Releases: watson-developer-cloud/swift-sdk
v0.23.4
This release fixes a bug with the updateClassifier
method in Visual Recognition.
v0.23.3
This release improves support for Swift 4.1 and Xcode 9.3.
v0.23.2
This release fixes a few bugs for Linux compatibility:
- Only
import CoreML
in supported environments - Include Assistant tests in
LinuxMain.swift
It also updates the readme for API changes introduced in v0.22.0.
v0.23.1
This release fixes a bug with created
and updated
dates when using the 2018-02-16
version of Assistant or Conversation.
v0.23.0
This update adds Apple Core ML support for Visual Recognition. With Core ML you can download a trained classifier model and use it for offline image classification.
v0.22.0
This update adds the Assistant service.
In addition, the following services have been updated to support their latest release. Note that some APIs have changed in order to improve uniformity between the Watson SDKs.
- Conversation
- Language Translator
- Natural Language Classifier
- Natural Language Understanding
- Personality Insights
- Text to Speech
- Tone Analyzer
- Visual Recognition
The following deprecated services have also been removed from the SDK:
- AlchemyDataNewsV1
- AlchemyLanguageV1
- AlchemyVisionV1
- DialogV1
- DocumentConversionV1
- PersonalityInsightsV2
- RelationshipExtractionV1Beta
- RetrieveAndRankV1
- TradeoffAnalyticsV1
v0.21.0
- Change uses of
NSLocalizedFailureReasonErrorKey
toNSLocalizedDescriptionKey
for Linux compatibility - Only lint the
Source/Service
folder for framework build targets (fixes a bug with carthage --no-use-binaries) - Fix style errors reported by
swiftlint
(fixes a bug with carthage --no-use-binaries) - Improve test infrastructure for simulator and Linux
v0.20.0
This release extracts Starscream to be an external dependency. Starscream adds support for WebSockets sessions and is required for use with the Speech to Text service. Because it is an external dependency, developers using the SDK must now link and embed it in their projects. For more information, see Readme: Speech to Text.
This release also includes the following additional changes:
- Deprecate Document Conversion
- Add conditional compilation by platform to improve future compatibility with Linux
- Set client WebSockets timeout to match the Speech to Text service (30s)
- Improve code style using SwiftLint
v0.19.0
This release includes the following changes:
- Update Conversation to support 2017-05-26 release with Codable models
- Update Natural Language Understanding to support 2017-02-27 release with Codable models
- Update Visual Recognition to deprecate similarity search operations
- Update Speech to Text to fix a bug with interim speaker labels
- Remove RestKit target (files are now included directly in each service's target)
- Update Package.swift for swift-tools-version:4.0
- Update Starscream dependency
- Update libopus dependency to v1.1.3
- Add extensions to support [String: Any] metadata with Codable models
- Add extensions to support additional properties with Codable models
v0.18.0
This release updates the SDK to use Swift 4.