Spaces:
Sleeping
Sleeping
File size: 986 Bytes
65ade69 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 | from typing import TypedDict, Annotated, List, Any
import operator
# The `AgentState` defines the data structure that is passed between nodes in our LangGraph.
# Every node in the graph will receive this state, and can return updates to it.
class AgentState(TypedDict):
# The original startup idea provided by the user.
idea: str
# We use `Annotated` with `operator.add` to create a "reducer".
# Since our 4 research agents will run in PARALLEL, they will all try to update the `results` key at the same time.
# Without `operator.add`, the last agent to finish would overwrite the others.
# With `operator.add`, LangGraph knows to *append* their results into a single list safely.
results: Annotated[List[dict[str, Any]], operator.add]
# The final evaluation score from 0-100
score: int
# A categorical verdict: "Go", "Caution", or "No-Go"
verdict: str
# A brief textual summary of the final report
summary: str
|