Complete Source Library¶
This page is the public source index for IINTS-AF. It collects the references used to shape SDK defaults, physiology assumptions, safety framing, dataset guidance, local AI workflows, and documentation.
Research-only boundary
These sources support transparent simulation and educational AI research. They are not a claim that IINTS-AF is clinically validated or certified for treatment decisions.
How To Use This Page¶
- Use this page when a reviewer asks: "Where did these assumptions come from?"
- Use Evidence Base for the shorter explanation of how sources map to SDK components.
- Use Diabetes Research Datasets for dataset acquisition and local AI planning.
- Use
iints sources --output-json results/source_manifest.jsonto export the packaged evidence subset from an installed SDK. - Use
iints data research-plan --output-dir data_packs/research_dataset_planto export the dataset-source plan.
Packaged Scientific Sources¶
These are the machine-readable evidence sources shipped in src/iints/presets/evidence_sources.yaml and exposed through iints sources.
| ID | Category | SDK component | Title | Link | Why it matters |
|---|---|---|---|---|---|
ada_2026_glycemic_goals |
guideline | validation_targets | Glycemic Goals and Hypoglycemia: Standards of Care in Diabetes—2026 | DOI 10.2337/dc26-S006 | Reference for TIR/TBR/TAR target framing and hypoglycemia thresholds in validation profiles. |
ada_2026_diabetes_technology |
guideline | cgm_and_aid_context | Diabetes Technology: Standards of Care in Diabetes—2026 | DOI 10.2337/dc26-S007 | Reference for CGM and AID operational context used in report language and safeguards. |
attd_2019_time_in_range |
consensus | metrics_targets | Clinical Targets for Continuous Glucose Monitoring Data Interpretation | DOI 10.2337/dci19-0028 | Basis for TIR/TBR/TAR interpretation in analytics and scorecards. |
nejm_2019_control_iq |
trial | aid_benchmarking | Six-Month Randomized, Multicenter Trial of Closed-Loop Control in Type 1 Diabetes | DOI 10.1056/NEJMoa1907863 | Reference envelope for realistic AID performance ranges in benchmark narratives. |
adapt_2022_ahcl |
trial | aid_benchmarking | Advanced hybrid closed loop therapy versus conventional treatment in adults with type 1 diabetes (ADAPT) | DOI 10.1016/S2213-8587(22)00212-1 | Adds modern AHCL outcomes for realistic comparator ranges. |
cobry_2010_meal_bolus_timing |
trial | prebolus_timing | Timing of Meal Insulin Boluses to Achieve Optimal Postprandial Glycemic Control | DOI 10.1177/193229681000400404 | Reference for pre-bolus timing defaults and scenario annotations. |
heise_2017_fiasp_pkpd |
pharmacology | insulin_action_profiles | A Faster-Onset Formulation of Insulin Aspart | DOI 10.1007/s40262-017-0510-8 | Reference for rapid insulin onset assumptions in therapy profiles. |
klaff_2020_urli_pkpd |
pharmacology | insulin_action_profiles | Ultra Rapid Lispro Demonstrates Accelerated Pharmacokinetics and Pharmacodynamics | DOI 10.1111/dom.14049 | Reference for ultra-rapid insulin variants and action-curve configuration. |
wentholt_2004_cgm_lag |
sensor | cgm_lag_and_validator | How glucose sensors can facilitate therapy in diabetes management | DOI 10.1089/dia.2004.6.615 | Background reference for physiologic CGM lag and fail-soft validator behavior. |
dalla_man_2007_meal_model |
model | virtual_patient_dynamics | Meal simulation model of the glucose-insulin system | DOI 10.1109/TBME.2007.893506 | Reference foundation for physiologic meal disturbance modeling. |
mandelbrot_1968_fractional_brownian |
model | cgm_noise_model | Fractional Brownian Motions, Fractional Noises and Applications | DOI 10.1137/1010093 | Mathematical anchor for long-memory, fractional-noise-style CGM noise approximations. |
campbell_1985_dawn_phenomenon |
physiology | circadian_egp_model | Pathogenesis of the Dawn Phenomenon in Patients with Insulin-Dependent Diabetes Mellitus | DOI 10.1056/NEJM198506063122302 | Reference for dawn-phenomenon physiology involving nocturnal growth hormone surges, glucose production, and impaired utilization. |
richter_2013_glut4_exercise |
physiology | exercise_glut4_model | Exercise, GLUT4, and skeletal muscle glucose uptake | DOI 10.1152/physrev.00038.2012 | Reference for exercise/contraction-stimulated GLUT4 translocation and insulin-independent skeletal-muscle glucose uptake. |
naslund_1999_glp1_gastric_emptying |
physiology | glp1_gastric_emptying_model | GLP-1 slows solid gastric emptying and inhibits insulin, glucagon, and PYY release in humans | DOI 10.1152/ajpregu.1999.277.3.R910 | Reference for GLP-1-associated slowing of gastric emptying used to document nonlinear meal absorption feedback. |
bergman_1979_minimal_model |
model | virtual_patient_dynamics | Quantitative estimation of insulin sensitivity | DOI 10.1152/ajpendo.1979.236.6.E667 | Foundation for the Bergman-style minimal-model insulin sensitivity and remote insulin-action states. |
hovorka_2004_nmpc_t1d |
model | hovorka_patient_model | Nonlinear model predictive control of glucose concentration in subjects with type 1 diabetes | DOI 10.1088/0967-3334/25/4/010 | Reference for the Hovorka-style compartmental glucose, insulin, and insulin-action structure used in research-mode physiology. |
gerich_1979_counterregulation |
physiology | endogenous_counterregulation | Hormonal mechanisms of recovery from insulin-induced hypoglycemia in man | DOI 10.1152/ajpendo.1979.236.4.E380 | Primary human physiology reference for glucagon and epinephrine roles in acute recovery from insulin-induced hypoglycemia. |
cryer_2013_haaf_mechanisms |
review | haaf_memory_model | Mechanisms of Hypoglycemia-Associated Autonomic Failure in Diabetes | DOI 10.1056/NEJMra1215228 | Reference for HAAF as a recurrent-hypoglycemia memory mechanism that blunts counterregulation and symptom awareness. |
cryer_2013_haaf_diabetes |
review | haaf_memory_model | Hypoglycemia-associated autonomic failure in diabetes | DOI 10.1016/B978-0-444-53491-0.00023-7 | Reference for antecedent hypoglycemia, sleep, and exercise lowering later autonomic and symptomatic responses. |
lv_2013_exogenous_glucagon_pk |
pharmacology | exogenous_glucagon_pkpd | Pharmacokinetics modeling of exogenous glucagon in type 1 diabetes mellitus patients | DOI 10.1089/dia.2013.0150 | Foundation for dual-hormone glucagon depot-to-plasma transport models in research simulations. |
haidar_2013_dual_hormone_ap |
trial | dual_hormone_context | Glucose-responsive insulin and glucagon delivery in adults with type 1 diabetes | DOI 10.1503/cmaj.121265 | Clinical research context for dual-hormone insulin-plus-glucagon closed-loop systems. |
haidar_2013_insulin_glucagon_pk |
pharmacology | exogenous_glucagon_pkpd | Pharmacokinetics of Insulin Aspart and Glucagon in Type 1 Diabetes during Closed-Loop Operation | DOI 10.1177/193229681300700610 | Reference for subcutaneous insulin and glucagon pharmacokinetics during closed-loop operation. |
hummel_2018_renal_glucose_handling |
physiology | renal_glucose_clearance | Physiology of renal glucose handling via SGLT1, SGLT2 and GLUT2 | DOI 10.1007/s00125-018-4656-5 | Reference for renal glucose reabsorption physiology and transporter context behind smooth renal-threshold modeling. |
defronzo_2013_renal_reabsorption_splay |
physiology | renal_glucose_clearance | Characterization of Renal Glucose Reabsorption in Response to Dapagliflozin | DOI 10.2337/dc13-0387 | Reference for threshold, transport maximum, and splay concepts used when documenting renal glucose clearance. |
visentin_2018_uvapadova |
model | simulation_validation | The University of Virginia/Padova Type 1 Diabetes Simulator Matches the 2014 DMMS.R | DOI 10.1177/1932296818757747 | Reference for simulator validation methodology and in-silico credibility. |
mujahid_2024_generative_t1d_simulator |
model | simulation_realism | Generative deep learning for the development of a type 1 diabetes simulator | DOI 10.1038/s43856-024-00476-0 | Reference for modern data-driven T1D simulator realism framing and the limits of purely physiological approximations. |
riddell_2017_exercise_consensus |
consensus | exercise_stress_scenarios | Exercise management in type 1 diabetes: a consensus statement | DOI 10.1016/S2213-8587(17)30014-1 | Reference for exercise stress scenario defaults and safety interpretation. |
marling_2020_ohiot1dm |
dataset | predictor_training_data | OhioT1DM Dataset for Blood Glucose Level Prediction: Update 2020 | source | Primary public dataset used for subject-level predictor training and evaluation. |
idc_2025_agp_report_overview |
consensus | agp_style_reporting | Guide to Understanding the Ambulatory Glucose Profile (AGP) Report | source | Reference for AGP-style report layout: time-in-ranges, glucose metrics, modal-day percentile profile, and daily glucose profiles. |
Dataset Sources¶
These are the dataset references registered in src/iints/data/datasets.json. Some are bundled or public-download sources; others require manual download, approval, or a data-use agreement. The SDK records them for provenance but does not bypass source access rules.
| Dataset ID | Access | Source | Name | Link | Citation / use |
|---|---|---|---|---|---|
sample |
bundled | IINTS-AF | IINTS Sample CGM (Bundled) | n/a | Bundled full-day demo trace with meal and bolus annotations for quickstarts, reporting, and offline demos. |
ohio_t1dm |
request | UNC Charlotte Smart Health Lab | OhioT1DM Dataset | source | Eight weeks each for 12 people with type 1 diabetes; includes CGM, fingersticks, basal/bolus insulin, meals, exercise, sleep, work, stress, illness, and fitness-band physiology. |
diatrend |
request | Dartmouth College / Scientific Data | DiaTrend Dataset | source | Controlled-access CGM and insulin pump dataset from 54 people with type 1 diabetes, with carb logs, bolus data, and selected pump settings. |
t1d_uom |
manual | University of Manchester / Zenodo | T1D-UOM Longitudinal Multimodal Dataset | source | Twelve-week multimodal dataset from 17 people with type 1 diabetes, including CGM, basal and bolus insulin, meal macronutrients, physical activity, and sleep. |
t1d_granada |
request | University of Granada / Zenodo | T1DiabetesGranada Dataset | source | Longitudinal flash glucose, demographic, diagnostic, and biochemical dataset spanning four years and 736 people with type 1 diabetes. |
aide_t1d |
public-download | Jaeb Center for Health Research | AIDE T1D Public Dataset | source | Automated Insulin Delivery in Elderly with Type 1 Diabetes (AIDE T1D) public dataset. |
pedap |
public-download | Jaeb Center for Health Research | PEDAP Public Dataset | source | Pediatric Artificial Pancreas (PEDAP) public dataset (Release 5). |
azt1d |
manual | Mendeley Data | AZT1D: A Real-World Dataset for Type 1 Diabetes | source | Real-world T1D dataset (CGM + insulin + meals) from 25 individuals on AID systems. |
hupa_ucm |
manual | Mendeley Data | HUPA-UCM Diabetes Dataset | source | Free-living T1D dataset with CGM, insulin, meals, and activity data. |
openaps_data_commons |
request | OpenAPS | OpenAPS Data Commons | source | Community-contributed open APS data commons. |
tidepool_bigdata |
request | Tidepool | Tidepool Big Data Donation | source | Large-scale real-world diabetes data donation program. |
niddk_central |
request | NIDDK | NIDDK Central Repository | source | NIH/NIDDK centralized repository of clinical diabetes studies. |
t1d_exchange |
request | Jaeb Center / T1D Exchange | T1D Exchange Clinic Registry | source | Large clinical registry for type 1 diabetes research. |
d1namo |
manual | Zenodo / AISLab HES-SO | D1NAMO Open Dataset | source | Exploratory multimodal feature research after local checksum and manual schema review. |
t1dexi |
manual | Jaeb Center / Vivli / Public Study Websites | Type 1 Diabetes Exercise Initiative (T1DEXI) | source | Primary external dataset for exercise and activity stress testing of local AI models. |
t1dexip |
manual | Jaeb Center / Public Study Websites | Type 1 Diabetes Exercise Initiative Pediatric Study (T1DexiP) | source | Hold-out pediatric generalization dataset, not a direct training replacement for adult cohorts. |
dclp3_idcl |
public-download | NIDDK Central Repository / Jaeb Center | International Diabetes Closed Loop Trial (DCLP3 / iDCL) | source | External validation and benchmark-style comparisons after conversion to the IINTS standard schema. |
jaeb_loop |
public-download | Jaeb Center / Public Study Websites | Loop Observational Study Public Dataset | source | External real-world AID validation after local schema conversion and subject-level splitting. |
metabonet |
mixed | MetaboNet / arXiv | MetaboNet Consolidated T1D Dataset | source | Best long-term target for broad external validation once local access and terms are documented. |
glucose_ml |
collection | arXiv / Glucose-ML authors | Glucose-ML Dataset Collection | source | Use as a research map to choose external datasets and benchmark protocols; do not treat it as one homogeneous training set. |
Documentation-Only Implementation Sources¶
These references are used in narrative docs, setup guides, local-AI explanation, and best-effort device-emulation notes. They are intentionally documented separately from the packaged scientific manifest.
Local AI Runtime And Model Setup¶
| ID | Source | Link | Used for |
|---|---|---|---|
ollama_linux_install |
Ollama Linux installation documentation | docs.ollama.com/linux | Local LLM setup guidance |
mistral_2025_ministral_3_announcement |
Mistral AI, Introducing Mistral 3 | mistral.ai/news/mistral-3 | Local model-family context |
mistral_2025_ministral_3_3b |
Mistral AI model docs | Ministral 3 3B | Small local model guidance |
mistral_2025_ministral_3_8b |
Mistral AI model docs | Ministral 3 8B | Mid-size local model guidance |
mistral_2025_ministral_3_14b |
Mistral AI model docs | Ministral 3 14B | Larger local model guidance |
mistral_2026_adjustable_reasoning |
Mistral AI documentation | Adjustable reasoning | Serverless replacement settings and reasoning_effort |
mistral_2026_small_4 |
Mistral AI model docs | Mistral Small 4 | Cloud small-model migration target |
mistral_2026_medium_35 |
Mistral AI model docs | Mistral Medium 3.5 | Cloud strong-model migration target |
Device Emulation And Pump-Context Sources¶
| ID | Source | Link | Used for |
|---|---|---|---|
bergenstal_2020_780g |
Bergenstal et al., hybrid closed-loop safety study | DOI 10.1056/NEJMoa2003479 | 780G context and benchmark narrative |
fda_k193510_780g |
FDA 510(k) K193510 | FDA accessdata | 780G regulatory/device reference |
medtronic_780g_user_guide |
Medtronic Diabetes product documentation | medtronicdiabetes.com | User-guide terminology and mode descriptions |
brown_2019_control_iq_dtt |
Brown et al., Control-IQ performance paper | DOI 10.1089/dia.2019.0226 | Control-IQ benchmark context |
fda_k191289_control_iq |
FDA 510(k) K191289 | FDA accessdata | Control-IQ regulatory/device reference |
idcl_nct03563313 |
ClinicalTrials.gov iDCL trial record | NCT03563313 | Closed-loop trial context |
control_iq_user_guide |
Tandem Diabetes Care product documentation | tandemdiabetes.com | Control-IQ user-facing terminology |
assert_omnipod_5 |
Insulet / Omnipod ASSERT trial page | omnipod.com/assert-trial | Omnipod 5 context |
onset_omnipod_5 |
Insulet / Omnipod ONSET trial page | omnipod.com/onset-trial | Omnipod 5 type 2 diabetes context |
fda_k203467_omnipod5 |
FDA 510(k) K203467 | FDA accessdata | Omnipod 5 regulatory/device reference |
omnipod5_user_guide |
Insulet / Omnipod product documentation | omnipod.com | User-guide terminology and system context |
Hardware And Safety-Engineering Context¶
| ID | Source | Link | Used for |
|---|---|---|---|
fda_infusion_pump_software_safety_research |
FDA infusion pump software safety research | FDA | Pico pump lab framing: simulate first, verify before hardware |
eu_ai_pact |
European Commission AI Pact information | digital-strategy.ec.europa.eu | EU AI governance framing for research artifacts |
Source Maintenance Rules¶
- Add scientific sources to
src/iints/presets/evidence_sources.yamlwhen they affect SDK assumptions, validation targets, or report interpretation. - Add datasets to
src/iints/data/datasets.jsonwhen they affect research data planning or local AI workflows. - Add docs-only setup/device links here and in Evidence Base when they explain setup or product context but are not part of the packaged evidence manifest.
- Prefer DOI links for papers and official source pages for datasets, regulators, and vendor documentation.
- Do not cite a source as clinical validation unless the SDK actually performs a clinical validation protocol.
Export Commands¶
iints sources --output-json results/source_manifest.json
iints data research-plan --output-dir data_packs/research_dataset_plan
iints evidence build --run normal=results/live_demo/results/01_normal_run --output-dir results/evidence_bundle