Spaces:
Sleeping
Sleeping
File size: 1,654 Bytes
f8734b3 94bbd69 fdfde63 144f199 bcfaa84 b48a433 bcfaa84 94bbd69 fdfde63 94bbd69 c2d02ae 94bbd69 bcfaa84 b6b7990 c2d02ae 94bbd69 bcfaa84 94bbd69 bcfaa84 94bbd69 bcfaa84 | 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 | from langgraph.graph import StateGraph
from langchain_core.tools import Tool
from langchain_community.chat_models import ChatOpenAI
from langchain_community.tools import TavilySearchResults
def build_graph():
graph = StateGraph(dict)
# Setup tools
search = TavilySearchResults()
def search_step(state):
question = state.get("question")
if not question:
raise ValueError("Missing 'question' in state")
result = search.run(question)
state["search_result"] = result
return state
def llm_step(state):
llm = ChatOpenAI(model="gpt-3.5-turbo", temperature=0)
question = state.get("question")
search_info = state.get("search_result", "")
prompt = f"""
You are solving a GAIA benchmark question.
Here’s the question:
{question}
Here’s retrieved web info:
{search_info}
ONLY return the final exact answer (no explanation, no prefix).
"""
response = llm.invoke(prompt)
state["llm_output"] = response.content.strip()
return state
def formatter_step(state):
llm_output = state.get("llm_output")
if not llm_output:
raise ValueError("Missing 'llm_output' in state")
cleaned = llm_output.strip()
state["final_answer"] = cleaned
return state
# Add nodes
graph.add_node("search", search_step)
graph.add_node("llm", llm_step)
graph.add_node("formatter", formatter_step)
# Define flow
graph.set_entry_point("search")
graph.add_edge("search", "llm")
graph.add_edge("llm", "formatter")
graph.set_finish_point("formatter")
return graph
|