Skip to content

v1.5.10

v1.5.10 is the EUCYS demo and physiology-release patch. It makes the public demo shortcuts available from the released package, improves the updater feedback, and folds the new stress/Hovorka physiology work into the SDK.

Highlights

1. EUCYS demo shortcut in the release wheel

The released CLI now supports the audience-specific demo shortcuts directly:

iints demo doctor
iints demo eucys
iints demo booth

This fixes the mismatch where the source tree already supported iints demo eucys, but an installed release could still show Usage: iints demo [OPTIONS] and reject the positional preset.

2. Better iints update feedback

iints update now verifies the installed package version after pip completes and prints the fallback command for unreleased GitHub main builds:

iints update --yes
iints update --source github --yes

Use the GitHub source only when you need a feature before the PyPI release workflow has finished publishing.

3. Hovorka patient model

The SDK now includes an experimental 19-state Hovorka-style patient model for research simulations:

from iints.core.patient.patient_factory import PatientFactory

patient = PatientFactory.create_patient(patient_type="hovorka")

The model includes glucose compartments, gut absorption, subcutaneous insulin depots, plasma insulin, insulin-action states, exogenous glucagon depots/plasma/effect, HAAF memory, and an exercise-driven GLUT4/NIMGU state. It also supports stress and exercise modifiers through the same patient interface used by the simulator.

The dawn/circadian EGP term is explicitly gated by dawn_phenomenon_strength, so the Hovorka-style model does not add hidden always-on circadian drift in default runs.

4. Stress/illness physiology events

Scenario files can now include physiological stress/illness events:

{
  "start_time": 1080,
  "event_type": "stress",
  "value": 0.6,
  "duration": 120
}

Stress mode reduces insulin sensitivity and raises endogenous glucose production in supported patient models. This is useful for research demos where glucose rises without a meal because of illness, cortisol/adrenaline effects, or other stress physiology.

5. More realistic built-in presets

The built-in presets have been recalibrated so they avoid impossible one-minute spikes and flat synthetic plateaus while still showing realistic free-living variation. The reference/free-living presets now include a mild stress physiology block, and overnight risk presets include enough physiological variation to be useful for research/demo discussions.

6. Publication-style plots

AGP/report plots and cockpit visuals now use cleaner, publication-oriented styling with high-DPI output, colorblind-conscious palettes, lighter target bands, and reduced visual noise. AGP asset export now also writes vector SVG files beside the PNGs, and the report extras include SciencePlots for reproducible publication-style Matplotlib rendering.

7. Dependency hardening

The FastAPI dependency now excludes the 0.136.3 build flagged by pip-audit, so fresh installs should avoid that audit failure while still accepting compatible stable FastAPI releases.

8. Source-backed physiology documentation

The release notes and docs now include a clearer source trail for the mathematical and physiological models used or planned in the SDK:

  • Bergman minimal-model insulin sensitivity and remote insulin action.
  • Hovorka-style glucose, insulin, and insulin-action compartments for research simulations.
  • Dalla Man meal disturbance modeling for post-prandial dynamics.
  • Experimental hypoglycemia counterregulation and HAAF memory in the Hovorka-style model.
  • Experimental exogenous glucagon PK/PD for dual-hormone pump research, with simulator safety caps.
  • Experimental long-memory CGM noise, GLP-1 gastric-emptying feedback, optional dawn/circadian EGP, and GLUT4/NIMGU exercise physiology.
  • Renal glucose handling as an experimental smooth threshold/splay-style model for high-glucose clearance realism.

See Physiology Reference, Hypoglycemia Science Model, Evidence Base, and Complete Source Library.

Upgrade

iints update --yes
hash -r
iints --version
iints demo eucys --dry-run

If PyPI has not finished publishing yet:

iints update --source github --yes
hash -r
iints --version

Validation

Release checks include targeted physiology, preset realism, demo shortcut, update dry-run, docs, and packaging validation.