Scale computer-use 2.0 with open-source drivers, cross-OS sandboxes, and benchmarks for training, evaluation, and data generation.
Rust agentai-agentapplecomputer-usecomputer-use-agentcontainerizationcuadesktop-automationhacktoberfestlumemacosmanusoperatorswiftvirtualizationvirtualization-frameworkwindowswindows-sandbox
"""Unit tests for the cli-shape lane (stdlib only)."""
from __future__ import annotations
import unittest
import cli_shape as cs
SPEC = {
"cua": {
"name": "cua",
"global_options": [
{"name": "json", "takes_value": False},
{"name": "daemon", "type": "DAEMON", "takes_value": True},
],
"commands": [
{
"name": "sandbox",
"aliases": ["sb"],
"usage": "cua sandbox [OPTIONS] <COMMAND>",
"arguments": [],
"options": [],
"flags": [],
"subcommands": [
{
"name": "create",
"usage": "cua sandbox create [OPTIONS] [IMAGE] [-- <COMMAND>...]",
"arguments": [
{"name": "IMAGE", "is_optional": True},
{"name": "COMMAND", "is_optional": True, "repeatable": True},
],
"options": [
{"name": "on", "type": "ON", "takes_value": True},
{
"name": "name",
"short_name": "n",
"type": "NAME",
"takes_value": True,
},
{"name": "format", "type": "md | json", "takes_value": True},
],
"flags": [
{"name": "vm", "short_name": "y"},
{"name": "x", "short_name": "x"},
],
"subcommands": [],
},
{
"name": "exec",
"usage": "cua sandbox exec [OPTIONS] <NAME> <COMMAND>...",
"arguments": [
{"name": "NAME", "is_optional": False},
{"name": "COMMAND", "is_optional": False, "repeatable": True},
],
"options": [],
"flags": [],
"subcommands": [],
},
{
"name": "mcp",
"usage": "cua sandbox mcp [OPTIONS] <NAME> <SERVICE> <COMMAND>",
"arguments": [
{"name": "NAME", "is_optional": False},
{"name": "SERVICE", "is_optional": False},
],
"options": [],
"flags": [],
"subcommands": [
{
"name": "tools",
"arguments": [],
"options": [],
"flags": [],
"subcommands": [],
}
],
},
],
}
],
},
"cua-driver": {
"name": "cua-driver",
"commands": [
{"name": "serve", "arguments": [], "options": [], "flags": [], "subcommands": []}
],
},
}
def problems(code: str, lang: str = "bash") -> list[str]:
return cs.check_block(code, lang, SPEC)[1]
class CliShapeTest(unittest.TestCase):
def test_valid_commands(self):
for code in [
"cua sb create linux --name dev --on cloud",
"cua --json sb create linux -n dev -- python -m http.server 8000",
"cua sandbox create linux --format=json -yx",
"cua sb exec dev uname -a",
"cua sb exec <name> ls -la /tmp | head",
"cua sb mcp web mcp tools",
"X=1 sudo cua sb create \\\n linux --vm",
"echo hi && cua sb create linux; cua sb --help",
'TOKEN="$(cua --json sb create linux)"',
"cua-driver serve",
"cua-driver list_apps '{}'",
"cat <<'EOF' > x\ncua sb bogus\nEOF\ncua sb create linux",
"# a comment\ncua --version",
]:
with self.subTest(code=code):
self.assertEqual(problems(code), [])
def test_invalid_commands(self):
for code, needle in [
("cua sb creat linux", "unknown command 'creat'"),
("cua sb create linux --nope", "unknown option --nope"),
("cua sb create linux --name", "needs a value"),
("cua sb create linux --format yaml", "takes md | json"),
("cua sb create linux --vm=1", "takes no value"),
("cua sb exec", "missing <NAME>"),
("cua sb", "missing subcommand"),
("cua", "missing command"),
("cua sb create a b", "unexpected argument"),
("cua-driver sreve", "unknown command"),
]:
with self.subTest(code=code):
found = problems(code)
self.assertTrue(any(needle in p for p in found), found)
def test_console_blocks_check_prompted_lines_only(self):
code = "$ cua sb create linux\ncreated dev\n$ cua sb ls\nNAME STATUS\n"
self.assertEqual(
cs.check_block(code, "console", SPEC),
(
2,
[
"'cua sb ls': cua sandbox: unknown command 'ls' (expected one of create, exec, mcp)"
],
),
)
def test_powershell_continuation(self):
self.assertEqual(problems("cua sb create linux `\n --name dev", "powershell"), [])
def test_blocks_without_invocations(self):
self.assertEqual(cs.check_block("docker ps\nls cua", "bash", SPEC), (0, []))
if __name__ == "__main__":
unittest.main()