Dedicated inference functionalities

[rarora-it]

For Update:

TestRunDedicatedInferenceUpdate - Updates an endpoint with a JSON spec file; verifies correct request is sent to the service
TestRunDedicatedInferenceUpdate_WithHuggingFaceToken -Updates with both spec and and hugging face token, flag; verifies secrets are included in the request
TestRunDedicatedInferenceUpdate_MissingID - Returns an error when no dedicated inference ID argument is provided

Update all commands

Update with a JSON spec
doctl dedicated-inference update --spec spec.json
Update with a YAML spec and Hugging Face token
doctl dedicated-inference update --spec spec.yaml --hugging-face-token ""
Update using stdin
cat spec.json | doctl dedicated-inference update --spec -
Show only specific columns
doctl di u --spec spec.yaml --format ID,Name,Status

For list:

TestRunDedicatedInferenceList - Lists all endpoints with no filters (empty region/name)
TestRunDedicatedInferenceList_WithRegion - Lists endpoints filtered by -- region nyc2
TestRunDedicatedInferenceList_WithName - Lists endpoints filtered by --name

List commands

List all endpoints
doctl dedicated-inference list
Filter by region
doctl dedicated-inference list --region nyc2
Filter by name
doctl dedicated-inference list --name my-endpoint
Short alias with format
doctl di ls --format ID,Name,Region,Status

For CreateToken:

TestRunDedicatedInferenceCreateToken - Creates an auth token with -- token name and verifies the correct request and token display
TestRunDedicatedInferenceCreateToken_MissingID - Returns an error when no dedicated inference ID is provided

Create tokens commands

Create a new auth token
doctl dedicated-inference create-token --token-name
Show only ID and Value
doctl dedicated-inference create-token --token-name --format ID,Value

For ListTokens:

TestRunDedicatedInferenceListTokens - Lists auth tokens for a given dedicated inference ID; verifies two tokens are displayed
TestRunDedicatedInferenceListTokens_MissingID - Returns an error when no dedicated inference ID is provided

List token commands

List all tokens for an endpoint
doctl dedicated-inference list-tokens
Show specific columns
doctl dedicated-inference list-tokens --format ID,Name,CreatedAt

For RevokeTokens:

TestRunDedicatedInferenceRevokeToken - Revokes a token with --force verifies both diID and tokenID are passed to the service
TestRunDedicatedInferenceRevokeToken_MissingArgs - Returns an error when no arguments are provided at all
TestRunDedicatedInferenceRevokeToken_MissingTokenID - Returns an error when only the dedicated inference ID is provided (missing token ID)

Revoke token endpoints

Revoke a token (with confirmation prompt)
doctl dedicated-inference revoke-token
Revoke without confirmation
doctl dedicated-inference revoke-token --force

For GetSizes:

TestRunDedicatedInferenceGetSizes - Retrieves GPU sizes; verifies the response includes GPU slug, pricing, regions, CPU, memory, and GPU details

Get sizes all commands

List all available GPU sizes and pricing
doctl dedicated-inference get-sizes
Using alias
doctl di gs
Show only slug, price, and regions
doctl dedicated-inference get-sizes --format GPUSlug,PricePerHour,Regions

for GetGPUModelConfig:

TestRunDedicatedInferenceGetGPUModelConfig - Retrieves GPU model configs; verifies two models returned with slug, name, gated status, and compatible GPU slugs

GPU all commands

List all supported GPU model configurations
doctl dedicated-inference get-gpu-model-config
Using alias
doctl di ggmc
Show only model slug and gated status
doctl dedicated-inference get-gpu-model-config --format ModelSlug,IsModelGated

Also added the integration tests for all.

[anup-deka]

LGTM 🚀