Spaces:
No application file
No application file
File size: 2,286 Bytes
1d5e089 |
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 |
from dotenv import load_dotenv
load_dotenv()
import os
from langchain_groq import ChatGroq
from langchain_community.tools.tavily_search import TavilySearchResults
from langgraph.prebuilt import create_react_agent
from langchain_core.messages import HumanMessage, SystemMessage, AIMessage
def get_response_from_ai_agent(llm_id, query, allow_search, system_prompt):
"""
Create and invoke an AI agent with optional search capabilities
"""
try:
# Initialize LLM with proper configuration
llm = ChatGroq(
api_key=os.environ.get("GROQ_API_KEY"),
model_name=llm_id
)
# Setup tools based on allow_search flag
tools = []
if allow_search:
tools.append(TavilySearchResults(
api_key=os.environ.get("TAVILY_API_KEY"),
max_results=2
))
# Create the agent
agent = create_react_agent(
model=llm,
tools=tools
)
# Prepare the initial messages
initial_messages = [
SystemMessage(content=system_prompt),
HumanMessage(content=query)
]
# Create proper state with messages
state = {
"messages": initial_messages,
"next_steps": [],
"structured_response": None
}
# Invoke agent with proper state
response = agent.invoke(state)
# Handle response
if isinstance(response, dict):
# Extract messages from response
if "messages" in response:
messages = response["messages"]
# Get the last AI message
ai_messages = [msg for msg in messages if isinstance(msg, AIMessage)]
if ai_messages:
return ai_messages[-1].content
# Check structured response
elif "structured_response" in response:
return response["structured_response"]
return "I apologize, but I couldn't generate a proper response. Please try again."
except Exception as e:
print(f"Debug - Error in get_response_from_ai_agent: {str(e)}")
raise Exception(f"Agent error: {str(e)}") |