GUPRI

GUPRI (Global Unique Persistent Resource Identifier) is Roche’s internal identifier backbone. Every artifact produced by the roche-data pipeline — entities, contracts, APIs, policies — receives a globally unique, resolvable URI through GUPRI. This enables unambiguous cross-system references: when Collibra, Snowflake, and an API all reference the same entity, they share the same GUPRI identifier.

Connection Details

Property	Value
URL	TBD (awaiting A02 resolution)
Auth method	TBD (to be clarified with GUPRI team)
Network	Roche corporate network (VPN required)
Access task	A02
GitHub issue	#16
URI prefix	`urn:gupri:roche`

Environment Variables

Variable	Source	Description
`GUPRI_BASE_URL`	Vault `common/gupri`	Base URL for GUPRI REST API
`GUPRI_URI_PREFIX`	Vault `common/gupri`	URI namespace prefix (`urn:gupri:roche`)

CLI Modules

Module	Usage
`rdt-model-gupri`	Registers new artifacts, resolves existing identifiers
`rdt-model-contract`	Embeds GUPRI URIs in data contract metadata
`rdt-model-api`	References GUPRI URIs in OpenAPI spec `info.x-gupri-id`
`rdt-model-register`	Uses GUPRI URIs for Collibra lineage relationships
`rdt-model-sdk`	Embeds GUPRI URIs in SDK metadata
`rdt-model-event`	Includes GUPRI URIs in lifecycle event payloads

URI Format

urn:gupri:roche:{entity}:{artifact-type}:v{version}

Examples:

urn:gupri:roche:waste-tracking:entity:v1
urn:gupri:roche:waste-tracking:contract:v1
urn:gupri:roche:waste-tracking:api:v1
urn:gupri:roche:waste-tracking:policy:v1

Access Verification

Script: scripts/access/check-gupri.sh

Required tools: curl

Checks performed:

HTTP connectivity to GUPRI endpoint (if GUPRI_BASE_URL is set)
HTTP status code validation (200/204 = pass, 401/403 = auth needed)

Expected state: skip — A02 not yet resolved.

Current Status

Access: Blocked — awaiting service account provisioning from GUPRI team
Workaround: StubGupriClient generates deterministic URIs using the pattern urn:gupri:roche:{entity}:{artifact}:v1
Impact: Does not block development — all modules work with stub-generated URIs in --dry-run mode

Output

The rdt-model-gupri module produces models/{entity}/gupri.yaml:

entity_id: waste-tracking
entity_uri: urn:gupri:roche:waste-tracking:entity:v1
contract_uri: urn:gupri:roche:waste-tracking:contract:v1
api_uri: urn:gupri:roche:waste-tracking:api:v1
created_at: "2026-05-01T10:00:00Z"