koenieeeee
22 lines
1.1 KiB
Python
22 lines
1.1 KiB
Python
from llm_communicate import LLM_Communicator
|
|
from prompt_techniques import LLM_chain_of_thought, LLM_few_shot_learning, LLM_prompt_data, LLM_role_play
|
|
from logger import ResearchLogger
|
|
|
|
|
|
def ask_different_prompts(input_text: str, prompt_tech: str, log:ResearchLogger, temperature, incorrect: bool = False) -> str:
|
|
prompt_data = LLM_prompt_data(input_text)
|
|
llm_communicator = LLM_Communicator()
|
|
fsl_prompt = LLM_few_shot_learning(prompt_data, incorrect)
|
|
cot_prompt = LLM_chain_of_thought(prompt_data, incorrect)
|
|
rp_prompt = LLM_role_play(prompt_data, incorrect)
|
|
result = ""
|
|
if prompt_tech.strip() == "Few-shot-learning": #['Few-shot learning', 'Chain-of-thought', 'Role play']
|
|
result = llm_communicator.ask_llm_to_convert(fsl_prompt, temperature)
|
|
elif prompt_tech.strip() == "Chain-of-thought":
|
|
result = llm_communicator.ask_llm_to_convert(cot_prompt, temperature)
|
|
else: #Role play
|
|
result = llm_communicator.ask_llm_to_convert(rp_prompt, temperature)
|
|
|
|
llm_communicator.write_log_output(log)
|
|
|
|
return result |