First Run

Use this page to verify the SDK and create one small, deterministic simulation.

Time: about five minutes after Python is installed.

Result: a folder containing a time series, report, metadata, and reproducibility information.

1. Create An Environment

macOS or Linux:

python3 -m venv .venv
source .venv/bin/activate

Windows PowerShell:

py -m venv .venv
.venv\Scripts\Activate.ps1

2. Install IINTS-AF

python -m pip install --upgrade pip
python -m pip install "iints-sdk-python35[full,mdmp]"

3. Verify The Installation

iints --version
iints doctor --smoke-run

Resolve blocking errors before continuing. Warnings about optional hardware or AI components are expected when those components are not installed.

4. Run The Demo

iints demo quick --output-dir results/first_run

The command uses a bundled patient, scenario, clinical baseline algorithm, and deterministic seed. It does not connect to a person or medical device.

5. Inspect The Result

List the generated files:

ls results/first_run

On Windows:

Get-ChildItem results\first_run

You should see at least:

• results.csv
• report.pdf
• run_manifest.json

The command may also create metadata, an audit trail, an evidence bundle, and an editable example algorithm.

Open report.pdf, but also inspect results.csv and the metadata. A graph is a summary, not the complete evidence.

6. Confirm Reproducibility

Run the same preset with an explicit seed in a second folder:

iints demo quick --seed 42 --output-dir results/repeat_run

The run metadata should record the same controlled inputs. See Understand A Run for the correct review order.

If It Fails

iints doctor --full --suggest
iints demo quick --dry-run --output-dir results/first_run

Then use Troubleshooting. Include the command, SDK version, Python version, operating system, and complete error when reporting a problem.

Next Step

Read Core Concepts, then complete the Complete First Workflow with your own project folder.