Command Reference

Use this page when you know the task and want the public command family quickly, without reading the full technical manual.

Read before: Choose Your Path if you are not sure which command family you need.

Need the shortest practical list? Use the Command Cheatsheet.

Read next: Technical Reference for deeper integration details.

Command Navigation

iints menu (or iints hub)

Choose a maintained command domain and print reliable entry points without invoking a workflow automatically.

iints map (or iints overview)

Print the maintained command domains. Typer-generated --help remains authoritative for options.

Beginner-Friendly Entry Points

iints guide

Run the short starter demo or open command navigation.

iints start

Print a goal-based first-run plan, or run the safe starter action.

Common forms:

iints start
iints start --goal project --run
iints start --goal edge
iints start --goal data

iints onboard

Show the one recommended path from machine check to first study bundle.

iints onboard
iints onboard --run-safe-steps

The safe mode runs doctor, a full demo, demo-data import, and a realism check, then prints the two study commands you can run once you choose an algorithm.

iints demo

The main live-demo button. It exports showable code, prints the presenter story, runs the demo in audience-safe stage mode, writes a technical log off-screen, and produces the cue card, artifact map, presenter guide, poster, and rerun script. By default it also builds evidence_bundle/ with a public README, model card, run index, and copied proof artifacts.

Common forms:

iints demo
iints demo doctor
iints demo eucys
iints demo booth
iints demo --audience jury --output-dir results/live_demo
iints demo --audience clinical
iints demo --dry-run          # rehearsal/preflight only, not the live call
iints demo --no-evidence      # fast rehearsal without the public proof bundle
iints demo --technical        # show raw execution detail for debugging
iints demo --simulation-only --quick
iints demo --simulation-only --full

Story presets:

• iints demo doctor starts with a clinical safety discussion: virtual patient, normal day, meal stress, risky context, and supervisor decision.
• iints demo eucys frames the run as one experiment with a research question, hypothesis, three runs, and evidence bundle.
• iints demo booth frames the run as a public digital-patient story: scenario changes, algorithm suggestion, safety check, and visual proof.

iints demo-live

Explicit alias for the same live presentation engine used by iints demo.

It exports showable Python code, prints an audience-aware opening talk track, runs the live demo, writes PRESENTER_GUIDE.md, DEMO_STORY.md, DEMO_CUE_CARD.md, DEMO_ARTIFACTS.md, and RUN_LIVE_DEMO.sh, then lists the poster plus proof artifacts to open next. Non-SDK story modes keep code as the proof layer instead of the first thing shown.

Common forms:

iints demo-live
iints demo-live --output-dir results/live_demo
iints demo-live --no-run
iints demo-live --prepare-ai
iints demo-live --audience clinical
iints demo-live --audience engineering
iints demo-live --story doctor
iints demo-live --story eucys
iints demo-live --story booth

iints quickstart

Create a ready-to-run project folder.

iints quickstart --project-name iints_quickstart

The generated project is self-contained: it includes patients/stable_patient.yaml, a scenario file, and an editable starter algorithm so you can run locally without depending on packaged patient assets.

iints run --wizard

Interactive custom run builder.

iints version

Compare the active Python environment with the latest stable PyPI release.

iints version --refresh
iints version --offline --json
iints version --refresh --fail-if-outdated --fail-if-unknown

The report distinguishes installed distribution metadata from active source code and includes the Python executable, CLI path, package location, release source, and check time. Exit code 2 means a verified update is available; exit code 3 means release status could not be verified when --fail-if-unknown is active.

iints update

Update the current Python environment to the newest SDK release.

Common forms:

iints update
iints update --dry-run
iints update --repair --force-reinstall --yes
iints update --no-cache-dir --yes
iints update --source github --github-ref stable --yes
iints update --extras full,mdmp,research,edge

The command prints the exact python -m pip install -U ... invocation before it changes anything. By default --source auto uses the stable PyPI release and never silently falls back to GitHub main. Use --dry-run during a live demo setup check, or --check for a status-only automation check.

iints delete

Remove IINTS from the current machine/environment with a visible deletion plan.

Common forms:

iints delete --dry-run
iints delete --yes
iints delete --everything --dry-run
iints delete --everything --yes
iints delete --source-checkout --yes
iints delete --local-outputs --yes
iints delete --no-packages --path results/old_iints_run --yes

Default behavior removes the active Python SDK packages plus user-level IINTS config, plugin, and cache folders. --everything also includes known generated output folders in the current directory and a detected local IINTS-SDK source checkout. It still does not guess private datasets, external-drive research archives, or unrelated virtual environments.

Core Simulation Commands

iints run

Run one simulation.

Examples:

iints run --preset baseline_t1d
iints run --algo algorithms/example_algorithm.py --scenario scenarios/example_scenario.json
iints run --dry-run --preset baseline_t1d

iints run-full

One-line run with full output bundle.

iints run-parallel

Run a matrix of scenarios in parallel.

iints benchmark

Compare algorithms across standard workloads.

Extension Commands

iints plugin install

Install a local algorithm plugin without editing SDK source code.

iints plugin install algorithms/my_algo.py
iints algorithms list

The SDK copies the file into the local plugin home and records it in ~/.iints/plugins/registry.json. For tests or portable environments, set IINTS_PLUGIN_HOME to another folder.

iints plugin register

Register extension files by kind.

iints plugin register algo algorithms/my_algo.py
iints plugin register patient-model patient_models/my_model.py --name "My Model"
iints plugin register data-source data_sources/my_importer.py
iints plugin register validator validators/my_check.py

Algorithm plugins become visible in iints algorithms list. Patient model, data source, and validator plugins are registered for discovery/documentation hooks so the SDK can grow without source-code edits.

iints plugin list

Show local extension plugins.

iints plugin list
iints plugin list --kind algorithm

iints plugin uninstall

Remove a local plugin registry entry.

iints plugin uninstall "My Algorithm"
iints plugin uninstall "My Algorithm" --remove-file

iints patientmodel list

Show built-in and locally registered patient models.

iints patientmodel list

Study / Research Commands

iints study-protocol

Write the official study protocol bundle.

iints run-study

Run the scientific benchmark matrix.

iints analyze

Aggregate a study directory.

iints compare-study

Compare two study outputs.

iints poster-study

Generate poster-ready figures from a study.

iints research regenerative panels

Inspect the bundled regenerative-islet protein panels and optionally export the required evidence plan:

iints research regenerative panels \
  --panel beta_cell_identity_and_function \
  --output-json results/regenerative/panels.json

iints research regenerative compare

Run a provenance-preserving descriptive protein comparison between normalized SC-islet and primary-islet observations:

iints research regenerative compare \
  --dataset data/standardized_islet_proteomics.csv \
  --panel beta_cell_identity_and_function \
  --normalization-note "Joint normalization within one experiment" \
  --output-dir results/regenerative/protein_comparison

The command does not calculate treatment efficacy, approve a cell product, or map protein abundance into graft-model parameters.

iints research regenerative import-proteomics

Standardize MaxQuant proteinGroups.txt, DIA-NN report.tsv, or wide matrix files into the regenerative comparator contract:

iints research regenerative import-proteomics \
  --input-file data/maxquant/proteinGroups.txt \
  --sample-metadata data/sample_annotations.csv \
  --source-id PXD001539 \
  --format maxquant \
  --output-csv data/standardized_islet_proteomics.csv

Data Commands

iints data list

Show public data packs.

iints data fetch

Fetch a pack into a local directory.

iints data research-plan

Generate the curated diabetes dataset acquisition plan for local AI research.

iints data research-plan --output-dir data_packs/research_dataset_plan

Use --dataset <id> repeatedly to generate a focused plan for a subset.

iints data certify

Run data certification.

iints data realism-check

Judge whether a glucose trace looks physiologically plausible for research or demo use. Supports: - --reference free_living_t1d, --reference azt1d, or --reference hupa_ucm - --output-json results/realism_report.json - --output-html results/realism_dashboard.html - --min-realism-verdict needs_review

iints data import-cgmacros

Import and standardize the multimodal CGMacros dataset (Nature Scientific Data, 2025): - Synchronizes Abbott FreeStyle Libre Pro and Dexcom G6 Pro sensors. - Extracts discrete meals with exact macronutrient breakdown (carbs, protein, fat, fiber, calories). - Merges participant clinical screening (bio.csv).

iints data import-cgmacros \
  --input-dir data/cgmacros_raw \
  --output-dir data/cgmacros_standardized

iints research ppgr-benchmark

Benchmark 2-hour Postprandial Glycemic Response (PPGR) estimators: - Uses a subject-grouped split; row-level random splitting is rejected. - Compares carb-only, macronutrient, and measured-context ridge baselines. - Adds the IINTS GlucoFM reproduction only when a trained checkpoint and measured 24-hour pre-meal histories are supplied. - Computes trajectory MAE, RMSE, Pearson \(r\), Peak Glucose MAE, and Time-to-Peak error.

iints research ppgr-benchmark \
  --meals-file data/cgmacros_standardized/cgmacros_meals.csv \
  --glucofm-checkpoint models/glucofm-reproduction/glucofm_encoder.pt \
  --sensor dexcom \
  --output-dir results/ppgr_benchmark

iints research cgm-jepa-embed

Extract a 96-dimensional latent representation from a 24h simulation or CGM trace using CGM-JEPA (arXiv:2605.00933):

iints research cgm-jepa-embed \
  --input results/baseline/results.csv \
  --output-dir results/cgm_jepa_embedding

iints research cgm-jepa-experiment

Run a 100-simulation virtual patient parameter sweep (varying \(S_I\)) and test latent manifold alignment and sensor noise robustness (\(\cos \theta\)):

iints research cgm-jepa-experiment \
  --output-dir results/cgm_jepa_study \
  --n-simulations 100 \
  --sweep-param insulin_sensitivity

iints research cgm-jepa-confounder

Run the Physiological Confounder Benchmark (50 paired cohorts) testing whether observational CGM foundation models confound divergent underlying biological states:

iints research cgm-jepa-confounder \
  --output-dir results/cgm_jepa_confounder \
  --num-pairs 50

Medical Device Safety & FDA Verification Commands

iints safety fda-list

List all verified medical device recall cases from the US FDA database registered in IINTS-AF:

iints safety fda-list

iints safety fda-benchmark

Run the complete FDA Adverse Event Benchmark evaluating unmitigated automated controllers vs the IINTS-AF Dual-Guard Safety Supervisor across real device failures (Tandem, MiniMed, Omnipod, Dexcom):

iints safety fda-benchmark \
  --output-dir results/fda_safety_study

iints import-carelink, iints import-nightscout, iints import-tidepool

Import real-world CGM sources.

Edge / Booth Commands

iints edge doctor

Preflight for Raspberry Pi or UNO Q.

iints edge quickstart

Create the easiest Pi or UNO Q demo project and optionally start the Linux-side runtime.

iints edge setup

Generate an edge project scaffold.

iints edge deploy

Scaffold, upload, install, and start a Raspberry Pi edge project in one command.

iints edge offline-bundle

Build a USB-friendly offline install tarball for Raspberry Pi or UNO Q setups.

iints edge study

Run a reproducible multi-seed study directly on the current edge machine.

iints edge long-study

Run a multi-day or multi-week YAML-driven study directly on the Pi, with rolling day profiles and export-friendly nested outputs. Use --resume to continue from the next incomplete day after a reboot.

iints edge study-snapshot

Create a .tar.gz snapshot of a long-study folder for crash recovery or USB backup.

iints edge study-export

Package a long-study folder into a transfer-ready zip archive for another device.

iints edge remote-status, iints edge remote-reset, iints edge remote-stop

Run common Raspberry Pi maintenance commands remotely over SSH.

Jetson Endurance Commands

iints jetson doctor

Check Jetson-like hardware probes, thermal zones, and NVIDIA tooling before a long headless run.

iints jetson endurance start

Run a headless adversarial endurance study.

iints jetson endurance start \
  --algo algorithms/example_algorithm.py \
  --predictor models/lstm_predictor.pt \
  --duration 7d \
  --output-dir results/jetson_7day \
  --profile mixed_adversarial \
  --seed 42 \
  --checkpoint-interval 360 \
  --hardware-sample-interval 60

Add --wall-clock when the study horizon should consume real time instead of finishing as fast as possible. A run such as --duration 1d --wall-clock therefore lasts about 24 real hours and writes a training-ready research/ bundle next to the normal endurance outputs.

iints jetson endurance status

Show progress, current glucose, TIR so far, interventions, critical events, the latest checkpoint, resume count, and wall-clock ETA.

iints jetson endurance monitor

Print the same status repeatedly with --watch.

iints jetson endurance stop

Request a safe stop and optional report finalization.

iints jetson endurance export

Package the complete endurance folder into a transfer-ready .zip.

iints jetson endurance finalize-research

Train post-run local research models from one endurance bundle and write a held-out closed-loop evaluation report.

iints jetson endurance install-service

Write a systemd service file with automatic --resume for multi-day Jetson runs. Pass --wall-clock here too when the generated service should preserve real-time pacing.

Local AI Research Commands

iints research blend-datasets

Blend already prepared real datasets into one source-aware predictor dataset.

iints research prepare-ohio

Prepare a local OhioT1DM XML folder into a gitignored processed dataset. Do not commit the raw OhioT1DM-volledig/ folder to GitHub:

iints research prepare-ohio \
  --input-dir /path/to/OhioT1DM-volledig \
  --splits train \
  --output data_packs/public/ohio_t1dm_full/processed/ohio_train.csv \
  --report data_packs/public/ohio_t1dm_full/processed/ohio_train_quality_report.json

iints research glucose-model build-dataset

Normalize one or more prepared glucose datasets into the dedicated iints-glucose-forecast-v0 training contract:

iints research glucose-model build-dataset \
  --input data_packs/public/ohio_t1dm_full/processed/ohio_train.csv \
  --input results/realism_learning_10k/research/predictor_training.csv \
  --labels ohio_full,sim_10k \
  --profile long \
  --output-dir models/iints-glucose-forecast-v0/dataset

iints research glucose-model train

Train the dedicated glucose-forecast model and optionally build a Hugging Face-ready export folder:

iints research glucose-model train \
  --data models/iints-glucose-forecast-v0/dataset/glucose_training_dataset.csv \
  --config models/iints-glucose-forecast-v0/dataset/glucose_model_config.yaml \
  --output-dir models/iints-glucose-forecast-v0 \
  --epochs 220 \
  --comparison-dir results/glucose_model_comparison \
  --export-hf

iints research glucose-model compare

Compare transparent baselines and trained MSE/Band/PINN checkpoints against physiology-aware gates:

iints research glucose-model compare \
  --data data_packs/public/ohio_t1dm_full/processed/ohio_test.csv \
  --config models/iints-glucose-forecast-v0/dataset/glucose_model_config.yaml \
  --model mse=models/glucose_mse/predictor.pt \
  --model pinn=models/iints-glucose-forecast-v0/predictor.pt \
  --mc-samples 30 \
  --output-dir results/glucose_model_comparison

iints research glucose-model export-hf

Package predictor.pt, training_report.json, the model config, privacy/limitations notes, examples, comparison metrics, a research-only model card, and a redacted public dataset manifest for Hugging Face:

iints research glucose-model export-hf \
  --model-dir models/iints-glucose-forecast-v0 \
  --dataset-manifest models/iints-glucose-forecast-v0/dataset/glucose_dataset_manifest.json \
  --comparison-dir results/glucose_model_comparison \
  --repo-id IINTS/iints-glucose-forecast-v0

iints research glucose-model jetson-train-hf

Continue training an existing Hugging Face glucose model on Jetson with warm-start, candidate comparison, and local champion promotion:

iints research glucose-model jetson-train-hf \
  --repo-id IINTS/iints-glucose-forecast-v0 \
  --dataset models/iints-glucose-forecast-v0/dataset/glucose_training_dataset.csv \
  --dataset-manifest models/iints-glucose-forecast-v0/dataset/glucose_dataset_manifest.json \
  --work-dir models/jetson_hf_training \
  --max-trials 1 \
  --epochs 2 \
  --batch-size 64 \
  --upload-mode none

Use --upload-mode pr after review to upload a promoted champion as a Hugging Face pull request. The command uploads model artifacts and redacted metadata only, not raw private dataset rows.

This command accepts the native IINTS glucose-forecast bundle format containing predictor.pt; it is not a generic adapter for GlucoFM, CGM-JEPA, or arbitrary Hugging Face backbones. Use iints research glucofm-pretrain for the independent GlucoFM reproduction.

iints research build-control-dataset

Combine one or more run bundles into a supervised controller teacher dataset.

iints research train-controller

Train the first auditable local controller baseline from safe-action labels.

iints research train-neural-controller

Train the stronger PyTorch controller from the same supervised safe-action labels.

iints research evaluate-controller

Compare a learned controller against the clinical baseline on held-out presets and seeds.

iints research local-ai-lab

Combine completed Jetson/simulator runs into one local AI workspace: predictor dataset, controller-teacher dataset, dataset card, local controller models, optional predictor training, and held-out controller evaluation.

iints research train-local-ai

Friendly alias for the same local AI workspace command. Prefer this name in new demos and docs:

iints research train-local-ai \
  --run day1=results/jetson_research_day \
  --output-dir results/local_ai_lab

Full workflow: Jetson Endurance Mode.

Results Management

iints results

Index all run-level results.csv files and every generated artifact under a results root:

iints results --root results

This writes a compact management bundle:

• run_index.csv
• artifact_inventory.csv
• RESULTS_INDEX.md
• result_manager_manifest.json
• results_index.xlsx when spreadsheet export is available

Use --include-raw only when you want one combined long table for downstream local-AI/data analysis:

iints results --root results/research_realism_sweep_20260603_02 --include-raw

The same command is also available as iints research results-index for research workflows.

Evidence Commands

iints run-doctor

Preflight an algorithm, patient YAML, scenario JSON, duration, time step, and output folder before a long run:

iints run-doctor \
  --algo algorithms/example_algorithm.py \
  --patient-config-path patients/stable_patient.yaml \
  --scenario-path scenarios/clinic_safe_baseline.json \
  --duration 1440 \
  --time-step 5

It catches missing files, validation errors, aggressive glucose drift, likely pre-meal hypoglycemia, and output path problems.

iints evidence build

Build a public research evidence bundle from one or more completed runs:

iints evidence build \
  --run normal=results/live_demo/results/01_normal_run \
  --run stress=results/live_demo/results/02_meal_stress \
  --output-dir results/live_demo/evidence_bundle

Optional inputs:

• --local-ai-dir results/local_ai_lab
• --pump-bundle-dir bundles/pico_bench_bundle

iints report --style agp

Generate an AGP-style research PDF from a dense simulation or CGM CSV:

iints report \
  --results-csv results/one_day/results.csv \
  --style agp \
  --png \
  --svg \
  --subject-name "stable demo run" \
  --bundle-dir results/one_day/agp_report

This writes agp_report.pdf, agp_summary.json, agp_assets/agp_profile.png, agp_assets/agp_profile.svg, agp_assets/daily_profiles.png, and agp_assets/daily_profiles.svg. When a run contains explainable_events, the AGP asset folder also includes xai_events.txt for human review and xai_events.json for downstream analysis. The layout includes glucose statistics, time-in-ranges, an AGP-style modal-day percentile plot, and daily glucose profiles.

iints safety-visualize

Create a standalone HTML safety visualizer from a run CSV:

iints safety-visualize \
  --results-csv results/one_day/results.csv \
  --output-html results/one_day/safety_visualizer.html \
  --output-json results/one_day/safety_visualizer.json

Normal iints run, iints presets run, run_full(...), and run_simulation(...) already create realism_report.json, realism_dashboard.html, safety_visualizer.html, and safety_visualizer.json inside the run bundle.

Pico Pump Bench Commands

iints pump init

Create a bench-only Pico pump lab workspace.

iints pump compile

Package a simulated SDK algorithm with locked, non-actuating Pico firmware:

iints pump compile \
  --algorithm algorithms/pico_bench_algorithm.py \
  --output-dir bundles/pico_bench_bundle

iints pump bench-test

Validate the bundle before upload:

iints pump bench-test \
  --bundle-dir bundles/pico_bench_bundle \
  --output-json bundles/pico_bench_bundle/bench_test_report.json

iints pump upload

Copy the locked bundle to a writable Pico/CircuitPython-style drive after explicit bench-only confirmation.

FPGA Safety-Core Commands

iints fpga start

Run the easiest FPGA quickstart: create a lab scaffold and run the golden mock safety-core demo.

iints fpga start --output-dir results/fpga_start

iints fpga doctor

Check whether FPGA mode can run locally. Mock transport works without hardware; serial transport is optional and requires pyserial.

iints fpga setup

Create a bench-only FPGA lab workspace:

iints fpga setup --output-dir iints_fpga_lab

The workspace includes a safety contract, demo events, a golden night_hypo_risk scenario, JSON-lines protocol description, Verilog scaffold, Verilog smoke test, JSON-lines bridge stub, FPGA_STORY.md, and a mock-demo shell script.

iints fpga simulate

Run a software-reference versus FPGA-style safety-core comparison:

iints fpga simulate \
  --events iints_fpga_lab/scenarios/night_hypo_risk.json \
  --output-dir results/fpga_mock_run

Use --transport serial --port /dev/ttyUSB0 when a real FPGA bridge is available.

iints fpga export-events

Convert an existing IINTS results CSV into FPGA event JSON:

iints fpga export-events \
  --results-csv results/my_run/results.csv \
  --output-events results/fpga_events.json

iints fpga replay

Convert an existing IINTS results CSV and immediately run the FPGA comparison:

iints fpga replay \
  --results-csv results/my_run/results.csv \
  --output-dir results/fpga_replay

iints fpga compare

Read fpga_comparison.json and fail if hardware-style output diverged from the SDK reference:

iints fpga compare --run-dir results/fpga_mock_run

iints fpga report

Print the report location and summary:

iints fpga report --run-dir results/fpga_mock_run

iints fpga demo

Create the lab scaffold and run a mock FPGA safety-core demo in one command:

iints fpga demo --output-dir results/fpga_demo

The demo bundle includes reviewer-friendly top-level files: events.csv, results.json, manifest.json, and report.md.

Maker Faire Commands

iints makerfaire up

Start the Pi booth flow.

iints makerfaire autostart

Prepare booth autostart files.

iints makerfaire watchdog

Recover the booth runtime if it stops.

Diagnostics

iints doctor

Basic and full environment checks.

iints doctor
iints doctor --full --suggest
iints doctor --smoke-run

doctor reports the installed SDK version, active Python executable, install path, and available command groups. This is the fastest way to catch an old Python interpreter that silently resolves only legacy SDK releases.

iints profiles

iints profiles presets
iints profiles create --name stable_patient --preset stable-demo
iints profiles create --name endurance_patient --preset endurance

Starter presets are provided for stable-demo, stress-test, and endurance.

MDMP Cryptography And Provenance

iints mdmp encrypt-data

Encrypt a local dataset with ChaCha20-Poly1305 authenticated encryption. Text passphrases are processed with a random per-envelope scrypt salt:

iints mdmp encrypt-data --input data/patient_cgm.csv --output data/patient_cgm.enc

iints mdmp decrypt-data

Decrypt and authenticate encrypted dataset files, ensuring tamper-resistance:

iints mdmp decrypt-data --input data/patient_cgm.enc --output data/restored.csv

iints mdmp sign

Sign an MDMP dataset passport card with Ed25519. The SDK does not currently implement or claim an ML-DSA/post-quantum signature:

iints mdmp sign --card results/passport.json --key keys/mdmp_private_v1.pem

iints mdmp verify

Verify an MDMP dataset passport card signature, expiry, and SHA-256 fingerprint:

iints mdmp verify results/passport.json --dataset data/patient_cgm.csv

iints mdmp validate

Validate dataset against an MDMP contract schema and range rules:

iints mdmp validate contracts/cgm_contract.yaml data/patient_cgm.csv --min-mdmp-grade research_grade

Foundation Models & Multi-Sensor Research Commands

iints research glucofm-pretrain

Pretrain the independent, paper-aligned IINTS GlucoFM v2 reproduction on mask-preserving daily windows:

iints research glucofm-pretrain \
  --source data/processed/cgm.csv \
  --glucose-column glucose_mgdl \
  --timestamp-column timestamp \
  --subject-column subject_id \
  --output-dir models/glucofm-reproduction

iints research glucofm-embed

Extract a checkpoint-backed 128-dimensional representation. Official Google weights are not bundled:

iints research glucofm-embed \
  --input-file data/patient_run.csv \
  --checkpoint models/glucofm-reproduction/glucofm_encoder.pt \
  --glucose-column glucose_mgdl \
  --timestamp-column timestamp \
  --output-file results/glucofm/embedding.csv

iints research foundation-arena

Compare measured model-evaluation artifacts from one shared group-disjoint benchmark. The command contains no built-in scores:

iints research foundation-arena \
  --result results/evaluations/glucofm.json \
  --result results/evaluations/jepa.json \
  --output-dir results/foundation_arena

iints research visualize

Generate the GlucoFM method schematic and add result figures only when their evidence artifacts are supplied:

iints research visualize \
  --arena-result results/evaluations/glucofm.json \
  --arena-result results/evaluations/baseline.json \
  --confounder-evidence results/confounder_pairs.csv \
  --dual-sensor-evidence results/paired_sensors.csv \
  --safety-trace results/safety_trace.csv \
  --output-dir results/scientific_visualizations

Without these inputs, missing result panels are marked not generated; the command never inserts synthetic performance scores.

iints research eucys-playbook

Build the jury dossier from the same evidence contracts. Clarke EGA input must contain reference_mgdl and predicted_mgdl:

iints research eucys-playbook \
  --prediction-evidence results/held_out_predictions.csv \
  --arena-result results/evaluations/glucofm.json \
  --arena-result results/evaluations/baseline.json \
  --confounder-evidence results/confounder_pairs.csv \
  --dual-sensor-evidence results/paired_sensors.csv \
  --safety-trace results/safety_trace.csv \
  --output-dir results/eucys_jury_dossier

iints data download-cgmacros

Download or generate the complete 45-participant CGMacros multi-sensor open science cohort with macronutrient meal events:

iints data download-cgmacros \
  --output-dir data/cgmacros_cohort \
  --participants 45

Full Details

For every option and advanced workflow, continue to: - Choose Your Path - CLI & Advanced Reference - Scientific Workflow - Study Analysis - CGM Foundation Models & Benchmarking - CGMacros Multi-Sensor Pipeline - OpenFDA Safety Benchmarks