G4 Studio Test Wright – Run Your G4 Bots as Tests in VS Code

Discover, plan, run, and report your G4 bots — right inside VS Code.

License

Overview

G4 Studio Test Wright turns the bots of a G4 project into a test suite.

Every bot file under the project's bots/ folder is a test, and every subfolder is a suite. Test Wright finds them, lets you build a run plan by dragging bots and suites into it, runs the plan against the G4 engine, and saves every run with its G4 reports and an Allure report — no Python, no Java, no extra tools.

It is the VS Code sibling of the desktop G4 Test Wright workbench, and it works side by side with the G4 Engine extension.

Get Started

Open a G4 project (the folder that holds manifest.json and bots/) in VS Code.
Make sure the G4 engine is running.
Click the G4 Test Wright flask in the Activity Bar — Test Wright opens in a tab. (With a tab open, the flask shows the side bar, whose Open Test Wright button opens more tabs.)
Press Discover Bots, drag bots into the Testing Plan, and press Run.

👉 New to Test Wright? The Run Your Bots learning path walks you through it, click by click.

Documentation

📗 Run Your Bots — discover bots, build and save plans, run them, read the results, open Allure reports, and schedule runs.
🧩 Webview components — for contributors: how the pages are built from g4-* components.

Key Capabilities

Bots are tests — .json, .g4, and .g4bot files under bots/ are discovered as tests; subfolders become suites. Tags come from each bot's reference.tags.
Drag-and-drop plans — drag bots or whole suites into the plan, reorder them, tick the ones to run, and save the plan to the repository or a file.
Runs against the G4 engine — each bot is sent to the engine with the project's manifest.json filling in what the bot leaves out. Run one bot at a time or in parallel, and stop a run at any time.
Environments — pick one of the workspace environments/ files; it is sent to the engine and set as the bots' default environment.
Honest statuses — a bot fails only when an assertion fails. An exception without a failed assertion is a warning (Allure broken).
Run history — every run is saved under reports/<run>/ with a run.json summary and one G4 report per bot. Open any run, drill into a bot's report, or drag the run back into the plan to rerun it.
Allure 3 built in — runs write Allure results to reports/allure-<run>, and Test Wright builds the Allure 3 report for one run or merges all of them, in a VS Code tab or your browser. No Java needed.
Scheduling — schedule up to five runs; they fire even when the Test Wright tab is closed.

How It Works

Test Wright is one VS Code webview page backed by services in the extension host:

Test Wright page (webview)          Extension host
──────────────────────────          ─────────────────────────────────────────────
Bots tree, plan, repositories  ⇄   Discovery      bots/ → tests and suites
                                    Request build  bot + manifest + environment
                                    Run service    POST /api/v4/g4/automation/invoke
                                    Evaluator      response → passed/failed/warning/broken
                                    Repositories   reports/<run>/run.json, *.g4rpt
                                    Allure         allure-<run>/results → Allure 3 report
                                    Scheduler      timed runs (kept in the workspace)

What is sent to the engine

For each bot, Test Wright builds the request from the bot file and manifest.json:

The manifest token always wins (authentication.token).
For everything else the bot wins; manifest values only fill keys the bot leaves out (authentication, driverParameters, settings).
Client-only manifest sections (clientLogConfiguration, clientLogSettings, clientReportSettings, recorderSettings, queueManagerSettings) are never sent.
The selected environment is uploaded to the engine and set as settings.environmentsSettings.defaultEnvironment.

The bot file itself is never changed.

How a status is decided

Status	When
Passed	The engine answered and no assertion failed.
Failed	At least one assertion evaluated to `false`.
Warning	No assertion failed, but a plugin raised an exception. Shown as broken in Allure.
Broken	The engine could not be reached, or answered with an error or an empty response.
Canceled	The run was stopped before the bot finished.

Run Folder

Every run creates one folder under the project's reports folder, and its Allure files go to a sibling folder with the allure- prefix:

reports/
├── 2026-10-08-213643907/          # run name: yyyy-MM-dd-HHmmssfff
│   ├── run.json                   # run summary: settings, tally, one entry per bot
│   ├── a.g4rpt                    # G4 automation report per bot (Base64 JSON)
│   └── smoke.tagged.g4rpt
├── allure-2026-10-08-213643907/
│   ├── results/                   # written when Collect is on
│   └── report/                    # built when you open the Allure report
└── allure-merged/                 # built when you open the report of several runs

Right-click a run.json in the Explorer for Open Run Report and Convert Run to Allure Results.

Project folders

When VS Code starts in a G4 project (a manifest.json is found), Test Wright creates the folders it writes to if they are missing:

Folder	Holds
`<project>/.repositories`	Saved run plans
`<project>/reports`	Runs and their Allure folders. The `settings.clientReportSettings.reportsFolder` manifest setting moves it to `<that folder>/reports`.

The project folder is the workspace folder, or its src folder when manifest.json is there.

Project Structure

g4-test-wright-vscode-extension/
├── languages/                  # g4log grammar
├── resources.assets/           # icons
├── resources.components/       # webview pages and g4-* components (see its README)
├── resources.docs/             # learning paths
├── resources.fonts/
└── src/
    ├── clients/                # G4 engine HTTP client
    ├── commands/               # VS Code commands (CommandBase pattern)
    ├── constants/
    ├── extensions/             # utilities, webview page and component helpers
    ├── logging/
    ├── models/
    ├── providers/              # Test Wright panel, run report tabs
    ├── services/               # discovery, request builder, evaluator, run, repositories, Allure, scheduler
    └── extension.ts

Develop

npm install
npm run compile      # webpack build into dist/
npm run lint

Press F5 in VS Code to start an Extension Development Host. In that host, G4 Test Wright: Show Component Gallery (Development) shows every component on one page.

Requirements

Visual Studio Code ^1.105.0
A compatible G4 Engine backend — local (via the G4 sandbox) or remote
Node.js — only for extension development/build

G4 Studio Test Wright

g4-api