Add lineage v1beta1 schema

[jorgee]

Summary

• Adds the JSON Schema (Draft 2020-12) for the lineage data model wire format emitted by nf-lineage (LinEncoder) at lineage/v1beta1/schema.json.
• Discriminated oneOf over the six LinSerializable subtypes (FileOutput, TaskOutput, TaskRun, Workflow, WorkflowOutput, WorkflowRun), each wrapped in the runtime envelope {version, kind, spec} with version and kind as const discriminators.
• Shared types (Checksum, DataPath, Parameter, Workflow) deduplicated via $defs/$ref; OffsetDateTime/Path modelled as string (with date-time format where applicable); model withSerializeNulls(true) accommodated via ["X", "null"] typing.
• Class titles and descriptions sourced from the corresponding GroovyDoc in the model classes.

Tests

lineage/v1beta1/tests/:

• Valid — valid_file_output.json, valid_task_output.json, valid_task_run.json, valid_workflow_run.json. The first three are real .data.json documents from an nf-core/rnaseq run; the workflow_run is trimmed for readability.
• Invalid — invalid_missing_envelope.json (bare spec body at root), invalid_wrong_kind.json (unknown kind), invalid_wrong_version.json (unsupported version).

Wired lineage/v1beta1 into validate.sh; updated README.md to list the lineage and module schemas.

Test plan

• ./validate.sh passes locally — confirmed: all 7 lineage cases behave as expected (4 valid pass, 3 invalid rejected); rest of repo unaffected.
• Additionally validated against 1485 .data.json files from a full nf-core/rnaseq lineage run — all pass.

Source of truth

The schema is generated by a Gradle task in nextflow-io/nextflow (./gradlew :nf-lineage:generateLineageSchema). When the lineage model changes, regenerate there and re-sync this repo.

[pditommaso]

Any chance to make it the schema a bit more readable (maybe just asking claude to reformat it)?

[jorgee]

Any chance to make it the schema a bit more readable (maybe just asking claude to reformat it)?

I have fixed some issues with arrays split into several lines and put the title and description at the beginning of the definitions.

I also tried to compress the anyOf with type: null and refs but the forced linting is expanding it again. If the applied fixes are not enough, an alternative could be adding extra definitions for these anyOf nullable statements. (such as ChecksumNullable). It could improve the readability but at the same time it will add types that are not part of the lineage model. So, not sure what is better.