w3ai-sdk 0.0.7

W3AI Flatform API

w3ai-sdk

Description

Using w3ai-sdk package to querry some of data from W3AI Flatform using APIKey. For example , Getting user balance, creating a task, getting task result, getting statistics and history of models, checking API permission ...

Installation

pip install w3ai-python-sdk

Setup

To start, simply require the W3AI-SDK and set up an instance with your W3AI API Keys.Please checking out your W3AI Page. In the example below show you how to using the SDK.

from 'w3ai-python-sdk' import W3AIClient
client = W3AIClient(API_KEY, SERVER_ENDPOINT)

Usage

API Key Permission

client.get_api_key_permission(API_KEY)

Response :

{
  status: 'success',
  data: {
    limit_models: true,
    models: [ '7da922fe-02b9-4534-b218-c5b7c79d2195' ]
  }
}

Models infomation

client.get_model_info(MODEL_ID)

Response :

{
  status: 'success',
  data: {
    model_id: '7da922fe-02b9-4534-b218-c5b7c79d2195',
    api_price: 0.12342134,
    model_description: 'this is best model',
    commit_hash: 'd1c366e0c6bbfdd28cb31301852030c113013576',
    input_format: { input: [Object] },
    output_format: { output: [Object] }
  }
}

Model serving info

client.check_model_is_serving(MODEL_ID)

Response :

{ status: 'success', data: { consumers: 1, serving: true } }

Task executing cost info

client.get_model_task_cost(MODEL_ID)

Response :

{
  status: 'success',
  message: 'Success',
  data: { cost: 0.011000000000000001, symbol: '$', unit: 'USD' }
}

Creating a task

Argument exmaple:

FILE_LIST = [
    {
      "key": "input",
      "data": "file_stotage_url",
      "name": "image.jpg"
    },
    {
      "key": "input2",
      "data": "file_stotage_url2",
      "name": "image2.jpg"
    }
  ]

INPUT_PARAMS = {
  "input": {
    "type": "file",
    "samples": [],
    "required": true,
    "mime_type": [
      "image/*"
    ]
  }
}

MODEL_ID = "7da922fe-02b9-4534-b218-c5b7c79d2195"

client.create_task(files= FILE_LIST, input_params=INPUT_PARAMS, model_id= MODEL_ID)

Response :

{
  status: 'success',
  message: 'Success',
  data: "e52fa152-351d-4a40-a4ce-83a4ff11d87e"
}

Getting task result

client.get_task_result(task_id = TASK_ID)

Response :

{
  "result": {
    "additionalProp1": {}
  },
  "status": "sucess",
  "success": true
}