ashmatics-datamodels 0.3.2

Canonical Pydantic data models for AshMatics healthcare applications - FDA vocabulary, regulatory schemas, and clinical AI use cases

AshMatics Core DataModels

Version: 0.3.1

Canonical Pydantic data models for AshMatics healthcare applications.

Changelog

v0.3.1 (2026-01-25) — ASHKBAPP-66

• Added PROCESS_DOCUMENTATION to GovernanceCategory enum
• This is a core CAI framework category required for MCP service compatibility

Overview

This library provides the single source of truth for data contracts across the AshMatics ecosystem:

• Knowledge Base (KB)
• CoreApp
• ashmatics-tools SDK
• AI Watch applications

Features

• FDA Vocabulary: OpenFDA-aligned schemas for manufacturers, clearances, classifications, recalls, adverse events
• MongoDB Document Schemas: Three-tier structure for all kb_* collections (evidence, regulatory, model cards, products, manufacturers, use cases)
• Governance Document Models: Clinical AI Governance Framework artifacts (policies, SOPs, work products, process documentation)
• Use Case Taxonomy: Clinical AI use case categorization
• Rich Validation: Built-in validators for regulatory identifiers (K numbers, product codes)
• Database Agnostic: Pure Pydantic models, no ORM coupling
• Type Safe: Full type hints with mypy support

Installation

# From git (recommended for now)
pip install git+https://github.com/AsherInformatics/ashmatics-core-datamodels.git

# Or add to pyproject.toml
# dependencies = [
#     "ashmatics-datamodels @ git+https://github.com/AsherInformatics/ashmatics-core-datamodels.git",
# ]

Quick Start

from ashmatics_datamodels.fda import (
    FDA_ManufacturerBase,
    FDA_510kClearance,
    FDA_DeviceClass,
    ClearanceType,
)

# Create a manufacturer
manufacturer = FDA_ManufacturerBase(
    manufacturer_name="Medical AI Corp",
    applicant="Medical AI Corp",
)

# Create a 510(k) clearance with validation
clearance = FDA_510kClearance(
    k_number="K240001",  # Validated format
    clearance_date="2024-08-15",
    device_name="AI-Chest Scanner",
    device_class=FDA_DeviceClass.CLASS_2,
)

Package Structure

ashmatics_datamodels/
├── common/          # Base models, validators, regulators, frameworks
├── fda/             # FDA vocabulary (manufacturers, clearances, classifications, recalls, adverse events)
├── documents/       # MongoDB document schemas (three-tier structure)
├── use_cases/       # Clinical AI use case taxonomy
└── utils/           # Parsing and normalization utilities

Documentation

📚 Full Documentation (when published)

Or build locally:

uv pip install -e ".[docs]"
uv run mkdocs serve

Design Documents

• Phase 1: FDA & Common Schemas
• Phase 2: MongoDB Document Schemas
• Complete Migration Plan

License

Apache 2.0 - See LICENSE for details.

Contributing

This is an internal Asher Informatics library. For questions, contact [e-mail].