iints-sdk-python35 0.1.9

A pre-clinical Edge-AI SDK for diabetes management validation.

IINTS-AF SDK

Intelligent Insulin Titration System for Artificial Pancreas

IINTS-AF is a safety-first simulation and validation platform for insulin dosing algorithms. It lets you test AI or classical controllers on virtual patients, enforce deterministic safety constraints, and generate audit-ready clinical reports before anything touches a real patient.

In one session you can:

• Run a clinic-safe preset and compare against PID and standard pump baselines
• Import real-world CGM CSV into a standard schema + scenario JSON
• Use the bundled demo CGM data pack (zero setup)
• Export a clean PDF report plus full audit trail (JSONL/CSV)
• Stress-test sensor noise, pump limits, and human-in-the-loop interventions
• Generate patient profiles with ISF/ICR + dawn phenomenon

Who it’s for:

• Diabetes researchers and clinicians validating new control strategies
• ML engineers benchmarking AI controllers with medical safety rails
• Developers building decision-support systems for closed-loop insulin delivery

Installation

Install the SDK directly via PyPI:

pip install iints-sdk-python35

Quick Start (CLI)

iints quickstart --project-name iints_quickstart
cd iints_quickstart
iints presets run --name baseline_t1d --algo algorithms/example_algorithm.py

One-line full run (CSV + audit + PDF + baseline):

iints run-full --algo algorithms/example_algorithm.py \
  --scenario-path scenarios/clinic_safe_baseline.json \
  --output-dir results/run_full

Import real-world CGM data:

iints import-data --input-csv data/my_cgm.csv --output-dir results/imported

Try the bundled demo data pack:

iints import-demo --output-dir results/demo_import

Or run the full demo workflow (import + run + report) in one script:

python3 examples/demo_quickstart_flow.py

Quick Start (Python)

import iints
from iints.core.algorithms.pid_controller import PIDController

outputs = iints.run_simulation(
    algorithm=PIDController(),
    scenario="scenarios/example_scenario.json",
    patient_config="default_patient",
    duration_minutes=720,
    seed=42,
    output_dir="results/quick_run",
)

Notebook Guide

Hands-on Jupyter notebooks live in examples/notebooks/

• Quickstart end-to-end run
• Presets + scenario validation
• Safety supervisor behavior
• Audit trail + PDF report export
• Baseline comparison + clinical metrics
• Sensor/pump models + human-in-the-loop
• Optional Torch/LSTM usage
• Ablation study (with/without Supervisor)

Documentation

• Product manual: SDK_COMPREHENSIVE_GUIDE.md
• Notebook index: examples/notebooks/README.md
• Technical README: TECHNICAL_README.md
• API Stability: API_STABILITY.md

Ethics & Safety

This SDK is for research and validation. It is not a medical device and does not provide clinical dosing advice.

“Code shouldn’t be a secret when it’s managing a life.” — Bobbaers Rune