Data Curation Checker

A tool for validating FHIR R4 Observation and DiagnosticReport resources. It checks structure, terminology, references, and laboratory-specific rules. You can run it in the browser or use the validation logic without the UI (e.g. from Node, scripts, or CI).

---

Running the Web UI

• Development: npm start — serves at http://localhost:4200/
• Production build: npm run build — output in dist/
• GitHub Pages: The app is deployed via GitHub Actions to https://jan92.github.io/Data_Curation_Checker/ when you push to main or master.

---

Supported input formats

• FHIR JSON Bundle (e.g. type: "collection" with entry[]) containing Observation and/or DiagnosticReport
• JSON array of Observation and/or DiagnosticReport resources
• NDJSON (one JSON resource per line)
• Single Observation or DiagnosticReport (JSON object)

---

What is validated

• Observation: Structure, required fields (status, code), value[x] vs dataAbsentReason, organizer rules; CodeableConcepts, references, dates, URLs.
• DiagnosticReport: Required status, code; recommended category, subject, effective/issued, performer, result (Observation references); reference and CodeableConcept checks. See FHIR DiagnosticReport.
• Laboratory (Observation, LOINC-coded): Parameter-specific rules for many common lab parameters (e.g. glucose, creatinine, HbA1c, CBC, liver, cardiac, thyroid, coagulation, tumor markers, urine, hormones, vitamins):
  • UCUM units and reference ranges, critical low/high
  • Specimen, method, timing, interpretation; panel/component, reflex, delta checks; status workflow, performer, device, dataAbsentReason
• Structure: Bundle format, resource relationships, reference integrity.

---

Using the checks without the UI

The validation logic lives in a plain TypeScript module with no Angular dependency. You can:

• Import and call it from any Node or TypeScript project.
• Run the CLI to validate files or stdin and get a JSON report.

Programmatic usage

Import and call:

import { validateFhirObservations } from './src/app/checks/fhir-observation-checks';

const report = validateFhirObservations(content, {
  source: 'File',           // optional: 'File' | 'Text' | 'stdin' | 'Input'
  sourceDetail: 'path.json' // optional: file path or other detail
});

Signature:

validateFhirObservations(
  content: string,
  options?: { source?: string; sourceDetail?: string }
): ValidationReport

ValidationReport:

{
  parseResult: {
    ok: boolean;
    type: string;              // e.g. 'JSON Object', 'NDJSON', 'JSON Array'
    resources: Observation[];
    diagnosticReports: DiagnosticReport[];
    error?: string;            // set when ok is false
  };
  issues: CheckIssue[];        // { severity, label, detail, location }
  checkResults: CheckResult[]; // { label, status, statusLabel, detail }
}

Types are exported from ./src/app/checks/fhir-observation-checks and ./src/app/checks/types.

CLI (headless)

The CLI reads from a file or stdin and prints a ValidationReport as JSON.

From a file:

npm run check:cli -- --file ./observations.json

From stdin:

cat observations.json | npm run check:cli

Requirements: ts-node is used as a dev dependency. Run npm install first.

---

Project layout

• src/app/checks/ — Validation logic (framework-agnostic):
  • fhir-observation-checks.ts — FhirObservationChecker, validateFhirObservations
  • types.ts — CheckStatus, CheckResult, CheckIssue, ParseResult, ValidationReport, ValidateOptions
• src/app/models/fhir.types.ts — FHIR R4 type definitions used by the checks
• scripts/run-checks-cli.ts — CLI entry point
• tsconfig.cli.json — tsconfig for the CLI (Node, CommonJS)

---

License

See repository for license information.