From 16d1bf0da0e46a9885d7cba9b12cdae93fac065a Mon Sep 17 00:00:00 2001 From: Yizhou Chi Date: Mon, 2 Sep 2024 09:59:47 +0800 Subject: [PATCH] Rename insight generate to instruction generator --- expo/MCTS.py | 2 +- expo/experimenter/aug.py | 8 ++++---- .../{InsightGenerate.py => instruction_generator.py} | 8 ++++---- expo/run_exp_augmentation.py | 8 ++++---- 4 files changed, 13 insertions(+), 13 deletions(-) rename expo/insights/{InsightGenerate.py => instruction_generator.py} (91%) diff --git a/expo/MCTS.py b/expo/MCTS.py index 5c448e3ac..42c93f650 100644 --- a/expo/MCTS.py +++ b/expo/MCTS.py @@ -3,7 +3,7 @@ import math import os import pandas as pd from expo.research_assistant import ResearchAssistant -from expo.insights.InsightGenerate import InsightGenerator +from exp_optimizer.expo.insights.instruction_generator import InstructionGenerator from expo.dataset import get_split_dataset_path, generate_task_requirement from expo.evaluation.evaluation import evaluate_score from expo.utils import mcts_logger, load_execute_notebook, get_exp_pool_path diff --git a/expo/experimenter/aug.py b/expo/experimenter/aug.py index 28643d47f..299c053cd 100644 --- a/expo/experimenter/aug.py +++ b/expo/experimenter/aug.py @@ -2,7 +2,7 @@ from experimenter import Experimenter from expo.MCTS import create_initial_state from expo.dataset import generate_task_requirement from expo.utils import mcts_logger, load_execute_notebook, get_exp_pool_path -from expo.insights.InsightGenerate import InsightGenerator +from exp_optimizer.expo.insights.instruction_generator import InstructionGenerator from expo.research_assistant import ResearchAssistant EXPS_PROMPT = """ @@ -21,12 +21,12 @@ class AugExperimenter(Experimenter): state = create_initial_state(self.args.task, start_task_id=1, data_config=self.data_config, low_is_better=self.args.low_is_better, name="") user_requirement = state["requirement"] exp_pool_path = get_exp_pool_path(self.args.task, self.data_config, pool_name="ds_analysis_pool") - exp_pool = InsightGenerator.load_analysis_pool(exp_pool_path) + exp_pool = InstructionGenerator.load_analysis_pool(exp_pool_path) if self.args.aug_mode == "single": - exps = InsightGenerator._random_sample(exp_pool, self.args.num_experiments) + exps = InstructionGenerator._random_sample(exp_pool, self.args.num_experiments) exps = [exp["Analysis"] for exp in exps] elif self.args.aug_mode == "set": - exp_set = InsightGenerator.sample_instruction_set(exp_pool) + exp_set = InstructionGenerator.sample_instruction_set(exp_pool) exp_set_text = "\n".join([f"{exp['task_id']}: {exp['Analysis']}" for exp in exp_set]) exps = [exp_set_text] * self.args.num_experiments else: diff --git a/expo/insights/InsightGenerate.py b/expo/insights/instruction_generator.py similarity index 91% rename from expo/insights/InsightGenerate.py rename to expo/insights/instruction_generator.py index 35bed976a..4f4155ff8 100644 --- a/expo/insights/InsightGenerate.py +++ b/expo/insights/instruction_generator.py @@ -27,7 +27,7 @@ from expo.utils import load_data_config, mcts_logger, clean_json_from_rsp DATA_CONFIG = load_data_config() -class InsightGenerator: +class InstructionGenerator: data_config = DATA_CONFIG @staticmethod @@ -68,7 +68,7 @@ class InsightGenerator: @staticmethod def load_analysis_pool(file_path, task_id=None): - data = InsightGenerator.load_json_data(file_path) + data = InstructionGenerator.load_json_data(file_path) for item in data: if "task_id" not in item: raise ValueError("task_id is not found in the analysis pool") @@ -79,14 +79,14 @@ class InsightGenerator: @staticmethod async def generate_new_instructions(task_id, original_instruction, max_num, file_path): - data = InsightGenerator.load_analysis_pool(file_path, task_id) + data = InstructionGenerator.load_analysis_pool(file_path, task_id) new_instructions = [] if len(data) == 0: mcts_logger.log("MCTS", f"No insights available for task {task_id}") return [original_instruction] # Return the original instruction if no insights are available for item in data[:max_num]: insights = item["Analysis"] - new_instruction = await InsightGenerator.generate_new_instruction(original_instruction, insights) + new_instruction = await InstructionGenerator.generate_new_instruction(original_instruction, insights) new_instructions.append(new_instruction) return new_instructions diff --git a/expo/run_exp_augmentation.py b/expo/run_exp_augmentation.py index f4d22093f..edf01c164 100644 --- a/expo/run_exp_augmentation.py +++ b/expo/run_exp_augmentation.py @@ -3,7 +3,7 @@ from expo.research_assistant import ResearchAssistant import asyncio from expo.utils import DATA_CONFIG, get_exp_pool_path from expo.dataset import generate_task_requirement -from expo.insights.InsightGenerate import InsightGenerator +from exp_optimizer.expo.insights.instruction_generator import InstructionGenerator from expo.MCTS import create_initial_state from expo.evaluation.evaluation import evaluate_score import json @@ -48,12 +48,12 @@ async def main(task_name, use_reflection=True, mode="single", num_experiments=2) user_requirement = generate_task_requirement(task_name, data_config) exp_pool_path = get_exp_pool_path(task_name, data_config, pool_name="ds_analysis_pool") - exp_pool = InsightGenerator.load_analysis_pool(exp_pool_path) + exp_pool = InstructionGenerator.load_analysis_pool(exp_pool_path) if mode == "single": - exps = InsightGenerator._random_sample(exp_pool, num_experiments) + exps = InstructionGenerator._random_sample(exp_pool, num_experiments) exps = [exp["Analysis"] for exp in exps] elif mode == "set": - exp_set = InsightGenerator.sample_instruction_set(exp_pool) + exp_set = InstructionGenerator.sample_instruction_set(exp_pool) exp_set_text = "\n".join([f"{exp['task_id']}: {exp['Analysis']}" for exp in exp_set]) exps = [exp_set_text] * num_experiments else: