Milab-6847: fix calculations with empty abundance

[erohinaelena]

Greptile Summary

The PR fixes usage normalization for empty or all-null abundance data by preserving the analysis table’s declared schema and updates the associated build tooling.

• Replaces the untyped TSV build with buildForPT() and lets wf.frame consume its embedded format and schema metadata.
• Updates CI identities to V/J Gene Usage and moves CI from Node 20 to Node 22.
• Upgrades @platforma-sdk/block-tools from 2.12.6 to 2.14.3 and records patch releases for the workflow and block packages.
• Important touched terms:
  • Abundance — the per-clonotype numeric weight used for weighted usage; its declared numeric type is now preserved even when every value is null or no rows remain.
  • Weighted usage — a gene group’s abundance divided by the matching sample’s total abundance; the change prevents its division expression from receiving a string-typed abundance column.
  • Unweighted usage — a gene group’s row count divided by the sample’s total row count; its calculation is unchanged.
  • PTabler schema — metadata describing table column and axis types; it is now carried by the table produced with buildForPT().
  • Analysis table — the sample, clonotype, V gene, J gene, and abundance input to the calculation stage; its construction changes from inferred TSV typing to declared typing.

Confidence Score: 5/5

The PR appears safe to merge, with no concrete blocking or independently actionable non-blocking issue identified.

The schema-aware table producer and frame consumer are changed together, their headers remain aligned with all usage aggregations, and the tooling updates reveal no established reachable failure.

Important Files Changed

Filename	Overview
workflow/src/main.tpl.tengo	Builds the calculation input with buildForPT() so declared axis and column types survive empty or all-null datasets.
workflow/src/usage.lib.tengo	Loads the structural PTabler input directly instead of overriding it as a plain TSV.
pnpm-workspace.yaml	Upgrades @platforma-sdk/block-tools to 2.14.3.
.github/workflows/build.yaml	Corrects the application identity and moves the shared build workflow to Node 22.
.github/workflows/mark-stable.yaml	Corrects the stable-release application identity and moves that workflow to Node 22.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart LR
  A[Resolved sample, clonotype, V, J, and abundance columns] --> B[Build analysis table with buildForPT]
  B --> C[Schema-carrying PTabler input]
  C --> D[Load with wf.frame]
  D --> E[Calculate per-sample totals]
  D --> F[Aggregate V, J, and V/J groups]
  E --> G[Normalize weighted and unweighted usage]
  F --> G
  G --> H[Publish typed usage results]

Loading

_{Reviews (1): Last reviewed commit: "MILAB-6847: run CI on node 22, fix copy-..." | Re-trigger Greptile}

Context used (3)

• Knowledge Base — Gene usage workflow
• Knowledge Base — Workflow orchestration
• Knowledge Base — Gene usage calculations