DigitalSoma v3.0.0 — Species-aware VeDDRA engine

What's new in v3.0.0

Major — VeDDRA overhaul

• New subpackage digitalsoma/veddra/ with official EMA Rev.16 term vocabulary
• Species-aware adverse event screen: 4 to 13 rules per species, replacing the flat 6-rule universal list
• Corrected VeDDRA codes: all codes now use official EMA Rev.16 PT-level integers; v2.x codes were incorrect

Fixes

• spo2_pct URI corrected (SNOMED 103228002 — was wrongly using a LOINC code)
• ambient_temp_C and skin_temp_C now have distinct URIs
• rmr_W and rmr_kcal_day now have distinct URIs
• FHIR Observation categories are correct across all 21 properties
• LOINC codes added for vco2_L_min and rmr_kcal_day with reverse map entries
• soma_agent.py docstring corrected to match actual tool schemas

Licence change

CC BY 4.0 → CC BY-NC-SA 4.0 (non-commercial, ShareAlike)

Documentation

Full technical documentation

Install: pip install digitalsoma==3.0.0

DigitalSoma v2.2.0

DigitalSoma v2.2.0

Author: Dr. ir. Ali Youssef
ORCID: 0000-0002-9986-5324
Affiliation: Computational Bio-Ecosystems, Digital Agroecosystems Laboratory,
University of Manitoba & BioTwinRs Ltd., Winnipeg, Canada

Install

pip install digitalsoma

What's in this release

• Physics-based five-layer digital twin for animal physiology
• 6 ATR species: porcine, equine, bovine, ovine, canine, salmonid
• 6 built-in physiological solvers (Fick, Kleiber+Q10, Newton, RQ, HPA, VeDDRA)
• HL7 FHIR R4 integration — to_fhir_bundle(), from_fhir_bundle(), FHIRMapper
• 30 canonical properties mapped to LOINC + SNOMED CT + UCUM
• VeDDRA v2.2 adverse event screen — EMA EVVET3 / UK VMD / FDA-CVM
• 11 OpenAI-compatible LLM tool schemas
• PyPI packaging — pyproject.toml, MANIFEST.in, py.typed
• 36 FHIR tests, full documentation, porcine clinical simulation
• Zero runtime dependencies · Python ≥ 3.9 · CC BY 4.0