Data AppendΒΆ

The data_append session, is an implementation that appends additional dataset(s), to a previous implemented data_new session. Like the data_new session, there are two ways to upload data, to the /load-data endpoint:

Note: the content of each of the above examples, can be substituted for the data attribute, in a given POST request:

import requests

endpoint = 'https://192.168.99.101:9090/load-data'
headers = {
    'Authorization': 'Bearer ' + token,
    'Content-Type': 'application/json'
}

requests.post(endpoint, headers=headers, data=json_string_here)

Note: more information, regarding how to obtain a valid token, can be further reviewed, in the /login documentation.

The following properties define the above data attribute:

  • collection: collection of dataset documents, used as a reference to add additional dataset documents into
  • dataset_type: corresponds to one of the following types:
    • dataset_url: indication that the supplied dataset will be url references
    • file_upload: indication that the supplied dataset(s) will be defined as a json string within the dataset attribute
  • session_type: corresponds to one of the following session types:
    • data_new
    • data_append
    • model_generate
    • model_predict
  • model_type: the type of model to perform on:
    • svm
    • svr
  • session_id: corresponds to the id associated with the original data_new uploaded dataset(s), being appended to.
  • dataset: the supplied dataset, contingent upon the dataset_type