0.7.5

• add throttling on upload progress event handlers

0.7.4

• Job is now asynchronous
• Wait for successive sequential jobs (follow api changes)
• Use bluebird instead of swissknife
• Hide get signed url (for chunks) errors

0.7.3

• add bucket / container info to source creation request

0.7.2

• remove deduplicate of upload errors (because of CORS constraint)

0.7.1

• improve ES5 support
• avoid need for Symbol polyfill

0.7.0

new API error format

upload > 5Go

• chunks
• auto-retry (manage small network issues)
• manual retry
• cancel

0.6.1

• Fix Datasets#isParent not considering scoresets as parents

0.6.0

Follow addition of datasets types in kml-api

• type: subset
• type: scoreset
• type: uploaded_dataset
• See PredicSis/kml-api#341

0.5.0

Name splits

Function dataset.split() returns an promised object with two properties:

{
  "train": { "id": "...", },
  "test": { "id": "...", }
}

0.4.3

Build univariate supervised report with train subset instead of parent dataset.