generated: '2026-07-21'
method: derived
source: openapi/strand-ai-openapi-original.json
note: >-
  Entity-relationship graph derived from the OpenAPI schemas and operation flow. Ids are UUIDs
  (no documented id-prefixes). The core flow: an Upload becomes a ready Sample, a Sample feeds a
  prediction Job, and a completed Job produces OME-Zarr Results.
entities:
- name: Upload
  id: uploadId (uuid)
  fields: [gcsPath, uploadUrl, existing, status, widthPx, heightPx, dimensionsSource]
  note: resumable upload of a whole-slide image; de-duplicated on contentSha256
- name: Sample
  id: id (uuid)
  fields: [expiresAt, expiresAtSource, retentionChangedAt, retentionChangedBy, batchId]
  note: a stored, ready slide; subject to expiration/retention and Trash restore
- name: Job
  id: id (uuid)
  fields: [status, progress, reservedCredits, markers, model, metadata, createdAt, startedAt, completedAt, errorMessage, resultsAvailable]
  note: a Lattice inference job over a Sample for a set of markers
- name: Results
  id: resultUrl
  fields: [resultBasePath, expiresAt]
  note: signed OME-Zarr store output of a completed Job
relationships:
- from: Upload
  to: Sample
  type: has_one
  via: uploadId
  note: a finalized (ready) Upload materializes as a Sample
- from: Job
  to: Sample
  type: belongs_to
  via: uploadId
  note: submitPrediction references the uploadId of a ready Sample
- from: Job
  to: Results
  type: has_one
  via: jobId
  note: a completed Job exposes one OME-Zarr Results store (resultsAvailable)
- from: Sample
  to: Sample
  type: has_many
  via: batchId
  note: bulk expiration groups samples under a shared batchId