Solution published: update src/main.py

This commit is contained in:
2026-06-18 09:55:02 +00:00
parent 3e018c42bb
commit 08ab537f6b
+80 -56
View File
@@ -1,11 +1,12 @@
"""
LangGraph research brief generator.
LangGraph comparison agent.
Usage:
python -m src.main "Topic"
python -m src.main "Entity1,Entity2,Entity3"
"""
import os
import json
from typing import TypedDict, List, Dict
from langgraph.graph import StateGraph
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_tavily import TavilySearch
from dotenv import load_dotenv
@@ -13,90 +14,113 @@ from dotenv import load_dotenv
load_dotenv()
# ---------- State ----------
class BriefState(TypedDict):
topic: str
sections: List[str]
content: Dict[str, str]
brief: str | None
class CompareState(TypedDict):
entities: List[str] # 3 names for comparison
criteria: List[str] # 35 criteria
findings: Dict[str, Dict[str, str]] # entity -> criterion -> note
final_table: str | None
verdict: str | None
# ---------- Nodes ----------
# ---------- LLM and tools ----------
llm = ChatOpenAI(temperature=0)
search_tool = TavilySearch(api_key=os.getenv("TAVILY_API_KEY"))
async def plan_sections(state: BriefState) -> BriefState:
# ---------- Nodes ----------
async def plan_criteria(state: CompareState) -> CompareState:
prompt = (
f"You are a helpful assistant. Given the topic: {state['topic']}. "
"Generate 5 concise section headings for a research brief. Return a JSON array of strings."
f"You are a helpful assistant. Given the entities: {', '.join(state['entities'])}. "
"Generate 3 to 5 concise criteria for comparing these entities. "
"Return a JSON array of strings."
)
response = await llm.ainvoke({"role": "user", "content": prompt})
import json
try:
sections = json.loads(response["content"].strip())
criteria = json.loads(response["content"].strip())
except Exception as e:
raise ValueError(f"Failed to parse sections: {e}")
state["sections"] = sections
raise ValueError(f"Failed to parse criteria: {e}")
state["criteria"] = criteria
return state
async def fetch_section(state: BriefState) -> BriefState:
# find next section not yet fetched
for sec in state["sections"]:
if sec not in state["content"]:
query = f"{state['topic']} {sec}"
async def research_entity(state: CompareState) -> CompareState:
# Find next unprocessed entity-criterion pair
for entity in state["entities"]:
for criterion in state["criteria"]:
if entity not in state["findings"] or criterion not in state["findings"][entity]:
query = f"{entity} {criterion}"
result = await search_tool.ainvoke({"query": query, "max_results": 1})
snippet = result.get("results", [{}])[0].get("snippet", "")
content = snippet if snippet else "No relevant information found."
state["content"][sec] = content
note = snippet if snippet else "No relevant information found."
state.setdefault("findings", {})
state["findings"][entity] = state["findings"].get(entity, {})
state["findings"][entity][criterion] = note
return state
return state
async def build_brief(state: BriefState) -> BriefState:
lines = [f"# Research Brief: {state['topic']}\n"]
for sec in state["sections"]:
lines.append(f"## {sec}\n")
lines.append(state["content"][sec] + "\n")
lines.append("---\n")
lines.append("**Summary**\n")
# simple summary using LLM
summary_prompt = (
f"Based on the following sections:\n{''.join(lines)}\n"
"Provide a concise summary of the brief in 3-4 sentences."
async def build_table(state: CompareState) -> CompareState:
# Build markdown table
header = "| Criterion | " + " | ".join(state["entities"]) + " |"
separator = "|---|" + "---|" * len(state["entities"]) # simple separator
rows = []
for criterion in state["criteria"]:
cells = [criterion]
for entity in state["entities"]:
note = state["findings"].get(entity, {}).get(criterion, "")
cells.append(note.replace("\n", " "))
rows.append("| " + " | ".join(cells) + " |")
table = "\n".join([header, separator] + rows)
state["final_table"] = table
return state
async def verdict(state: CompareState) -> CompareState:
prompt = (
f"You have the following comparison table:\n{state['final_table']}\n"
"Based on this table, provide a concise verdict: which entity is best for a general-purpose use case, and why."
)
summary_resp = await llm.ainvoke({"role": "user", "content": summary_prompt})
lines.append(summary_resp["content"].strip() + "\n")
state["brief"] = "\n".join(lines)
response = await llm.ainvoke({"role": "user", "content": prompt})
state["verdict"] = response["content"].strip()
return state
# ---------- Graph ----------
builder = StateGraph(BriefState)
builder.add_node("plan_sections", plan_sections)
builder.add_node("fetch_section", fetch_section)
builder.add_node("build_brief", build_brief)
builder = StateGraph(CompareState)
builder.add_node("plan_criteria", plan_criteria)
builder.add_node("research_entity", research_entity)
builder.add_node("build_table", build_table)
builder.add_node("verdict", verdict)
builder.set_entry_point("plan_sections")
# after planning, fetch sections until all done
builder.set_entry_point("plan_criteria")
# After planning, loop research until all pairs processed
builder.add_conditional_edges(
"plan_sections",
lambda _: "fetch_section",
"plan_criteria",
lambda _: "research_entity",
)
builder.add_conditional_edges(
"fetch_section",
lambda state: "build_brief" if all(sec in state["content"] for sec in state["sections"]) else "fetch_section",
"research_entity",
lambda state: "build_table" if all(
criterion in state["findings"].get(entity, {}) for entity in state["entities"] for criterion in state["criteria"]
) else "research_entity",
)
builder.add_edge("build_brief", "END")
builder.add_edge("build_table", "verdict")
builder.add_edge("verdict", END)
app = builder.compile()
# ---------- Demo ----------
if __name__ == "__main__":
import argparse
parser = argparse.ArgumentParser(description="Generate a research brief on a topic.")
parser.add_argument("topic", nargs='?', default="Artificial Intelligence", help="Topic for the research brief")
parser = argparse.ArgumentParser(description="Compare three entities and produce a table and verdict.")
parser.add_argument("entities", nargs='?', default="Chroma,FAISS,Qdrant", help="Comma-separated list of three entities to compare")
args = parser.parse_args()
init_state: BriefState = {
"topic": args.topic,
"sections": [],
"content": {},
"brief": None,
entities = [e.strip() for e in args.entities.split(',') if e.strip()]
if len(entities) != 3:
raise ValueError("Please provide exactly three entities separated by commas.")
init_state: CompareState = {
"entities": entities,
"criteria": [],
"findings": {},
"final_table": None,
"verdict": None,
}
result = app.invoke(init_state)
print(result["brief"])
print("## Comparison Table\n")
print(result["final_table"], "\n")
print("## Verdict\n")
print(result["verdict"], "\n")