Untitled
unknown
plain_text
a year ago
6.9 kB
7
Indexable
#Running command: streamlit run Chatbot.py
import openai
import streamlit as st
import requests
import json
from pathlib import Path
from dotenv import load_dotenv
import os
load_dotenv()
OPENAI_API_KEY = os.getenv('OPENAI_API_KEY')
def load_custom_css():
custom_css = Path("styles.css")
st.markdown(f"<style>{custom_css.read_text()}</style>", unsafe_allow_html=True)
# Call the function to load the custom CSS
load_custom_css()
def get_customer_information(Customer_ID):
"""Get the customer information using the customer_ID argument"""
URL = "http://127.0.0.1:8000/customer/{}".format(Customer_ID)
r = requests.get(url = URL)
data = r.json()
print(URL)
return json.dumps(data[0])
openai_api_key = OPENAI_API_KEY
# with st.sidebar:
# openai_api_key = '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'
# "[Get an OpenAI API key](https://platform.openai.com/account/api-keys)"
# "[View the source code](https://github.com/streamlit/llm-examples/blob/main/Chatbot.py)"
# "[](https://codespaces.new/streamlit/llm-examples?quickstart=1)"
delimiter = "####"
system_message = f"""Instruction: Create a customized and CONSIZED POINTER backed by Data pulled and DO NOT create any facts.
Context: You are an assistant to Tele Sales agent for a AbInbev, a Beer company. Write a data backed reasons for selling more beer for my next customer whose details are as follows:
The details around the customer should be pulled using Customer ID.
If the user dosent provide the Customer ID then ask him to provide the Customer ID.
The customer service query will be delimited with \
{delimiter} characters.
Constraint:
1. Construct a selling story WITHIN 150 words
2. DO NOT create any facts.
3. Use ONLY customer information in the selling story.
4. NO Saluation or Greeting is required.
5. NEED ONLY 10 WORDS per pointer
6. CREATE ONLY 5 POINTERS
"""
# system_message = f'''**Instruction: Craft a Concise, Customized Sales Pitch Supported by Data**
# **Context:** You're an assistant aiding a Tele Sales agent at AbInbev, a beer company.
# Generate data-driven incentives to boost beer sales for the upcoming customer.
# Retrieve pertinent customer details via their Customer ID.
# If the Customer ID is missing, prompt for it.
# Use \{delimiter} to separate customer service queries.
# Constraint: Construct a selling story within 200 words
# '''
# user_message = f"""Can you tell me about customer with customer_ID 0010Y00001aLGsYQAW"""
functions = [
{
"name": "get_customer_information",
"description": "Get the customer information using the customer_ID argument",
"parameters": {
"type": "object",
"properties": {
"Customer_ID": {
"type": "string",
"description": "The customer Identifier, which is unique for each customer",
},
"Customer_Name": {
"type": "string",
"description": " This is the Customer Name",
},
},
"required": ["Customer_ID"],
},
}
]
st.title("💬 GenScribe: BDR Assistant")
counter =1
if "messages" not in st.session_state:
st.session_state["messages"] = [{"role": "assistant", "content": "Hey I am your Assistant !! Ready to help you with Selling Stories!!"}, {'role':'system','content': system_message}]
for msg in st.session_state.messages:
if (msg["role"] == "assistant" and msg["content"] is not None):
st.chat_message(msg["role"]).write(msg["content"].replace('####',''))
# counter+=1
elif( msg["role"] == "system"and counter>1):
st.chat_message(msg["role"]).write(msg["content"].replace('####',''))
counter+=1
elif (msg["role"] == "user" ):
st.chat_message(msg["role"]).write(msg["content"].replace('####',''))
counter+=1
# st.chat_message(msg["role"]).write(msg["content"])
if prompt := st.chat_input():
if not openai_api_key:
st.info("Please add your OpenAI API key to continue.")
st.stop()
openai.api_key = openai_api_key
st.session_state.messages.append({"role": "user", "content":f"{delimiter}{prompt}{delimiter}" })
st.chat_message("user").write(prompt)
response = openai.ChatCompletion.create(model="gpt-3.5-turbo", functions=functions,function_call="auto",messages=st.session_state.messages)
response_message = response.choices[0].message
# st.session_state.messages.append(response_message)
# st.chat_message("assistant").write(response_message.content)
if response_message.get("function_call"):
# Step 3: call the function
# Note: the JSON response may not always be valid; be sure to handle errors
available_functions = {
"get_customer_information": get_customer_information,
} # only one function in this example, but you can have multiple
function_name = response_message["function_call"]["name"]
fuction_to_call = available_functions[function_name]
function_args = json.loads(response_message["function_call"]["arguments"])
print("function_args : ", function_args)
function_response = fuction_to_call(
Customer_ID=function_args.get("Customer_ID"),
)
print(function_response)
# Step 4: send the info on the function call and function response to GPT
st.session_state.messages.append(response_message) # extend conversation with assistant's reply
st.session_state.messages.append(
{
"role": "function",
"name": function_name,
"content": function_response,
}
) # extend conversation with function response
second_response = openai.ChatCompletion.create(
model="gpt-3.5-turbo-0613",
messages=st.session_state.messages,
) # get a new response from GPT where it can see the function response
# return second_response
print(second_response)
print('****************Scenario 2*******************')
print(second_response.choices[0].message)
st.session_state.messages.append(second_response.choices[0].message)
st.chat_message("assistant").write(second_response.choices[0].message.content)
else:
print(response_message)
st.session_state.messages.append(response_message)
st.chat_message("assistant").write(response_message.content)
print('------------------------------------------------------')
print(st.session_state.messages)Leave a Comment