fix: overhaul the platform skeleton, add a serverless deployment track

An end-to-end audit found the repo could not build, test or run as shipped. This
fixes every finding, then adds a Cloud Run track so the demo costs about £1/month
idle instead of ~£150.

CI (red on its first run)
- api: setuptools could not build the package (flat layout with app/ and alembic/)
- web: missing @types/node; `vitest run` exited 1 with no test files
- pipeline: the stub run needed a gitignored VCF, and no process had a stub block
- ruff pinned, mypy configured, DB tests on real Postgres (pgserver locally, service in CI)

ML serving (scores were meaningless)
- the registered model now carries its own feature engineering and returns predict_proba,
  so serving sends raw columns and cannot drift from training
- resolve by registry alias (stages are deprecated in MLflow 3) and record the real
  version; re-scoring upserts instead of failing on the unique constraint
- ClinVar labels parsed from VEP's lowercase terms

Pipeline
- exact ref/alt recovered from a CHROM_POS_REF_ALT VCF ID; loading is idempotent
- job status reaches running/failed/succeeded, so the UI stops polling dead jobs
- DATABASE_URL travels in the environment or a Nextflow secret, never on a command line
- VEP cache and plugins staged as inputs; the gcp profile runs tasks on Google Batch

Deployment
- the API serves /api (matching the ingress); the web app reads its API URL at runtime
- migrations run in an init container under a Postgres advisory lock
- terraform: custom VPC shared with Batch, private Cloud SQL, API enablement, Workload
  Identity bindings, Secret Manager, deletion protection
- serverless track, now the default: Cloud Run services scaling to zero, a Cloud Run job
  for the Nextflow driver, and Neon or Cloud SQL behind one DATABASE_URL secret. GKE and
  Argo remain, behind -var deploy_kubernetes=true. See docs/cloud.md.

Correctness and security
- 409 on duplicate sample names, 422 on bad paging, natural chromosome ordering, wider
  VEP text columns, enum dropped on downgrade, the sample's assembly actually used
- vcf_uri restricted to gs:// objects or files under the data root, blocking option injection
- CORS restricted to configured origins; `make down` no longer deletes volumes

Data
- docs/data.md records the peer-reviewed, openly licensed sources (GIAB HG002, ClinVar,
  gnomAD) with citations and an honest evaluation plan; `make data` fetches a chr22 slice

Verified: api 50 tests, ml 18, loader 16, web 12; ruff, mypy, svelte-check, terraform
validate and both kustomize overlays clean.

infra/argo-workflows/annotate.yaml

@@ -1,37 +1,54 @@
-# Triggered by an Argo Events sensor listening on the Pub/Sub topic "vcf-uploaded".
+# Submitted by the Argo Events sensor in events.yaml for each "vcf-uploaded" Pub/Sub message.
+# The Nextflow driver runs here; its tasks run on Google Batch (see the gcp profile in
+# pipeline/nextflow.config). Image names are rewritten by the gcp overlay and bumped by CI.
 apiVersion: argoproj.io/v1alpha1
 kind: WorkflowTemplate
-metadata: { name: annotate-vcf, namespace: rarelens }
+metadata: { name: annotate-vcf }
 spec:
+  serviceAccountName: rarelens-pipeline
   entrypoint: nextflow
+  onExit: exit-handler
   arguments:
     parameters:
       - { name: job_id }
       - { name: vcf_uri }
+      - { name: assembly, value: GRCh38 }
   templates:
     - name: nextflow
-      inputs:
-        parameters: [{ name: job_id }, { name: vcf_uri }]
-      serviceAccountName: rarelens-pipeline
       container:
-        image: europe-west2-docker.pkg.dev/PROJECT/rarelens/pipeline:latest
-        command: [nextflow]
+        image: rarelens/pipeline
         args:
           - run
           - /pipeline/main.nf
           - -profile
           - gcp
           - --vcf
-          - "{{inputs.parameters.vcf_uri}}"
+          - "{{workflow.parameters.vcf_uri}}"
           - --job_id
-          - "{{inputs.parameters.job_id}}"
-          - --db_url
-          - "$(DATABASE_URL)"
-        envFrom: [{ secretRef: { name: api-secrets } }]
-        resources: { requests: { cpu: "2", memory: 4Gi } }
-    - name: score
-      # Optional GPU step for the deep-learning baseline; Autopilot schedules on an L4 node.
-      nodeSelector: { cloud.google.com/gke-accelerator: nvidia-l4 }
+          - "{{workflow.parameters.job_id}}"
+          - --assembly
+          - "{{workflow.parameters.assembly}}"
+        # GCP_PROJECT / GCS_BUCKET / GCP_REGION feed params in nextflow.config.
+        envFrom: [{ configMapRef: { name: pipeline-config } }]
+        resources: { requests: { cpu: "1", memory: 2Gi } }
+
+    # The loader marks success; anything else (Nextflow error, OOM, eviction) is marked here
+    # so the UI never polls a dead job forever.
+    - name: exit-handler
+      steps:
+        - - name: mark-failed
+            template: mark-failed
+            when: "{{workflow.status}} != Succeeded"
+    - name: mark-failed
       container:
-        image: europe-west2-docker.pkg.dev/PROJECT/rarelens/ml:latest
-        resources: { limits: { nvidia.com/gpu: 1 } }
+        image: rarelens/loader
+        command: [set_job_status.py]
+        args:
+          - --job-id
+          - "{{workflow.parameters.job_id}}"
+          - --status
+          - failed
+          - --log
+          - "Argo workflow {{workflow.name}} ended {{workflow.status}}"
+        envFrom: [{ secretRef: { name: pipeline-secrets } }]
+        resources: { requests: { cpu: 100m, memory: 256Mi } }