edgelens-robot

Robot Framework library for the EdgeLens Edge AI Device Farm. Keyword- driven wrapper around the Python SDK — write Request Device, Deploy Model, Run Inference, and let Robot drive the camera.

Counterpart to edgelens-pytest for pytest users and @edgelens/playwright-plugin for JS/TS.

Install

pip install edgelens-robot

Requires Python 3.9+ and Robot Framework 6.0+.

Quickstart

*** Settings ***
Library             EdgeLensLibrary    api_key=%{EDGELENS_API_KEY}
Suite Teardown      Release Device

*** Test Cases ***
Camera Renders The Test Pattern
    Request Device    os=stock-os
    Deploy Model    ./face.cvimodel
    Start Inference
    ${stats}=    Run Inference    duration=30
    Should Be True    ${stats}[fps][avg] > 15

Run it:

EDGELENS_API_KEY=el_test_xxx robot tests/

Keywords

Session lifecycle

Device operations

Display

Inference (stock OS only)

Library scope

Default is SUITE — one camera per suite, shared across every test in the file. The right tradeoff for read-only suites where allocation takes 30s+. Override on the Library line:

Library    EdgeLensLibrary    scope=TEST     # fresh camera per test
Library    EdgeLensLibrary    scope=GLOBAL   # one camera across the whole run

Configuration

CI

The edgelensai/edgelens-sdk image ships Robot Framework — drop the library in and you're set:

GitHub Actions

- uses: edgelens-ai/edgelens-sdk/ci/github-action@v0.2.1
  with:
    api-key: ${{ secrets.EDGELENS_API_KEY }}
    command: robot --output output.xml tests/
    junit-file: output.xml

GitLab CI

include:
  - remote: "https://raw.githubusercontent.com/edgelens-ai/edgelens-sdk/main/ci/gitlab/edgelens.gitlab-ci.yml"

edgelens-robot:
  extends: .edgelens:robot

Develop

python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"
pytest

See also

Pairs with edgelens-pytest and @edgelens/playwright-plugin.