File size: 4,544 Bytes
534c0fb
29e2b7b
 
 
 
534c0fb
29e2b7b
c8c86d2
534c0fb
29e2b7b
 
534c0fb
c8c86d2
9d6d510
eb58b7a
c8c86d2
 
534c0fb
 
 
 
 
 
9d6d510
c8c86d2
 
29e2b7b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
534c0fb
29e2b7b
 
 
 
 
 
534c0fb
29e2b7b
 
534c0fb
c8c86d2
 
75a7b93
534c0fb
 
 
29e2b7b
 
 
534c0fb
 
 
c8c86d2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
534c0fb
 
 
 
 
 
 
c8c86d2
534c0fb
 
 
 
 
 
 
760198a
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
import os
from datetime import datetime
import json
import gradio as gr

from getFlightData import getFlightInfo

from langchain.agents import initialize_agent, Tool, load_tools, ZeroShotAgent, AgentExecutor, ConversationalAgent
from langchain.chat_models import ChatOpenAI
from langchain.output_parsers import StructuredOutputParser, ResponseSchema
from langchain.prompts import PromptTemplate, ChatPromptTemplate, HumanMessagePromptTemplate
from langchain.memory import ConversationBufferMemory
from langchain import LLMChain
import openai

os.environ["OPENAI_API_KEY"] = "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"
os.environ["SERPER_API_KEY"] = "488a16eee2a618380bd1cfbf6ad4fcec9f0a3b76"

today = datetime.today().strftime("%Y-%m-%d")

memory = ConversationBufferMemory(memory_key="chat_history")
llm = ChatOpenAI(model_name="gpt-3.5-turbo", temperature=0)

openai.api_base = "https://oai.hconeai.com/v1"


def parsing_infoJson(string):
    #Define the response schemas we want to receive
    response_schemas = [
    ResponseSchema(name="Flight_number", description="Flight number contained in the user sentence, if you cant find it, then it's equal to 0"),
    ResponseSchema(name="Date", description=f"Date contained in the user sentence, if you can't find it then equal to 0. today's date is: {today}. date format has to be: YYYY-MM-DD"),     
    ]
    output_parser = StructuredOutputParser.from_response_schemas(response_schemas)

    #We now get a string that contains instructions on how to format the user input, we instert this into the prompt
    format_instructions = output_parser.get_format_instructions()
    
    prompt = PromptTemplate(
    template="extract the flight number and the date from the user input. if cant find them, then they are equal to 0 \n{format_instructions}\n{question}",
    input_variables=["question"],
    partial_variables={"format_instructions": format_instructions}
    )

    #send prompt to model
    _input = prompt.format_prompt(question=string)
    output = llm(_input.to_messages())

    result = output_parser.parse(output.content)  

    if result["Flight_number"] == 0:
        return("I need the flight number and date to give you more info")
    elif result["Date"] == 0:
        return("the user didnt provide flight number or date, ask him to do so")
    else:
        date = datetime.strptime(result["Date"].strip(), "%Y-%m-%d").date()
        info = getFlightInfo(result["Flight_number"], date)

        summarize = PromptTemplate(
        template="summarize this information to the user \n{question}",
        input_variables=["question"]
        )
        inp= summarize.format_prompt(question=str(info))
        out = llm(inp.to_messages())

        return out.content
    

premade_tools = load_tools(["google-serper"],llm=llm)
premade_tools[0].description = "Useful when a customer asks for anything related to atihad, always use this tool first to get the most accurate answer"
premade_tools[0].return_direct = False
custom_tools = [
    Tool(
        name = "getSpecificInfoFlight",
        func=parsing_infoJson,
        description= f"Useful for when the user wants to know specific more about their flight, input the flight number and date. today's date is: {today}. date format has to be: YYYY-MM-DD",
        return_direct=True
    ),
]

tools = premade_tools




####

prefix = """You are an AI assistant that works for the Airline company 'Etihad'. Your objective is to assist customer with their inquires. You have access to the following tools:"""
suffix = """Begin! "

Question: {input}
{agent_scratchpad}"""

prompt = ConversationalAgent.create_prompt(
    tools, 
    prefix=prefix, 
    suffix=suffix, 
    input_variables=["input", "agent_scratchpad"]
)

llm_chain = LLMChain(llm=llm, prompt=prompt)

####


tool_names = [tool.name for tool in tools]
agent = ConversationalAgent(llm_chain=llm_chain, required_tools=tool_names)

agent_executor = AgentExecutor.from_agent_and_tools(agent=agent, tools=tools, verbose=True)


#agent = initialize_agent(tools, llm, agent="conversational-react-description", verbose=True, memory=memory)







def main(user_input, history=[]):
    answer = agent_executor.run(user_input)
    history.append((user_input, answer))

    return history, history

inputs = gr.components.Textbox(lines=7, label="Chat with AI")
outputs = gr.components.Textbox(label="Reply")

iface = gr.Interface(fn=main, inputs=["text","state"], outputs=["chatbot","state"], title="AI Chatbot",
             theme="compact")

iface.launch()