front desk agent added, that give response in json

OAI_CONFIG_LIST.json

@@ -0,0 +1,11 @@
+[
+    {
+      "model": "gpt-3.5-turbo",
+      "api_key": "api-key-here"
+    },
+    {
+      "model": "gpt-4",
+      "api_key": "api-key-here"
+    }
+  ]
+  

download_verify_email.py

@@ -0,0 +1,62 @@
+import PyPDF2
+import requests
+import hashlib
+import io
+def verify_bank_pdf(pdf_url):
+    """
+    Download a PDF from a given URL and perform basic integrity checks.
+    
+    Args:
+    pdf_url (str): The URL of the PDF file to download and verify.
+    
+    Returns:
+    dict: A dictionary containing verification results.
+    """
+    try:
+        # Download the PDF
+        response = requests.get(pdf_url)
+        if response.status_code != 200:
+            return {"success": False, "message": f"Failed to download PDF. Status code: {response.status_code}"}
+        
+        pdf_content = response.content
+        
+        # Calculate MD5 hash of the PDF
+        md5_hash = hashlib.md5(pdf_content).hexdigest()
+        
+        # Read the PDF
+        pdf_reader = PyPDF2.PdfReader(io.BytesIO(pdf_content))
+        
+        # Extract metadata
+        metadata = pdf_reader.metadata
+        
+        # Check if the PDF is encrypted
+        is_encrypted = pdf_reader.is_encrypted
+        
+        # Get the number of pages
+        num_pages = len(pdf_reader.pages)
+        
+        # Check for digital signatures (this is a basic check and may not catch all types of signatures)
+        has_signature = '/Sig' in pdf_reader.trailer['/Root'].get('/AcroForm', {}).get('/Fields', {})
+        
+        return {
+            "success": True,
+            "md5_hash": md5_hash,
+            "is_encrypted": is_encrypted,
+            "num_pages": num_pages,
+            "has_signature": has_signature,
+            "metadata": {
+                "author": metadata.get('/Author', 'N/A'),
+                "creator": metadata.get('/Creator', 'N/A'),
+                "producer": metadata.get('/Producer', 'N/A'),
+                "subject": metadata.get('/Subject', 'N/A'),
+                "title": metadata.get('/Title', 'N/A'),
+                "creation_date": metadata.get('/CreationDate', 'N/A'),
+                "modification_date": metadata.get('/ModDate', 'N/A')
+            }
+        }
+    except Exception as e:
+        return {"success": False, "message": f"An error occurred: {str(e)}"}
+
+# Example usage
+# result = verify_bank_pdf("https://example.com/bank_statement.pdf")
+# print(result)

main.py

@@ -5,39 +5,55 @@ from dotenv import load_dotenv
 load_dotenv()  # take environment variables from .env.
 config_list = [
     {
-        'model': 'gpt-3.5-turbo-16k',
+        'model': 'gpt-3.5-turbo',
         'api_key': os.getenv('OPENAI_API_KEY'),
     }
 ]
 
 llm_config = {
-    "timeout": 300,
+    "timeout": 120,
     "seed": 42, # for caching. once task is run it caches the response,
     "config_list": config_list,
     "temperature": 0 #lower temperature more standard lesss creative response, higher is more creative
 
 }
 
-assistant = autogen.AssistantAgent(
-    name="laon_assistant",
+front_desk_assistant = autogen.AssistantAgent(
+    name="front_desk_assistant",
     llm_config=llm_config,
-    system_message="checks the bank documents. extract pdf using extract_pdf_skill.",
+    system_message="""You have a personality of monopoly banker. You have to ask questions and collect information from user
+    Questions that you have to ask : What is your name, How much do you want to borrow, What country do you live in,
+    What bank do you use, Do you have a job/proof of income, Do you have any history of not paying back loans?
+    once you collect all these answers, create a json response with following key 
+    {"first_name" : "",  last_name: "", "country" : "", "bank" : "", "income" : "", "history" : "", loan_amount : ""}
+    Ask user for confirmation that the details are right and want to proceed with it. Show the response in json format and save it to a file
+    """,
 )
 
+# assistant = autogen.AssistantAgent(
+#     name="laon_assistant",
+#     llm_config=llm_config,
+#     system_message="checks the bank documents. extract pdf using extract_pdf_skill.",
+# )
+
 user_proxy = autogen.UserProxyAgent(
     name="user_proxy",
-    human_input_mode="TERMINATE",
+    human_input_mode="ALWAYS",
     max_consecutive_auto_reply=3,
     is_termination_msg=lambda x: x.get("content", "").rstrip().endswith("TERMINATE"),
-    code_execution_config={"work_dir": "web"},
+    code_execution_config={
+        "last_n_messages": 3,
+        "work_dir": "code",
+        "use_docker": False,
+    },
     llm_config=llm_config,
     system_message="""Reply TERMINATE if the task has been solved at full satisfaction
     otherwise, reply CONTINUE, or the reason why the task is not solved yet."""
 )
 
 def main():
-    task = """"""
-    user_proxy.initiate_chat(assistant, message=task)
+    # task = """"""
+    user_proxy.initiate_chat(front_desk_assistant, message="I want to apply for a loan, please help me")
 
 if __name__ == "__main__":
     main()