Quick scenario import (advanced, 10 min)

This tutorial is for contributors and engineers who develop scenarios or automate imports in CI/CD. To upload a ZIP file directly in Portal, use Import a hands-on scenario in one shot (5 min).

By the end of this tutorial

• Install the current dhub2-examples repository and Python dependencies.
• Configure Manager API authentication safely.
• Import the retail scenario's metadata, code sources, and data files.
• Verify the CLI summary and the resulting Portal resources.

1. Prepare the repository and Python environment

Run these commands in a new working directory:

git clone https://github.com/dtonic/dhub2-examples.git
cd dhub2-examples
python -m venv .venv

Activate the virtual environment:

# macOS / Linux
source .venv/bin/activate

# Windows PowerShell
.venv\Scripts\Activate.ps1

On a Korean Windows console, the CLI can raise UnicodeEncodeError when printing its ✓ status marker. Enable UTF-8 mode in the same PowerShell window first:

$env:PYTHONUTF8='1'

In Command Prompt (cmd), run set PYTHONUTF8=1 instead.

Install the two packages used by the current CLI. There is no tools/requirements.txt to install.

python -m pip install httpx python-dotenv

2. Configure the Manager API and authentication

Create .env in the dhub2-examples repository root, where the tools and scenarios directories are visible. To sign in to the production D.Hub with email and password, use:

DHUB2_API_URL=https://manager.hub.dtonic.io/api/v1
DHUB2_EMAIL=<your-email>
DHUB2_PASSWORD=<your-password>

If you already have a bearer token, use it instead of email and password:

DHUB2_API_URL=https://manager.hub.dtonic.io/api/v1
DHUB2_TOKEN=<your-token>

DHUB2_API_URL is the Manager API base URL, including /api/v1, not the Portal URL. The .env file contains credentials, so never commit it to Git.

3. Import the scenario

Start with a dry run to inspect the files and execution order without making API calls:

python tools/import.py scenarios/retail_inventory_intelligence --dry-run

If the expected items appear, run the actual import:

python tools/import.py scenarios/retail_inventory_intelligence

The CLI performs the following work in dependency order:

1. 1 collection
2. 3 datasets and 2 Code items
3. 4 ontology Entities and 3 Relations
4. 2 Pipelines and 1 Dashboard
5. 1 Agent Tool and 1 Agent
6. 2 Code source uploads and 3 Parquet data uploads

In a clean environment, the current final summary is:

Created : 23
Skipped : 0
Failed  : 0

The count of 23 includes 18 resources, two Code source uploads, and three data uploads.

4. Verify the result in Portal

Open the retail_inventory_intelligence collection in Portal and verify:

• 3 datasets: retail_inventory, competitor_prices, and update_competitor_price
• 2 Code items and 2 Pipelines
• the inventory_overview Dashboard
• 4 ontology Entities and 3 Relations

The Data tabs of retail_inventory, competitor_prices, and update_competitor_price contain rows immediately after import. update_competitor_price is not created by a later Pipeline run; the importer creates the dataset and uploads its data as part of this scenario import. Open retail-assistant from the Agents list and confirm that its detail page links the ontology_graph_query tool in the same collection scope.

Continue with Retail Inventory Intelligence to use the imported resources.