lzy/sci-gui-agent-benchmark
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Disable a11y tree temporarily

Browse Source
This commit is contained in:
Timothyxxx
2024-03-18 21:43:35 +08:00
parent 1c9c5fd2ad
commit f992d1f694
2 changed files with 22 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
lib_run_single.py
View File

@@ -2,7 +2,7 @@ import datetime
import json
import logging
import os
import wandb
# import wandb
from wrapt_timeout_decorator import *
@@ -15,13 +15,13 @@ with open("./settings.json", "r") as file:
time_limit = data["time_limit"]
@timeout(time_limit, use_signals=False)
def run_single_example(agent, env, example, max_steps, instruction, args, example_result_dir, scores, run):
def run_single_example(agent, env, example, max_steps, instruction, args, example_result_dir, scores):
agent.reset()
obs = env.reset(task_config=example)
done = False
step_idx = 0
env.controller.start_recording()
str_table = wandb.Table(columns=["Screenshot", "A11T", "Modle Response", "Action", "Action timestamp", "Done"])
# str_table = wandb.Table(columns=["Screenshot", "A11T", "Modle Response", "Action", "Action timestamp", "Done"])
while not done and step_idx < max_steps:
response, actions = agent.predict(
instruction,
@@ -43,10 +43,10 @@ def run_single_example(agent, env, example, max_steps, instruction, args, exampl
_f.write(screenshot)
# get a11tree and save to wandb
thisrun_a11tree = env.controller.get_accessibility_tree()
str_table.add_data(wandb.Image(data_or_path=os.path.join(example_result_dir, f"step_{step_idx + 1}_{action_timestamp}.png"), caption=f"step_{step_idx + 1}_{action_timestamp}"),
thisrun_a11tree,
response, action, action_timestamp, done)
run.log({"Reward": reward})
# str_table.add_data(wandb.Image(data_or_path=os.path.join(example_result_dir, f"step_{step_idx + 1}_{action_timestamp}.png"), caption=f"step_{step_idx + 1}_{action_timestamp}"),
# thisrun_a11tree,
# response, action, action_timestamp, done)
# run.log({"Reward": reward})
with open(os.path.join(example_result_dir, "traj.jsonl"), "a") as f:
f.write(json.dumps({
"step_num": step_idx + 1,
@@ -62,11 +62,11 @@ def run_single_example(agent, env, example, max_steps, instruction, args, exampl
logger.info("The episode is done.")
break
step_idx += 1
run.log({"str_trajectory": str_table})
# run.log({"str_trajectory": str_table})
result = env.evaluate()
logger.info("Result: %.2f", result)
scores.append(result)
with open(os.path.join(example_result_dir, "result.txt"), "w", encoding="utf-8") as f:
f.write(f"{result}\n")
env.controller.end_recording(os.path.join(example_result_dir, "recording.mp4"))
run.log({"Result": result})
# run.log({"Result": result})