Reset exp-params and input-exp-params on $draft

[Chris0296]

Summary
$release splits an artifact's expansion parameters into two copies: the author-specified values are snapshotted into a new input-exp-params contained resource (referenced via the cqf-inputParameters extension), while the live exp-params (referenced via cqf-expansionParameters) is then enriched with processing-time values (e.g. pinned system-versions) for use during that release. There was previously no mechanism to undo this, drafting a released artifact left the processing-time values sitting in exp-params and the stale input-exp-params snapshot (plus its extension) hanging around indefinitely, even though they were purely an artifact of a completed release cycle. This could lead to duplicate extensions being applied.

This PR makes $draft reverse that split: it restores the author's original input back onto exp-params and removes the now-redundant input-exp-params resource and extension, so the artifact is back to a clean, author-only state in preparation for its next release.

Changes

• Added restoreInputExpansionParams(IKnowledgeArtifactAdapter) to new version specific DraftVisitor classes, wired into DraftVisitor.visit(), dispatched by FHIR version, alongside the existing release-label/approval-date reset that already happens at the start of drafting.

Tests

• Full $release -> $draft workflow, confirming exp-params goes from 3 processing-time-enriched parameters back down to the 1 author-specified parameter on $draft, and input-exp-params/its extension are gone.
• Unit test of happy path
• Unit test for case where authored params is empty, confirming exp-params is set to empty
• Unit test for case where input-exp-params extension is present but there is no matching contained resource, confirms cleanup still happens
• Unit test for no-op case where there is no input-exp-params extension, confirms artifact is left as is.

[github-actions]

Formatting check succeeded!

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
95.7% Coverage on New Code
15.6% Duplication on New Code

See analysis details on SonarQube Cloud