tonic-api 2.0.0

Wrappers around the Tonic.ai API

Overview

This library contains useful wrappers around the Tonic.ai API.

Usage

Instantiate the API wrapper using the following code:

from tonic_api.api import TonicApi

# Do not include trailing backslash in TONIC_URL
api = TonicApi(TONIC_URL, API_KEY)

Once instantiated, the following endpoints are available for consumption. Note that available endpoints and response types are limited. Available fields may be severely limited compared to the current Tonic API.

TonicApi:
    get_workspace(workspace_id) => Workspace

Workspace:
    id => string
    name => string
    models => Model[]

    train(model_id) => new job ID
    get_trained_model_by_training_job_id(job_id) => TrainedModel (or None if training is not done)
    get_training_status_by_training_job_id(job_id) => TrainingJobStatus
    get_most_recent_trained_model_by_model_id => TrainedModel

    describe() => debugger helper for printing fields

Model:
    id => string
    name => string
    query => string
    parameters => {}
    encodings => {}

    describe() => debugger helper for printing fields

TrainedModel:
    id => string
    job_id => string
    model => Model

    sample(num_rows) => pandas DataFrame (defaults to 1 row if num_rows not provided)
    sample_source(num_rows) => pandas DataFrame (defaults to 1 row if num_rows not provided). Note: NOT randomized. Upper limit is limited to row count in source.

    get_numeric_columns() => string[]
    get_categorical_columns() => string[]

    describe() => debugger helper for printing fields

TrainingJobStatus:
    state => string
    error => string[] or None
    tasks => {}

    current_epoch_progress() => dict or None

    describe() => debugger helper for printing fields