Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog and this project adheres to Semantic Versioning.

Unreleased

[v8.0.6] - 2022-10-07

Added

websocket
- Added possibility to override recorder

v8.0.5 - 2022-03-15

Fixed

websocket
- Fixed crash when calling cleanup with an undefined socket

v8.0.4 - 2022-02-09

Added

websocket
- Optional text index start parameter for prompts

v8.0.3 - 2021-10-13

Fixed

websocket
- Cleanup now removes listeners
recorder
- Trying to initialize twice will no longer throw an error

v8.0.2 - 2021-08-26

Fixed

websocket
- Don't reconnect on connect failure

v8.0.1 - 2021-05-13

Fixed

recorder
- Firefox and Safari browser will now record with the correct channel count

v8.0.0 - 2021-04-23

Changed

recorder
- Changed the recorder from audio-recorder-polyfill which uses only ScriptProcessorNode to extended-media-recorder which can use AudioWorkers.
websocket
- Depracted cleanup
- start now return a promise which is fulfilled when the backend actually is able to receive audio.

v7.0.0 - 2021-03-04

Changed

websocket
- its-api websocket recording endpoint

v6.0.0 - 2021-01-11

Added

websocket
- its-api websocket helper functions.

Removed

api
- Removed its-gae API

v5.7.0 - 2020-10-12

Changed

sdk
- Switch from travis-ci to github actions.

Fixed

api
- Make sure a link header will be passed down the line.

v5.6.1 - 2020-07-06

Changed

api
- Improved the connection and more specifically the disconnection. It will show an error if it was not closed normally. Previously the reason for closure was swallowed and not visible for the user.

v5.6.0 - 2020-04-24

Changed

recorder
- Add constraints to the call of createMediaStream. This way we can be sure a recording session will be done with the minimum recording quality (sampleRate) that the ITSLanguage backend needs to perform analysis on.

v5.5.1 - 2020-02-28

Fixed

api
- Version number in the main index.js file updated.
- Exposed websocket connection which apparently was hidden.

v5.5.0 - 2020-02-28

Fixed

api
- When stopping the recorder if no audio was send, make sure to clean up and close the websocket connection.

Changed

api
- Do not auto reconnect the websocket connection by default.
- Add closing of the WebSocket connection to the finally handler in stead of doing it in both the then and catch.
- Check the size of the audio blob to send. If it seems to only have a header it wont sent it (true for wave). Any other mimeType will check for size > 0.

v5.4.1 - 2020-02-11

Changed

api
- Close the WebSocket connection after using it with feedback.

v5.4.0 - 2020-02-04

Changed

api, sdk
- Update the travis configuration.
- Make sure to allow overriding the dataEvent to listen on a recorder.

v5.3.0 - 2020-01-10

Fixed

api
- Only unregister an autobahn registration if there is actually an registration registered.
- Reset the lastChunk after it is used.
- Really cleanup some event listeners on the recorder after they have been used for streaming audio to the backend.
- TypeError raised by makeWebsocketCall.
- Allow to pass groups as a string (next to an array) and disallow other types. This is for the getById function of the Progress API.

Changed

api, recorder, player, examples, sdk
- Bumped used package versions.
- Upgraded yarn.lock file.

v5.2.0 - 2019-12-19

Added

recorder
- Added a new plugin: BufferPlugin. This plugin buffers the microphone input for a given number of seconds. It also exposes an api to get (part of) the buffer.
examples
- Added the react-recorder-buffer example. This example shows an example of the recorder with buffer plugin enabled.

Changed

recorder
- Updated the recorder readme file to be more accurate.

Fixed

api
- Fix the stability of the test by taking the asynchronous approach that was used into account.

v5.1.0 - 2019-09-27

Added

api
- Better inform the user when the recorder is prepared and ready to process audio. We now send out an recorderready event if ready.
examples
- Added the react-speechchallenge example. This example shows a speech challenge example and uses the recorder to record audio for it. It is also possible to listen back to the recorded audio.

Changed

api
- The dataavailable and stop event that will be used by the sdk to send data to the backend will now remove the event handlers upon adding them. Just the handlers that it would add. This to make sure the recorder can be reused in stead of just once.
- Before registering an RPC to the autobahn session, all the current registered functions will be unregistered, if there are any. This to make sure only one registration is active at the time.
- The broadcaster event websocketserverreadyforaudio is a bit confusing. Sending it through the event emitter is deprecated and will be removed in a future version.
sdk
- Updated .gitignore file with more common patterns.

v5.0.1 - 2019-08-29

Fixed

player
- Bumped the peer dependency of player to the correct api version.

v5.0.0 - 2019-08-29

Changed

api, recorder, player, examples, sdk
- Bumped used package versions.
- Upgraded yarn.lock file.
- Set core-js@3 as peerDependency.
recorder
- Allow the default browser media types for recording, not just audio/wav. This can be done through the mimeType param on the createRecorder function.

Fixed

recorder
- Failing test after bumping versions. Fixed by using non hardcoded value for the recorder getSpecs call.

Added

examples
- Added the react-recorder example. This example shows how the recorder can be used.

Removed

recorder
- The addAsGlobal function is removed.
- The signature for createRecorder is changed. The params setToWindow and asObject are removed.
- Several exported defaults are removed.

v4.2.2 - 2019-08-13

Fixed

recorder
- The recorder.getAudioSpecs() was returning some hard-coded values that where not true on some devices. If the recorded and processed audio sounded like it was speedup: this was why.

v4.2.1 - 2019-06-21

Changed

api
- Allow multiple writes for speech recordings by using the timeslice feature of the recorder.

v4.2.0 - 2019-04-17

Added

api, recorder, player, examples, sdk
- Added prettier to the project.
recorder
- Made it possible to load plugins.
- Added the Amplitude plugin.
examples
- Add an example project to demonstrate the AmplitudePlugin.

Changed

api, recorder, player
- Bump dependencies to the latest available versions.
- Update babel config to make use of the latest babel features.
- Updated all changelog files to point at this specific file. This way all changes can be logged and be visible at one location.

v4.1.0 - 2019-03-27

Added

api
- Added support for creating speech recordings via the SDK.

v4.0.2 - 2019-02-11

Fixed

api
- Removed adding of the Content-Type header when not needed to add it.

v4.0.1 - 2018-12-17

Packages release

Added

api
- Function to take Choice Recognition without streaming.

v4.0.0 - 2018-12-12

Changed

api, recorder, player
- Documentation is now build with JSDoc in stead of esdoc.
- SDK package will not be published anymore.

Removed

api, recorder, player
- SDK as single package
- Stopwatch
- VolumeMeter

Fixed

api, recorder, player
- call to this.createAudioContext on AudioContext class needed to call a static member
- WebAudio recorder can be stopped from paused state.

Added

api, recorder, player
- api package: available through @itslanguage/api.
- player package: available through @itslanguage/player.
- recorder package: available through @itslanguage/recorder.
- Packages are published as UMD build to unpkg.com.
- Feedback API, streaming audio.

v3.1.2

Changed

Version bumps for some used packages
Update license file with the correct date

v3.1.1

Added

Delay for the recorder, start recording after 100ms.

v3.1.0

Added

New functions to do proper impersonation, does not require credentials.

Fixed

Run travis tests with sudo enabled.

Changed

Run travis tests on headless chrome.
Slack notifications on travis builds.
Improve README.md documentation.
Changed the getUserAuth and getOAuth2Token to use the new API auth functions.