File size: 1,643 Bytes
8360ec7 |
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 |
from core import register_solver, FactCheckerState, StandardTaskSolver
from factool import Factool
import os
##
#
# Factool Solver
#
# Notes:
# Factool requires 3 input parameters: prompt, response, and category.
# Category is always set to 'kbqa' (Knowledge Base Question Answering) for the purposes of this project.
# Because of employing a pipeline of its own, with specific search engine and analysis tools, Factool requires several API keys to be set as environment variables.
# That is:
# openai_key - OpenAI API key (https://beta.openai.com/)
# serper_key - Serper API key (https://serper.dev/)
# scrapper_key - Scrapper API key (https://www.scraperapi.com/)
# Additional parameters:
# llm_in_use - The OpenAI LLM in use (e.g. gpt-4)
#
##
@register_solver("factool_blackbox", "response", "claim_info")
class FactoolBlackboxSolver(StandardTaskSolver):
def __init__(self, args):
super().__init__(args)
self.input_prompt = args.get("input_prompt", None)
self.gpt_model = self.global_config.get("llm_in_use", "gpt-4")
# self.input_prompt = args["input_prompt"] if "input_prompt" in args else None
# self.gpt_model = args["llm_in_use"] if "llm_in_use" in args else "gpt-4"
def __call__(self, state: FactCheckerState, *args, **kwargs):
prompt = state.get(self.input_prompt)
response = state.get(self.input_name)
factool_instance = Factool(self.gpt_model)
inputs = [{"prompt": prompt, "response": response, "category": "kbqa"}]
claim_info = factool_instance.run(inputs)
state.set("claim_info", claim_info)
return True, state
|