- Visual Recognition instances created post May 22 2018, have a new authentication mechanism
- Added opt-out option for collection parsing of strings in Natural Language Classifier Node.
- Fix to collection check in Natural Language Classification Node allowing for . in domain names.
- Fix to defaulting name for NLU Node.
- Allow pre-check of audio format to be disabled in Speech to Text node.
- Migrate from deprecated getModels and getCustomizations methods in Speech to Text nodes.
- Implement update classifier in Visual Recognition Util node.
- In Visual Recognition Util node append 'positive_' to zip name if neither 'positive' not 'negative' not already there.
- Removed duplicated code in Visual Recognition Util node.
- Needed to stringify json before addDocument in Discovery node.
- Using Node.js v 6 features, so will not run on Node.js v 4 anymore
- Implemented classify collection on Natural Language Classifier node. The collection can be in the form of multiple sentences, an array of strings, or an array of objects.
- Move all Discovery calls to latest API version - 2017-11-07
- Updated calls to deprecated discovery methods addJsonDocument and getCollections
- Correct implemetation of passages related options
- Allow highlight option to be specified in Discovery overrides - msg.discoveryparams
- Rename Conversation Nodes to Assistant
- Use Assistant endpoint
- Move all Assistant calls to latest API version - 2018-02-16
- Move all Visual Recognition calls to lates API version - 2018-03-19
- Add French as a Visual Recognition classification response language
- Enable Opt-out option for Conversation Node.
- Implement time out option for response from Conversation Node.
- Bump to v 3.2.1 of watson-developer-cloud dependency.
- Fix to how API version is specified in Natural Language Understanding Node.
- Add Korean to list of languages in Natural Language Classifier Node.
- Added Mute option for STT Node warning status when running in Streaming mode
- Discard option for STT Node Streaming listening events.
- Added Auto-Connect mode for STT Node - Will keep connection to service alive, else connection is restored on demand.
- Fix to Discovery Query Builder Node to return schema options.
- Endpoint can now be specified in Natural Language Classifier Node
- Speech to Text node now reports all errors, including disconnects when running in stream mode.
- Allow input / output from Node-RED web-sockets for Speech to Text node. To enable select streaming mode. No token is needed as the Node handles this. Look out for sample flows and templates showing how to use this feature.
- Three new languages for Language Translator, Polish, Russian and Turkish in Neural Translation mode.
- Removed the code for the deprecated Language Translation Nodes.
- Visual Recognition fix for accept-language
- Allow STT Language to be dyamically configurable using msg.srclang
- Bump to watson-developer-cloud 3.0.4
- Document link updates
- IE fix
- Option to set Conversation credentials though msg.params
- Speech To Text fullresult now includes speaker_labels. Breaking change to any code that used fullresult due an extra level of indirection.
- Allow Korean as an option for Personality Insights
- Text Recognition option removed from Visual Recognition Node
- Move Retrieve and Rank nodes to deprecated list
- Removed Document Conversion, Alchemy, Concept Insights, Dialog, Tradeoff Analytics and Similarity Search Nodes
- Nodes deprecated in 0.5.x will be removed in 0.6.x releases
- Allow file name to be configured in Add Document node.
- Allow payload as json for Add Document node.
- New node for Add Document for Discovery.
- Natural Language Understanding Node fix for credentials check.
- Added experimental neural translation as a config option to translation node.
- Update Supported audio file check in Speech to Text Node.
- Remove Continuous flag from Speech to Text Node.
- Add Alternatives and Smart Formatting options to Speech to Text Node.
- Allow Conversation Workspace Manager node to be dynamically configured.
- CRUD methods for Entity Values and Dialog Nodes in Conversation Workspace Manager node.
- Implement methods to manage Entities in Conversation workspace Manager node.
- Create Collection Language fix in Discovery Node.
- Similarity Search Nodes moved into deprecated category.
- Allow language code to be specified on create collection in Discovery Node.
- Implement delete collection, delete environment in Discovery Node.
- Allow multiple interface versions for Tone Analyzer Node.
- Enable Content, en or fr, in Tone Analyzer Node.
- Bump to latest version of watson-developer-cloud node.js sdk
- Allow empty input into converse for Conversation Node
- Endpoint can now be specified in Natural Language Understanding, Discovery and Discover Query Builder Nodes
- Full Promises implementation for on input processing for Natural Language Understanding Node
- Fix to node.error invocation in Conversation node.
- Personality Insights on Bluemix needed new path to node_modules
- Fix to Personality Insights Node when running in Japanese mode.
- Bump Interface version to Discovery service to '2017-08-01'.
- Japanese supported by Translation service.
- Speech to Text service now reports UK English as 'en-GB'.
- Add button to allow model cache for STT node to be flushed.
- Added more of the supported audio formats for the TTS node.
- Retrieve and Rank node update to return document body.
- Add support for NLP Query in Discovery and Discovery Query Builder Nodes.
- Allowed detect_mode for Visual Recognition node to be set in msg.params
- Text to Speech speech on msg.payload option.
- Speech to Text transcription on msg.payload option
- Endpoint can now be specified in Conversation, Conversation Workspace Manager Language Identify, Language Translator, Speech to Text, STT Corpus Builder, Text to Speech, TTS Corpus Builder, Personality Insights and Tone Analyzer nodes.
- Visual Reconition Node, now accepts readstream on msg.payload
- Add passages parameter to Discovery Node
- Fix to Tone Analyzer to allow JSON as input
- Enabled Conversation Tone method to Tone Analyzer Node
- Discovery Node now supports: Create Environment, Create Configuration, Create Collection
- User Agent added to following nodes: Conversation, Conversation Workspace Manager, Discovery, Discovery Query Builder, Document Conversion, Language Translator, Language Translator Utility, Language Indentify, Natural Language Classifier, Natural Language Understanding, Personality Insights, Retrieve and Rank, Similarity Search, Speech to Text, STT Corpus Builder, Text to Speech, TTS Corpus Builder, Tone Analyzer, Visual Recognition
- Discovery Query Builder was not picking up searchable fields.
- Personality Insights Node reset to use '2016-10-20' version
- Natural Language Classifier Node migrated to use url based services utility to detect bound service.
- Natural Language Classifier Node Migrated to use Promises.
- Natural Language Classifier now able to use File Inject as input for create mode.
- Natural Language Classifier Node allows Classifier ID to be set dynamically in classify, list and remove modes.
- Natural Language Understanding Node now allows language and model overrides on msg.nlu_options.language, msg.nlu_options.entity_model and msg.nlu_options.relations_model.
- Trade-off Analytics Node moved to the deprecated list.
- Migrated Visual Recognition and Visual Recognition Util nodes to use promises, and ensure that all error responses are reported.
- Fix for Service / Name conflicts in Document Conversion Node
- Implement methods to manage Counter Examples in Conversation workspace Manager node.
- Removed V1 Personality Insights node.
- Removed V1 Language Indentification node.
- Removed V1 Language Translation node.
- Removed V1 Language Translation Utility node.
- Visual Recognition was overwriting msg.payload with 'look at msg.results'. Fixed so that msg.payload is left as is.
- Implement methods to manage for Intent and Example Input for Intent, in Conversation workspace Manager node.
- Deprecated Alchemy Nodes.
- Removed Conversation experimental node.
- Removed Discovery experimental node.
- Removed Tone Analyzer beta node.
- Removed Relationship Extraction node.
- Removed the V1 Visual Recognition node.
- New node for Natural Language Understanding
- watson-developer-cloud dependency forces node engine >= 4.5
- Nodes deprecated in 0.4.x will be removed in 0.5.x releases
- Bump version of Conversation interface to '2017-02-03'
- New node for Conversation Workspace Management
- Clean up fix in Speech to Text Node, which caused problems with long audio tracks.
- Add get pronunciation to Text to Speech Customization Node.
- Add delete customization to Speech to Text Customization Node.
- Add description to Info panel for Text to Speech Customizations Node.
- Tone node updated to use url based services utility to detect bound service.
- New Node for Text to Speech Customizations.
- Added option to select Customization in Text To Speech Node.
- Added support for word training in Speech to Text Customizations Node.
- New Node for Speech to Text Customizations - Initial support for corpus (sentences) training.
- Added option to select Customization in Speech to Text Node.
- Fixed bug in Conversation node, when no msg.params defined.
- Added Keyword and Entity emotion and sentiment as options on Alchemy Feature Extract Node.
- Allow Conversation credentials to be provided on msg.params
- Similarity Search and Visual Recognition Nodes updated to use url based services to detect bound service.
- Personality Insights node updated to latest December 15, 2016 GloVe profiles.
- Conversation, Alchemy Feature Extract and Date Extraction nodes updated to use url based services utility to detect bound service.
- Added diarization support to STT Node via the parameter speaker_labels
- STT and TTS node use url based services utility to detect bound service.
- New V1 Query Builder Node for the Discovery Node
- New services utilities to handle node-red app name clash problems
- New V1 Discovery Node
- Implementation of Discovery Node query, list and get details methods
- Fix to Translator Node which was ignoring any override values set in the msg object
- Fix to Document Conversion Node to preserve full msg object
- Fix to Language Translator Node train mode documentation
- New V1 Experimental Discovery Node
- Change color of V3 Personality Insights Node
- New V3 Personality Insights Node
- Deprecated V2 Personality Insights Node
- Fix to Language Translation Node to allow it work with File Inject
- Bugfixes to Language Translator Train Node
- Fix to Language Translator Train Node to allow it work with File Inject
- Added new Date Extraction Node
- Visual Recognition Node, now defaults detection setting and language on initial initialisation.
- Added Translation and Language Identify nodes for Language Translator Service (Blue Tiles)
- Deprecated Translation and Language Indentification nodes for Language Translation (Green Tiles)
- Translation Node supports German
- New Similarity Search (Beta) Node
- New Language Translation Util mode, to request available translation models
- Translation Node now has option to look for global translation settings
- Enhancements to Conversation Node
- Node has option to remember context
- Context support for single user and multi-user models
- Ability to reset context, including user specific reset
- Ability to splice in additional context variables.
- Added Concept support to News Node
- Name space fixes to Text to Speech Node
- Fixed how Document Conversion node was handling docx files
- Oops! fix to Speech to Text Node
- Name space fixes to Speech to Text Node
- The dialog for the Text to Speech service now loads available voices dynamically. This allows new voices and languages to be identified without requiring a further code change.
- Emergency fix for Watson Language Translation node. Bluemix credentials read too late.
- Emergency fix for node.js version compatibility problem, in payload-utils
- Corrected word count for Japanese text in Personality Insights nodes.
- Move Dialog to deprecated list.
- Conversation : added support for optional alternate_intents, output, entities and intents.
- Conversation : Bugfixes
- Retrieve and Rank : allow parameters to be passed into the nodes on the msg object
- Moved Relationship Extraction and Concept Insights nodes to the deprecated list.
- Updated dependency to watson-developer-cloud node-SDK v2.1.0
- Added in German and Japanese support to Natural Language Classifier node
- Visual Recognition V3 : added support of Accept-Language for Classify feature, new icon with pink background, icon label renamed (removing v3 from node name)
- Fixed document conversion node when filetype is not recognized
- Moved functions to utility module
- Oops! fix to Conversation V1 node
- New Conversation V1 node
- Bugfixes to Feature Extract node
- Bugfixes to Visual Recognition node
- Added in Japanese and Arabic support to Personality Insights node
- Updated dependency to watson-developer-cloud node-SDK 1.12.2
- New Conversation experimental node.
- Visual Recognition v3 : added a Delete All Classifiers feature in the util node. Added corresponding Flow in Starter.
- New palette category for deprecated Nodes, that are being retained for backward compatibity.
- Bugfixes to Speech to Text.
- New Visual Recognition V3 node to support the V3 GA API. This incorporates the features that were previously available as part of AlchemyAPI Vision.
- The transcription returned for the Speech to Text node now returns full (untruncated) transcription. The alternatives are returned in msg.fullresult. The dialog now loads available models dynamically. This will allow new speech models to be identified without requiring a further code change.
- The Retrieve and Rank node nows stores credentials in a configuration node, allowing the credentials to be shared acrosss a flow with multiple Retrieve and Rank nodes.
- New Tone Analyzer V3 node to support the V3 GA API.
- AlchemyAPI Image Analysis and Language nodes migrated from old Alchemy SDK to current Watson Developer Cloud SDK for Node.js
- The dialog for the Language Translation service now loads available models dynamically. This allows new translation models to be identified without requiring a further code change. This version allows customised translations models to be selected.