Prompt optimization operations

createpromptoptimizationjob(instance::MLFlow, experiment_id::String,
    source_prompt_uri::String, config::Dict{String,Any};
    tags::Array{PromptOptimizationJobTag}=PromptOptimizationJobTag[])

Create a new prompt optimization job.

This endpoint initiates an optimization run with the specified configuration. The optimization process runs asynchronously and can be monitored via getpromptoptimizationjob.

Arguments

• instance: MLFlow configuration.
• experiment_id: ID of the MLflow experiment to track the optimization job in.
• source_prompt_uri: URI of the source prompt to optimize (e.g., "prompts:/my-prompt/1").
• config: Configuration for the optimization job as a dictionary.
• tags: Optional tags for the optimization job.

Returns

An instance of type PromptOptimizationJob.

getpromptoptimizationjob(instance::MLFlow, job_id::String)

Get the details and status of a prompt optimization job.

Returns the job configuration, current status, progress statistics, and the best prompt if the optimization has completed.

Arguments

• instance: MLFlow configuration.
• job_id: The unique identifier of the optimization job (same as run_id).

Returns

An instance of type PromptOptimizationJob.

searchpromptoptimizationjobs(instance::MLFlow, experiment_id::String)

Search for prompt optimization jobs.

Returns a list of optimization jobs matching the specified filters.

Arguments

• instance: MLFlow configuration.
• experiment_id: ID of the MLflow experiment to search optimization jobs in.

Returns

Vector of PromptOptimizationJob entities.

cancelpromptoptimizationjob(instance::MLFlow, job_id::String)

Cancel an in-progress prompt optimization job.

If the job is already completed or cancelled, this operation has no effect.

Arguments

• instance: MLFlow configuration.
• job_id: The unique identifier of the optimization job to cancel.

Returns

An instance of type PromptOptimizationJob.

deletepromptoptimizationjob(instance::MLFlow, job_id::String)

Delete a prompt optimization job and its associated data.

This permanently removes the job and all related information.

Arguments

• instance: MLFlow configuration.
• job_id: The unique identifier of the optimization job to delete.

Returns

true if successful. Otherwise, raises exception.