3.6.0

Version 3.6.0

2017-03-07

• New: Automatic releases
• Change: Update OkHttp to v3.6.0
• New: Speech to Text: Add addCorpus() method, deprecate addTextToCustomizationCorpus() method
• New: Speech to Text: Add getCorpus() method
• New: Speech to Text: Add sort parameter to getWords() method
• New: Speech to Text: Add SupportedFeatures for SpeechModel object
• New: Speech to Text: Add and fix constructors for Word object
• New: Speech to Text: Add count to WordData object
• New: Speech to Text: Add confidence to SpeakerLabel object
• New: Speech to Text: Add warnings to RecognitionJob object
• Fix: Speech to Text: Enable customization_id for recognizeUsingWebSocket() method
• Fix: Speech to Text: Fix allowOverwrite parameter of addTextToCustomizationCorpus() method
• Fix: Speech to Text: Fix error field for WordData object
• New: Text to Speech: Add createCustomVoiceModel() and updateCustomVoiceModel() methods, deprecate saveCustomVoiceModel() method
• New: Text to Speech: Add addWords() and addWord() methods, deprecate saveWords() method
• New: Text to Speech: Add getWord() method
• New: Text to Speech: Add getCustomVoiceModel() method with CustomVoiceModel argument
• New: Text to Speech: Add deleteWord() method with String argument
• New: Text to Speech: Add words to CustomVoiceModel object
• New: Text to Speech: Add part_of_speech to CustomTranslation object
• Fix: Text to Speech: Enable customization results for getVoice() method
• New: Tradeoff Analytics: Add findPreferableOptions parameter to dilemmas() method
• New: Tradeoff Analytics: Add PreferableSolutions object to Resolution object
• New: Tradeoff Analytics: Add DominatingOption object to Solution object

3.5.3

Version 3.5.3

2016-12-15

• New: Added Discovery v1

3.5.2

Version 3.5.2

2016-12-08

• New: Added Speaker Labeling (diarization) added
• New: Added Conversation + Tone Analyzer example
• New: Added byte array support to classify an in memory image
• Fix: Text to Speech: getPronunciation() and getVoice now accept a customizationId parameter.

3.5.1

Version 3.5.1

2016-11-01

• Fix: Remove limitation when creating a Solr cluster

3.5.0

Change Log

Version 3.5.0

2016-10-19_20

• New: Personality Insights v3
• New: Speech to Text Customizations
• New: Conversation version date 2016-09-19
• Fix: Calling synthesize in Text to Speech now accepts a rate for PCM. #454
• Fix: READMEs now point to the right maven dependency #459
• Fix: AlchemyLanguage now has anchor dates for combine calls
• Fix: Retrieve and Rank topAnswer bug #456

3.4.0

Change Log

Version 3.4.0

2016-09-29

The SDK 3.4.0 is the first step into moving from one maven module into a multi-module project.
This change will allow users to specify the service they want to use skipping the other services.
For example, in order to use Retrieve and Rank in Gradle you need:

 'com.ibm.watson.developer_cloud:retrieve-and-rank:3.4.0'

• New: Move to a maven multi-module project
• Fix: Send default headers in WebSocket connections #422
• Fix: Skip token creation if skipAuthentication=true #440

3.3.1

Version 3.3.1

2016-09-07

• New: AlchemyLanguage now returns the sourceText in the combined call.
• New: Add french broadband model to Speech to Text
• Fix: Add extracted dates to Alchemy Language combined call #436
• Fix: Speech to text example closes the connection.
• Fix: The extracted dates to Alchemy Language combined call #436
• Fix: removed deprecated services: Concept Insights and Relationship Extraction.
• Fix: Add missing parameter in document conversation. #428
• Fix: Add smart_formatting to Speech to Text

3.3.0

Change Log

Version 3.3.0

2016-08-08

• New: Tests are ignored if config.properties doesn't have valid credentials
• New: Document Conversion can index documents directly into Retrieve and Rank
• New: Visual Recognition Re-training
• New: Conversation v1: MessageRequest now supports entities and intents
• New: AlchemyLanguage typedRelations entities and intents
• New: Speech to Text keyword_spotting
• Fix: Remove Relationship extraction v1 Beta
• Fix: Remove Tone Analyzer v3 Beta
• Fix: 🔥 update version_date for Visual Recognition to be 2016-05-20
• Fix: Remove warning when JDNI can't find credentials

v3.2.0

CHANGES

• New: Conversation v1
• New: Text To Speech now supports audio/basic, audio/i16 and ogg/vorgis
• New: AlchemyLanguage now has a TypedRelation method that extract relationships and entities from text
• Fix: Speech to Text now returns the voice description when calling getVoices()
• Fix: url from VCAP_SERVICES take presecent over the default url
• Fix: Removed Relationship Extraction v1-beta
• Fix: Removed Tone Analyzer v3-beta

🔆🔆🔆
🔆🔆🔆

v3.0.1

The SDK 3.0.1 introduced support for react (use rx()). Services can now be combined using react in a synchronous or asynchronous calls and calls more efficiently.

CHANGES

• New: The User-Agent can be customized
• New: Services added:
  • Visual Recognition v3
  • Conversation v1-experimental
  • Tone Analyzer v3
• New: profanity_filter added to Speech to Text
• New: Introduce reactive API calls (use rx())
• New: Added new voices to Speech to Text
• New: getVoice() in Speech to Text and Text to Speech
• New: keyword_spotting in Speech to Text
• New: AlchemyLanguage now provides typedRelations() from Relationship Extraction
• Fix: Error when running in Android and using JDNI
• Fix: Error when loading ContentItems from a file in Personality Insights
• Fix: Conceptual search in Concept Insights now return the user_fields
• Fix: Content-Type in Speech to text when using flac files
• Fix: Warning when calling Alchemy services regarding connections being leaked
• Fix: Added missing fields in AlchemyDataNews
• Fix: SSL certificate issues when running in Bluemix and using the IBM JDK

There were also important changes on how we test the services and lots of bug fixing. Here is the list with all the issues we close in this release
🔆🔆🔆
🔆🔆🔆

CONTRIBUTORS

This release was driven mainly by @max-vogler with some help from @GrapeBaBa. 🚀