generated: '2026-08-27'
method: derived
source: https://github.com/sightmachine/sightmachine-sdk
note: >-
  Derived from the first-party Python SDK's entity modules and the endpoint paths they
  call, not from an OpenAPI (none is published). Every entity and endpoint below appears
  verbatim in smsdk source at commit cb8fe167ed7f75975e6bc4f0f82306784ac2b169. Field-level
  schemas are deliberately NOT enumerated: they are tenant-specific and are discovered at
  runtime through get_machine_schema(), so listing them would be invention.
entities:
- name: MachineType
  description: >-
    A class of production asset (e.g. a lasercut station, a filler). Carries the metadata
    and the statistic set shared by every machine of that type.
  endpoint: /v1/selector/assets
  sdk_methods:
  - get_machine_type_names
  - get_machine_types
- name: Machine
  description: An individual physical asset on a line, belonging to a MachineType.
  endpoint: /v1/selector/assets
  sdk_methods:
  - get_machine_names
  - get_machine_schema
- name: Cycle
  description: >-
    One production cycle on a machine — the core time-series record of the platform, and
    the primary unit analysts retrieve.
  endpoint: /v1/datatab/cycle
  sdk_methods:
  - get_cycles
- name: Downtime
  description: A recorded interval in which a machine was not producing, with its reason coding.
  endpoint: /v1/datatab/downtime
- name: Part
  description: >-
    A discrete produced item, typed by part_type. Field metadata for a part type is fetched
    per type rather than globally.
  endpoint: /v1/datatab/part
  field_endpoint: /v1/selector/datatab/part/{part_type}/field
- name: Line
  description: An ordered arrangement of machines forming a production line.
  endpoint: /v1/datatab/line
- name: RawData
  description: Uncontextualised source records as ingested by FactoryTX before modelling.
  endpoint: /v1/datatab/raw_data
- name: KPI
  description: A computed production metric selectable as a chart axis.
  endpoint: /v1/selector/datavis/kpi/y_axis
- name: Dashboard
  description: >-
    A saved collection of panels/widgets. The Dashboard API resolves a dashboard into its
    fully-computed panel values for external consumption.
  endpoint: /v1/obj/dashboard/{dashboard_id}/full_values
  panel_fields:
  - data
  - id
  - name
  - status
  - type
- name: Workspace
  description: >-
    A schema-scoped development environment. Selectable through the SDK and exposed to
    ODBC clients as a PostgreSQL schema named workspace_{workspace_id}; production data
    lives in the `sightmachine` schema of the `tenant_storage` database.
  sdk_methods:
  - select_workspace_id
relationships:
- from: Machine
  to: MachineType
  type: belongs_to
  via: machine type name
- from: MachineType
  to: Machine
  type: has_many
- from: Cycle
  to: Machine
  type: belongs_to
  via: asset/machine identifier
- from: Downtime
  to: Machine
  type: belongs_to
  via: asset/machine identifier
- from: Line
  to: Machine
  type: has_many
- from: Part
  to: Cycle
  type: belongs_to
  via: cycle association on the production record
- from: Dashboard
  to: KPI
  type: has_many
  via: panel/widget configuration
- from: Workspace
  to: Cycle
  type: has_many
  via: schema scoping (workspace_{id} vs sightmachine)
gaps:
- >-
  No id-prefix convention is documented (unlike, say, Stripe's `cus_`/`ch_`), and no
  canonical object reference page exists.
- >-
  Per-entity field schemas are tenant-specific and only discoverable at runtime via
  get_machine_schema(); they are not published and are not reproduced here.