Domain Dictionaries

Domain Dictionaries

New in v0.3

infermap ships curated alias dictionaries for common business domains. Loading one boosts mapping accuracy on schemas that use standard terminology for that domain.

Available domains

Domain	Scope
generic	Common PII, customer, order, system fields. Loaded by default.
healthcare	HL7/FHIR/EHR: MRN ↔ patient_id, ICD/CPT codes, observations, medications, providers.
finance	ISO 20022 / banking / trading: account_id ↔ IBAN, amt ↔ amount, ccy ↔ currency.
ecommerce	Catalog / order / fulfillment: SKU ↔ product_id, qty ↔ quantity, shipping/billing variants.

Usage

Python

from infermap import MapEngine

engine = MapEngine(domains=["healthcare"])
result = engine.map("epic_extract.csv", "research_schema.csv")

TypeScript

import { MapEngine } from "infermap";

const engine = new MapEngine({ domains: ["healthcare"] });
const result = engine.mapSchemas(source, target);

Via config file

# infermap.yaml
domains:
  - healthcare
  - finance

What you get

Without the dictionary, mrn and patient_id are connected only via fuzzy name similarity (~0.27 confidence). With domains: ["healthcare"], the AliasScorer fires a direct match at 0.95.

Extending

Add custom aliases on top of shipped dictionaries:

domains:
  - healthcare
aliases:
  patient_id:
    - our_patient_uid
    - legacy_pt_id

See also

• Scorers — how the AliasScorer uses domain dictionaries
• Domain dictionaries docs page
• Example: 09_domain_dictionaries.py