Files
Artemis-Eapil/packages/artemis-client/README.md
T
chenmingxuan be6e8c2abd feat: release 2.2.0 with PC-only tests and execution reliability fixes
Unify task scheduling and queue state, fix Android input and observation recovery, preserve screenshot paths, and match final review criteria accurately. Add collapsible copyable results and publish the complete 2.1.0 upgrade notes. Validation: 3189 backend tests, 326 frontend tests, 24 client tests, production build and quality gates.
2026-10-27 01:13:41 +08:00

158 lines
5.1 KiB
Markdown

# Artemis Client
`artemis-client` is the remote Python SDK that runs on a developer machine. It
communicates with the Artemis daemon on a device host over HTTP and contains no
ADB, device-driver, agent, LLM, or image-processing code.
Runtime dependencies: **zero**. HTTP, JSON, and asynchronous adapters use only
the Python standard library.
Artemis 2.2.0 queues computer and Android tasks for serial execution. Legacy
concurrency options cannot enable parallel execution on that server version.
## Installation
Install from a local checkout:
```powershell
uv add --editable C:\path\to\artemis\packages\artemis-client
```
Install from the Git repository subdirectory:
```powershell
uv add "artemis-client @ git+https://github.com/google/artemis.git#subdirectory=packages/artemis-client"
```
## Connect to a Device Host
The current Artemis daemon administration API does not include built-in remote
authentication. Do not expose its port directly to the public internet. During
the initial rollout, connect through an SSH tunnel:
```powershell
ssh -L 8000:127.0.0.1:8000 user@artemis-host
```
Then connect the SDK to the local tunnel endpoint:
```python
from artemis_client import ArtemisClient
client = ArtemisClient("http://127.0.0.1:8000")
```
For production, place a reverse proxy in front of the daemon that enforces HTTPS
and bearer-token authentication and exposes only task-related APIs.
## Usage
PC-only tasks do not need Android or an emulator online. Enable and authorize a
computer observation plugin on the host first, then use the same target as the
web launcher's “仅 PC” checkbox:
```python
handle = await client.submit(
"Open the desktop browser and inspect the requested page",
profile="flash", # "pro" is also supported
execution_target="computer",
)
```
`run()` accepts the same argument. For task-like objects, use
`Task(goal="Check desktop", execution_target="computer")` from the full runtime
and pass it to `client.run_task()`. A saved Android device on the client is ignored
for PC-only tasks. Per-task `options` are sent as the server's `runtime_overrides`,
for example `options={"execution_target":"computer","plugins":[{"id":"cua-driver","profile":"default"}]}`.
Plugin IDs must exist on the host; omitting `plugins` selects enabled authorized
defaults for the target. Mixed tasks still require Android.
For interactive case design, use `client.exploration` with the same actions as the
web workspace and `mobile_exploration`. Use a stable request ID when retrying a write:
```python
session = await client.exploration(
"create",
request_id="explore-settings-create",
payload={"goal": "Explore the Settings search entry; ask when expectations are unclear"},
)
await client.exploration(
"start",
exploration_id=session["id"],
request_id="explore-settings-start",
)
state = await client.exploration("get", exploration_id=session["id"])
```
The default exploration model is `gpt-6.1-sol / high`. Timeouts preserve progress
and pause; `completed` does not imply a passing test. Read `runs[].test_result`.
See [exploration modes and actions](../../docs/exploration-mode.md) for manual
demonstrations, versioned cases and reruns.
```python
import asyncio
import os
from artemis_client import ArtemisClient
async def main() -> None:
client = ArtemisClient(
os.environ["ARTEMIS_BASE_URL"],
token=os.environ.get("ARTEMIS_TOKEN"),
)
handle = await client.submit(
"Open the Battery page in Settings and verify the battery percentage",
profile="flash",
device_serial="emulator-5554",
)
print("queued:", handle.task_id)
result = await client.wait_for_task(handle.task_id, timeout=600)
if not result.succeeded:
raise RuntimeError(result.error or result.status)
asyncio.run(main())
```
You can also submit a task and wait for it in one call:
```python
result = await client.run("Open Android Settings and check the Wi-Fi status")
```
Main methods:
- `health()` quickly checks whether the server and scheduler are alive.
- `readiness()` runs full device and toolchain diagnostics and may take longer.
- `capabilities()` negotiates API features and falls back for older servers.
- `list_devices()` lists devices available on the host.
- `submit()` submits an idempotent task and returns immediately.
- `get_task()` retrieves a task.
- `wait_for_task()` waits for a terminal state.
- `run()` submits a task and waits for completion.
- `stop()` stops one task.
By default, `submit()` generates a UUID on the client and sends it as the current
server's `session_id`. Network retries reuse the same `task_id`, preventing the
server from creating duplicate tasks.
## Compatibility
Version 0.1 supports these current Artemis daemon endpoints:
- `POST /api/run`
- `GET /api/sessions/{session_id}`
- `GET /api/status`
- `POST /api/stop`
- `GET /api/devices`
- `GET /api/system/readiness`
Capability discovery first attempts the future `GET /api/v1/capabilities`
endpoint. If an older server returns 404, the SDK uses its known baseline
capability set.
See [DESIGN.md](DESIGN.md) for the architecture and protocol evolution plan.