lzy/sci-gui-agent-benchmark
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

ver Jan5th

Browse Source 
debugged
This commit is contained in:
David Chang
2024-01-05 15:20:47 +08:00
parent 5fedf5b891
commit eeb8a120d6
17 changed files with 158 additions and 86 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
desktop_env/controllers/python.py
View File

@@ -3,6 +3,8 @@ from typing import Any, Dict
import requests
from desktop_env.envs.actions import KEYBOARD_KEYS
import logging
logger = logging.getLogger("desktopenv.pycontroller")
class PythonController:
def __init__(self, http_server: str, pkgs_prefix: str = "python -c \"import pyautogui; {command}\""):
@@ -17,7 +19,7 @@ class PythonController:
if response.status_code == 200:
return response.content
else:
print("Failed to get screenshot. Status code:", response.status_code)
logger.error("Failed to get screenshot. Status code: %d", response.status_code)
return None
def get_file(self, file_path: str):
@@ -26,10 +28,10 @@ class PythonController:
"""
response = requests.post(self.http_server + "/file", data={"file_path": file_path})
if response.status_code == 200:
print("File downloaded successfully")
logger.info("File downloaded successfully")
return response.content
else:
print("Failed to get file. Status code:", response.status_code)
logger.error("Failed to get file. Status code: %d", response.status_code)
return None
def execute_python_command(self, command: str) -> None:
@@ -38,7 +40,7 @@ class PythonController:
It can be used to execute the pyautogui commands, or... any other python command. who knows?
"""
command = self.pkgs_prefix.format(command=command)
payload = json.dumps({"command": command})
payload = json.dumps({"command": command, "shell": True})
headers = {
'Content-Type': 'application/json'
}
@@ -46,12 +48,12 @@ class PythonController:
try:
response = requests.post(self.http_server + "/execute", headers=headers, data=payload)
if response.status_code == 200:
print("Command executed successfully:", response.text)
logger.info("Command executed successfully: %s", response.text)
else:
print("Failed to execute command. Status code:", response.status_code)
logger.error("Failed to execute command. Status code: %d", response.status_code)
return response.json()
except requests.exceptions.RequestException as e:
print("An error occurred while trying to execute the command:", e)
logger.error("An error occurred while trying to execute the command: %s", e)
def execute_action(self, action: Dict[str, Any]):
"""

49
desktop_env/controllers/setup.py
View File

@@ -8,6 +8,10 @@ import os.path
from typing import Dict, List
from typing import Any, Union
import logging
logger = logging.getLogger("desktopenv.setup")
import traceback
class SetupController:
def __init__( self
@@ -88,12 +92,12 @@ class SetupController:
for chunk in response.iter_content(chunk_size=8192):
if chunk:
f.write(chunk)
print("File downloaded successfully")
logger.info("File downloaded successfully")
downloaded = True
break
except requests.RequestException as e:
print(f"Failed to download {url}. Retrying... ({max_retries - i - 1} attempts left)")
logger.error(f"Failed to download {url}. Retrying... ({max_retries - i - 1} attempts left)")
if not downloaded:
raise requests.RequestException(f"Failed to download {url}. No retries left. Error: {e}")
@@ -107,17 +111,18 @@ class SetupController:
}
)
headers = {"Content-Type": form.content_type}
print(form.content_type)
logger.debug(form.content_type)
# send request to server to upload file
try:
logger.debug("REQUEST ADDRESS: %s", self.http_server + "/upload")
response = requests.post(self.http_server + "/upload", headers=headers, data=form)
if response.status_code == 200:
print("Command executed successfully:", response.text)
logger.info("Command executed successfully: %s", response.text)
else:
print("Failed to upload file. Status code:", response.text)
logger.error("Failed to upload file. Status code: %s", response.text)
except requests.exceptions.RequestException as e:
print("An error occurred while trying to send the request:", e)
logger.error("An error occurred while trying to send the request: %s", e)
def _change_wallpaper_setup(self, path: str):
# if not config:
@@ -138,11 +143,11 @@ class SetupController:
try:
response = requests.post(self.http_server + "/change_wallpaper", headers=headers, data=payload)
if response.status_code == 200:
print("Command executed successfully:", response.text)
logger.info("Command executed successfully: %s", response.text)
else:
print("Failed to change wallpaper. Status code:", response.text)
logger.error("Failed to change wallpaper. Status code: %s", response.text)
except requests.exceptions.RequestException as e:
print("An error occurred while trying to send the request:", e)
logger.error("An error occurred while trying to send the request: %s", e)
def _tidy_desktop_setup(self, **config):
raise NotImplementedError()
@@ -165,11 +170,11 @@ class SetupController:
try:
response = requests.post(self.http_server + "/open_file", headers=headers, data=payload)
if response.status_code == 200:
print("Command executed successfully:", response.text)
logger.info("Command executed successfully: %s", response.text)
else:
print("Failed to open file. Status code:", response.text)
logger.error("Failed to open file. Status code: %s", response.text)
except requests.exceptions.RequestException as e:
print("An error occurred while trying to send the request:", e)
logger.error("An error occurred while trying to send the request: %s", e)
def _launch_setup(self, command: List[str]):
if not command:
@@ -181,11 +186,11 @@ class SetupController:
try:
response = requests.post(self.http_server + "/launch", headers=headers, data=payload)
if response.status_code == 200:
print("Command executed successfully:", response.text)
logger.info("Command executed successfully: %s", response.text)
else:
print("Failed to launch application. Status code:", response.text)
logger.error("Failed to launch application. Status code: %s", response.text)
except requests.exceptions.RequestException as e:
print("An error occurred while trying to send the request:", e)
logger.error("An error occurred while trying to send the request: %s", e)
def _execute_setup(self, command: List[str], stdout: str = "", stderr: str = ""):
if not command:
@@ -195,7 +200,7 @@ class SetupController:
headers = {"Content-Type": "application/json"}
try:
response = requests.post(self.http_server + "/launch", headers=headers, data=payload)
response = requests.post(self.http_server + "/execute", headers=headers, data=payload)
if response.status_code == 200:
results: Dict[str, str] = response.json()
if stdout:
@@ -204,13 +209,15 @@ class SetupController:
if stderr:
with open(os.path.join(self.cache_dir, stderr), "w") as f:
f.write(results["error"])
print( "Command executed successfully: {:} ->".format(" ".join(command))
, response.text
)
logger.info( "Command executed successfully: %s -> %s"
, " ".join(command)
, response.text
)
else:
print("Failed to launch application. Status code:", response.text)
logger.error("Failed to launch application. Status code: %s", response.text)
except requests.exceptions.RequestException as e:
print("An error occurred while trying to send the request:", e)
logger.error("An error occurred while trying to send the request: %s", e)
traceback.print_exc()
def _act_setup(self, action_seq: List[Union[Dict[str, Any], str]]):
# TODO

36
desktop_env/envs/desktop_env.py
View File

@@ -17,10 +17,12 @@ from desktop_env.controllers.setup import SetupController
# from desktop_env.evaluators import eval_funcs
from desktop_env.evaluators import metrics, getters
import logging
logger = logging.getLogger("desktopenv.env")
Metric = Callable[[Any, Any], float]
Getter = Callable[[gym.Env, Dict[str, Any]], Any]
def _execute_command(command: List[str]) -> None:
if command[:4] == ["vmrun", "-T", "ws", "start"]:
p = subprocess.Popen(command)
@@ -72,7 +74,7 @@ class DesktopEnv(gym.Env):
self._set_task_info(task_config)
# Initialize emulator and controller
print("Initializing...")
logger.info("Initializing...")
self._start_emulator()
self.host = f"http://{self._get_vm_ip()}:5000"
self.controller = PythonController(http_server=self.host)
@@ -98,26 +100,26 @@ class DesktopEnv(gym.Env):
output: List[str] = output.splitlines()
# if self.path_to_vm.lstrip("~/") in output:
if self.path_to_vm in output:
print("VM is running.")
logger.info("VM is running.")
break
else:
print("Starting VM...")
logger.info("Starting VM...")
_execute_command(["vmrun", "-T", "ws", "start", self.path_to_vm])
time.sleep(3)
except subprocess.CalledProcessError as e:
print(f"Error executing command: {e.output.decode().strip()}")
logger.error(f"Error executing command: {e.output.decode().strip()}")
def _get_vm_ip(self):
max_retries = 10
print("Getting IP Address...")
logger.info("Getting IP Address...")
for _ in range(max_retries):
try:
output = _execute_command(["vmrun", "-T", "ws", "getGuestIPAddress", self.path_to_vm]).strip()
print(f"IP address: {output}")
logger.info(f"IP address: {output}")
return output
except:
time.sleep(5)
print("Retrying...")
logger.info("Retrying...")
raise Exception("Failed to get VM IP address!")
def _save_state(self):
@@ -156,38 +158,38 @@ class DesktopEnv(gym.Env):
self.metric_options: Dict[str, Any] = self.evaluator.get("options", {})
def reset(self, task_config: Optional[Dict[str, Any]] = None, seed=None, options=None):
print("Resetting environment...")
logger.info("Resetting environment...")
print("Switching task...")
logger.info("Switching task...")
if task_config is not None:
self._set_task_info(task_config)
self.setup_controller.reset_cache_dir(self.cache_dir)
print("Setting counters...")
logger.info("Setting counters...")
self._traj_no += 1
self._step_no = 0
self.action_history.clear()
print("Setup new temp dir...")
logger.info("Setup new temp dir...")
self.tmp_dir = tempfile.mkdtemp(
prefix="{:d}.{:}.".format(self._traj_no, self.task_id),
dir=self.tmp_dir_base
)
os.makedirs(os.path.join(self.tmp_dir, "screenshots"))
print("Reverting to snapshot to {}...".format(self.snapshot_path))
logger.info("Reverting to snapshot to {}...".format(self.snapshot_path))
_execute_command(["vmrun", "-T", "ws", "revertToSnapshot", self.path_to_vm, self.snapshot_path])
time.sleep(5)
print("Starting emulator...")
logger.info("Starting emulator...")
self._start_emulator()
print("Emulator started.")
logger.info("Emulator started.")
print("Setting up environment...")
logger.info("Setting up environment...")
self.setup_controller.setup(self.config)
time.sleep(5)
print("Environment setup complete.")
logger.info("Environment setup complete.")
observation = self._get_obs()
return observation

7
desktop_env/evaluators/getters/misc.py
View File

@@ -5,6 +5,9 @@ import subprocess
import ctypes
import os
import logging
logger = logging.getLogger("desktopenv.getters.misc")
R = TypeVar("Rule")
def get_rule(env, config: R) -> R:
"""
@@ -40,7 +43,7 @@ def get_wallpaper():
process = subprocess.Popen(['osascript', '-e', script], stdout=subprocess.PIPE)
output, error = process.communicate()
if error:
print("Error:", error)
logger.error("Error: %s", error)
else:
return output.strip().decode('utf-8')
@@ -49,7 +52,7 @@ def get_wallpaper():
output = subprocess.check_output(["gsettings", "get", "org.gnome.desktop.background", "picture-uri"])
return output.decode('utf-8').strip().replace('file://', '').replace("'", "")
except Exception as e:
print("Error:", e)
logger.error("Error: %s", e)
return None
os_name = platform.system()

11
desktop_env/evaluators/metrics/chrome.py
View File

@@ -5,6 +5,9 @@ import sqlite3
from playwright.sync_api import sync_playwright
import logging
logger = logging.getLogger("desktopenv.metrics.chrome")
"""
WARNING:
1. Functions from this script assume that no account is registered on Chrome, otherwise the default file path needs to be changed.
@@ -36,7 +39,7 @@ def get_default_search_engine():
'Google')
return search_engine
except Exception as e:
print(f"Error: {e}")
logger.error(f"Error: {e}")
return "Google"
@@ -61,7 +64,7 @@ def get_cookie_data():
return cookies
except Exception as e:
print(f"Error: {e}")
logger.error(f"Error: {e}")
return None
@@ -85,7 +88,7 @@ def get_bookmarks():
return bookmarks
except Exception as e:
print(f"Error: {e}")
logger.error(f"Error: {e}")
return None
@@ -117,7 +120,7 @@ def get_extensions_installed_from_shop():
manifest = json.load(file)
manifests.append(manifest)
except json.JSONDecodeError:
print(f"Error reading {manifest_path}")
logger.error(f"Error reading {manifest_path}")
return manifests

6
desktop_env/evaluators/metrics/docs.py
View File

@@ -4,6 +4,8 @@ from typing import List, Dict, Any
from docx import Document
from docx.enum.text import WD_PARAGRAPH_ALIGNMENT
import logging
logger = logging.getLogger("desktopenv.metric.docs")
def find_default_font(config_file_path, rules):
"""Find the default font in LibreOffice Writer."""
@@ -23,7 +25,7 @@ def find_default_font(config_file_path, rules):
for value in prop.findall('value', namespace):
default_font = value.text
except Exception as e:
print(f"Error: {e}")
logger.error(f"Error: {e}")
return 1 if default_font == expected_font else 0
@@ -192,4 +194,4 @@ def compare_contains_image(docx_file1, docx_file2):
# print(result)
# config_path = "/home/[username]/.config/libreoffice/4/user/registrymodifications.xcu"
# print(find_default_font("Ani", config_path))
# print(find_default_font("Ani", config_path))

6
desktop_env/evaluators/metrics/general.py
View File

@@ -18,13 +18,13 @@ def check_csv(result: str, rules: Dict[str, List[Dict[str, str]]]) -> float:
float
"""
expect_metrics = [False] * len(rules["except"])
expect_metrics = [False] * len(rules.get("expect", []))
unexpect_metric = True
with open(result) as f:
reader = csv.DictReader(f)
for rcd in reader:
for i, r in enumerate(rules["expect"]):
for i, r in enumerate(rules.get("expect", [])):
expect_metrics[i] = expect_metrics[i] or _match_record(r, rcd)
unexpect_metric = unexpect_metric and all(_match_record(r, rcd) for r in rules["unexpect"])
unexpect_metric = unexpect_metric and all(_match_record(r, rcd) for r in rules.get("unexpect", []))
return float(all(expect_metrics) and unexpect_metric)

14
desktop_env/evaluators/metrics/table.py
View File

@@ -10,6 +10,8 @@ from typing import Dict, List
from typing import Any, Union
from numbers import Number
import logging
logger = logging.getLogger("desktopenv.metric.table")
def compare_table(actual: str, expected: str, **options) -> float:
"""
@@ -32,7 +34,7 @@ def compare_table(actual: str, expected: str, **options) -> float:
df1 = pd.read_excel(expected)
df2 = pd.read_excel(actual)
metric: bool = df1.equals(df2)
print("Normal Contents Metric: {:}".format(metric))
logger.debug("Normal Contents Metric: {:}".format(metric))
features: List[str] = options.get("features", [])
for ftr in features:
@@ -43,12 +45,12 @@ def compare_table(actual: str, expected: str, **options) -> float:
sp1 = load_sparklines(actual)
sp2 = load_sparklines(expected)
new_metric: bool = sp1 == sp2
print("Sparkline Metric: {:}".format(new_metric))
logger.debug("Sparkline Metric: {:}".format(new_metric))
elif ftr=="chart":
charts1 = load_charts(workbook1, **options)
charts2 = load_charts(workbook2, **options)
new_metric: bool = charts1 == charts2
print("Chart Metric: {:}".format(new_metric))
logger.debug("Chart Metric: {:}".format(new_metric))
elif ftr=="number_format":
number_formats1: List[str] = [ c.number_format.lower()\
for col in workbook1.active.iter_cols()\
@@ -61,7 +63,7 @@ def compare_table(actual: str, expected: str, **options) -> float:
if c.data_type=="n"
]
new_metric: bool = number_formats1==number_formats2
print("Number Format Metric: {:}".format(new_metric))
logger.debug("Number Format Metric: {:}".format(new_metric))
else:
raise NotImplementedError("Unsupported xlsx feature: {:}".format(ftr))
metric = metric and new_metric
@@ -79,7 +81,7 @@ def check_sheet_list(result: str, rules: List[Dict[str, Any]]) -> float:
expected_name: str = worksheet_names[r["sheet_idx"]]
actual_name: str = r["sheet_name"]
metric: bool = expected_name == actual_name
print("Assertion: {:d}.{:} is {:} - {:}".format(r["sheet_idx"], actual_name, expected_name, metric))
logger.debug("Assertion: {:d}.{:} is {:} - {:}".format(r["sheet_idx"], actual_name, expected_name, metric))
passes = passes and metric
elif r["type"] == "sheet_data":
if isinstance(r["sheet_idx0"], int):
@@ -99,7 +101,7 @@ def check_sheet_list(result: str, rules: List[Dict[str, Any]]) -> float:
sheet_idx: int = int(sheet_idx)
df2: pd.DataFrame = pd.read_excel(file_name, sheet_idx)
metric: bool = df1.equals(df2)
print("Assertion: {:} == {:} - {:}".format(r["sheet_idx0"], r["sheet_idx1"], metric))
logger.debug("Assertion: {:} == {:} - {:}".format(r["sheet_idx0"], r["sheet_idx1"], metric))
passes = passes and metric
else:
raise NotImplementedError("Unimplemented sheet check: {:}".format(r["type"]))

3
desktop_env/evaluators/metrics/utils.py
View File

@@ -11,6 +11,9 @@ from openpyxl.chart._chart import ChartBase
from typing import Dict, List, Set
from typing import Any
import logging
logger = logging.getLogger("desktopenv.metrics.utils")
_xlsx_namespaces = [ ("x14", "http://schemas.microsoft.com/office/spreadsheetml/2009/9/main")
, ("xm", "http://schemas.microsoft.com/office/excel/2006/main")
]

13
desktop_env/evaluators/metrics/vlc.py
View File

@@ -5,6 +5,9 @@ from xml.etree import ElementTree
import pygetwindow as gw
import pyautogui
import logging
logger = logging.getLogger("desktopenv.metrics.vlc")
def read_vlc_config(setting_name):
"""
Reads the VLC configuration file to check for a specific setting.
@@ -24,7 +27,7 @@ def read_vlc_config(setting_name):
config_path = paths.get(os_type)
if not config_path or not os.path.exists(config_path):
print("VLC config file not found for this operating system.")
logger.warning("VLC config file not found for this operating system.")
return None
try:
@@ -33,7 +36,7 @@ def read_vlc_config(setting_name):
if line.startswith(setting_name):
return line.strip()
except IOError as e:
print(f"Error reading config file: {e}")
logger.error(f"Error reading config file: {e}")
return None
@@ -53,7 +56,7 @@ def get_vlc_playing_info(host='localhost', port=8080, password='password'):
return status, file_info
return status, None
except Exception as e:
print(f"Error: {e}")
logger.error(f"Error: {e}")
return None, None
@@ -78,10 +81,10 @@ def is_vlc_fullscreen():
except IndexError:
# VLC window not found
print("VLC window not found.")
logger.error("VLC window not found.")
return False
except Exception as e:
print(f"An error occurred: {e}")
logger.error(f"An error occurred: {e}")
return False

10
desktop_env/server/main.py
View File

@@ -16,17 +16,19 @@ app = Flask(__name__)
pyautogui.PAUSE = 0
pyautogui.DARWIN_CATCH_UP_TIME = 0
logger = app.logger
@app.route('/setup/execute', methods=['POST'])
@app.route('/execute', methods=['POST'])
def execute_command():
data = request.json
# The 'command' key in the JSON request should contain the command to be executed.
command = data.get('command', [])
shell = data.get('shell', False)
command = data.get('command', "" if shell else [])
# Execute the command without any safety checks.
try:
result = subprocess.run(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
result = subprocess.run(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=shell, text=True)
return jsonify({
'status': 'success',
'output': result.stdout,
@@ -94,7 +96,7 @@ def capture_screen_with_cursor():
# Use the screencapture utility to capture the screen with the cursor
subprocess.run(["screencapture", "-C", file_path])
else:
print(f"The platform you're using ({user_platform}) is not currently supported")
logger.warning(f"The platform you're using ({user_platform}) is not currently supported")
return send_file(file_path, mimetype='image/png')
@@ -190,7 +192,7 @@ def download_file():
return "File downloaded successfully"
except requests.RequestException as e:
print(f"Failed to download {url}. Retrying... ({max_retries - i - 1} attempts left)")
logger.error(f"Failed to download {url}. Retrying... ({max_retries - i - 1} attempts left)")
return f"Failed to download {url}. No retries left. Error: {e}", 500