-                            },
-                            "InvoiceNumber": {
-                                "type": "string",
-                                "title": "Invoice Number",
-                                "description": "The number of the invoice"
-                            },
-                            "VendorAddress": {
-                                "type": "string",
-                                "title": "Vendor Address",
-                                "description": "The address of the vendor"
-                            },
-                            "InvoiceCurrency": {
-                                "type": "string",
-                                "title": "Invoice Currency",
-                                "description": "The currency used in the invoice (e.g., USD, INR, AUD)"
-                            }
-                        },
-                        "required": [
-                            "LineItems",
-                            "TaxAmount",
-                            "VendorGST",
-                            "VendorName",
-                            "InvoiceDate",
-                            "TotalAmount",
-                            "InvoiceNumber",
-                            "VendorAddress",
-                            "InvoiceCurrency"
-                        ],
-                        "additionalProperties": False,
-                        "description": "Schema for extracting structured invoice data"
-                    }
-                }
-            },
+                "json_schema": json_schema
             temperature=0.5,
             max_tokens=16384
         )
@@ -265,13 +170,19 @@ def extract_text_from_file(
                 "message": "Document Retrieved from MongoDB.",
                 "document": existing_document
             }
+        # Fetch dynamic schema based on document type
+        schema_doc = db.schema_collection.find_one({"document_type": document_type})
+        if not schema_doc:
+            raise ValueError("No schema found for the given document type")
+
+        json_schema = schema_doc.get("json_schema")    
             
         # Retrieve file from S3 and determine content type
         content_type = get_content_type_from_s3(file_key)
 
         # Extract and parse invoice data
         file_data, _ = fetch_file_from_s3(file_key)
-        extracted_data = extract_invoice_data(file_data, content_type)
+        extracted_data = extract_invoice_data(file_data, content_type,json_schema)
 
         # Store extracted data in MongoDB
         document = {