Files
sci-gui-agent-benchmark/evaluation_examples/examples/multi_apps/26660ad1-6ebb-4f59-8cba-a8432dfe8d38.json
yuanmengqi e433f35c1f feat: standardize configuration fields across all evaluation examples
- Add `fixed_ip` field to all 369 JSON files in examples directory
  - Set to `true` for 8 files listed in google_chrome.json multi_apps
  - Set to `false` for remaining 361 files
- Add `possibility_of_env_change` field to 363 JSON files missing this field
  - Set to "low" for newly added fields
  - Preserve existing values (4 medium, 2 high) for 6 files that already had this field

This ensures consistent configuration schema across all evaluation examples
while maintaining backward compatibility with existing settings.
2025-07-16 13:45:34 +00:00

77 lines
1.8 KiB
JSON

{
"id": "26660ad1-6ebb-4f59-8cba-a8432dfe8d38",
"snapshot": "multiapps",
"instruction": "I want to test the quality of the network environment my laptop is currently in. Please measure my network situation through speedtest.net, copy the results in speedtest.net/results, and save them to ~/Test/Speed/results.txt (if the dir does not exist, create it). Each metric occupies one line, with the metric name and its value separated by a single space.",
"source": "https://www.speedtest.net/",
"config": [
{
"type": "launch",
"parameters": {
"command": [
"google-chrome",
"--remote-debugging-port=1337"
]
}
},
{
"type": "launch",
"parameters": {
"command": [
"socat",
"tcp-listen:9222,fork",
"tcp:localhost:1337"
]
}
},
{
"type": "chrome_open_tabs",
"parameters": {
"urls_to_open": [
"https://www.speedtest.net/"
]
}
},
{
"type": "activate_window",
"parameters": {
"window_name": "Google Chrome"
}
},
{
"type": "execute",
"parameters": {
"command": [
"python",
"-c",
"import pyautogui; import time; time.sleep(0.5);"
]
}
}
],
"trajectory": "trajectories/",
"related_apps": [
"os",
"browser"
],
"evaluator": {
"func": "file_contains",
"result": {
"type": "vm_file",
"path": "/home/user/Test/Speed/results.txt",
"dest": "results.txt"
},
"expected": {
"type": "rule",
"rules": {
"expected": [
"Ping",
"Download",
"Upload"
]
}
}
},
"proxy": true,
"fixed_ip": false,
"possibility_of_env_change": "low"
}