- Change: Update OkHttp to v3.6.0
2016-12-15
- New: Added Discovery v1
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()
andgetVoice
now accept acustomizationId
parameter.
2016-11-01
- Fix: Remove limitation when creating a Solr cluster
2016-10-19
- 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 arate
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
2016-09-30
- New: Add Language Translator
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
2016-09-07
- New: AlchemyLanguage now returns the
sourceText
in thecombined
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 close the connection when finish 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
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 be2016-05-20
- Fix: Remove warning when JDNI can't find credentials
2016-08-02
- New: Conversation v1
- New: Text To Speech now supports
audio/basic
,audio/i16
andogg/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 precedence over the default url - Fix: Removed Relationship Extraction v1-beta
- Fix: Removed Tone Analyzer v3-beta
2016-06-08
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.
- 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 usingflac
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
2016-04-22
The SDK 3.0 is designed around a new API that offers asynchronous callbacks in addition to synchronous blocking calls. It uses modern Java patterns like immutability and chained builders.
- New: Deprecated methods were removed
- Fix: Websockets issues during a recognition using Speech to Text
- New: Automatic snapshot deployments to Sonatype
- New: Chained builders for Speech to text and Tradeoff Analytics
2016-04-18
- Fix: Switch from String to Integer in AlchemyVision face recognition
- New: Enumerations for Languages in Language Translation
- Fix: Error in AlchemyLanguage when language wasn't specified