Package org.fhirfrog.frog.testscriptrender

Class TestScriptRenderCli

java.lang.Object
org.fhirfrog.frog.testscriptrender.TestScriptRenderCli
public final class TestScriptRenderCli extends Object
Minimal, directory-in/HTML-files-out CLI over this module's renderer - not a general framework, just enough to let another build (e.g. an IG's own CI) turn a folder of TestScript resource files into browsable HTML fragments without a JVM dependency of its own.

Usage: TestScriptRenderCli <input-dir> <output-dir>

Scans every *.json file directly under <input-dir> (not recursively) and renders only the ones whose resourceType is actually TestScript - source directories like au-core-tests' input/resources/ mix TestScript definitions with plain fixture instances (Patient, Observation, ...) in the same flat folder with no filename convention distinguishing them, so filtering has to look at content, not the filename.

For each matching file (by its FHIR id, not its filename), writes <output-dir>/<id>.md containing the self-contained <section> fragment from TestScriptHtmlRenderer verbatim - ready to drop straight into an IG Publisher markdown pagecontent page (or any other static page) with no further wrapping. Deliberately .md, not .xml: the IG Publisher infers "xml" page generation (XSLT-templated) from a bare .xml extension rather than "html" (proper XHTML, brought across unchanged), and running an already-rendered HTML fragment through that XSLT transform fails outright - confirmed against a real IG Publisher build. A .md page's raw HTML block is passed through unprocessed by the markdown renderer, which is the effect actually wanted here.