Easy Predictive Analytics can also use the REST API to update a prediction model (perform model retraining with the latest data). This process is executed by the series of API operations indicated below. 1.Retrain Data Upload API: Uploads data and executes retraining 2.Retrain Status Check API: Checks the status of retraining and retrieves a URL for downloading the results file after training 3.Retrain Model Switch API: Switches the model with API created to the retrained model

Retrain Status Check API

Retrieves the status of retraining. This enables you to retrieve the URL for downloading the retraining results file when retraining is complete.

URL

https://developer-api.easy-predictive-analytics.sony.com/v1/external/retrain_jobs/{retrain_job_id}

The HTTP GET method is used.

Request

header

path parameter

query parameter

Not used.

body

Not used.

Response

body

The format is indicated below.

{
  "status": str,
  Optional("model_id"): int,
  Optional("message"): str,
  Optional("download_url"): str,
  "metadata": {
    "retrain_job_id": str
  }
}

error message

Execution Example 1

An example of executing the Retrain Status Check API in Python is indicated below.

This assumes that the Get retrain job dataset upload url API has been executed.

import requests
​
retrain_job_id = 'YOUR_RETRAIN_JOB_ID'
api_url = f'https://developer-api.easy-predictive-analytics.sony.com/v1/external/retrain_jobs/{YOUR_RETRAIN_JOB_ID}'
api_key = 'YOUR_API_KEY'
​
headers = {
    'x-api-key': api_key
}
​
response = requests.get(api_url, headers=headers)
response_json = response.json()
status = response_json['status']

download_url = None
while True:
  if response_json['status'] == 'job_finished':
    download_url = response_json['download_url']

  if download_url:
    response = requests.get(download_url)
    with open(result_file_path, 'w') as f:
      print(response.content)
      

Execution Example 2

An example of executing the curl command from the command prompt of the Retrain Status Check API is indicated below. This assumes that the Get retrain job dataset upload url API has been executed.

1.Launch the command prompt.

2. Change the view to UTF-8.
$ chcp 65001

3. Make a request.
$ curl -o "response.json" -H "x-api-key:{YOUR_API_KEY}" -X GET -G https://developer-api.easy-predictive-analytics.sony.com/v1/external/retrain_jobs/{retrain_job_id}
ex) $ curl -o "response.json" -H "x-api-key:xxxx-xxx-xxx-xxxx" -X GET -G https://developer-api.easy-predictive-analytics.sony.com/v1/external/retrain_jobs/{xxxx-xxxx-xxxx-xxxx}

4. Check the status in the response.json file.
If the status is job_finished, retrieve download_url.

5. Retrieve the retraining results.
$ curl -G GET {download_url retrieved in 4}