1 | # Changelog
|
2 |
|
3 | ### v0.8.3
|
4 |
|
5 | - Add `media` field onto the character message type.
|
6 |
|
7 | ### v0.8.2
|
8 |
|
9 | - Add `tapToContinue` to message history type.
|
10 |
|
11 | ### v0.8.1
|
12 |
|
13 | - Add `timestamp` to messages returned from `getMessageHistory`.
|
14 | - Improved type for `getMessageHistory`.
|
15 |
|
16 | ### v0.8.0
|
17 |
|
18 | - Can now specify `playthroughToken` to re-use a playthrough instead of creating a new one when connecting.
|
19 | - Can now fetch message history of the playthrough using `charisma.getMessageHistory`.
|
20 |
|
21 | ### v0.7.3
|
22 |
|
23 | - Fix `IMessageCharacter.speech` type.
|
24 |
|
25 | ### v0.7.2
|
26 |
|
27 | - `ISynthesisConfig` and some additional types are now exported.
|
28 |
|
29 | ### v0.7.1
|
30 |
|
31 | - `speech` now takes a config object as well as a boolean. It can specify the audio encoding to use and whether it returns the raw audio data, or a link to an audio file.
|
32 |
|
33 | ### v0.7.0
|
34 |
|
35 | - BREAKING: The `reply` event has been renamed to `message`, and now has a `type` field on the payload to distinguish between `character` and `media` events. Other fields have been refactored, such as `character` and `speech`. Please consult [src/types.ts](src/types.ts) to find the new message format.
|
36 | - A new `tap` event is available for the client to send.
|
37 |
|
38 | ### v0.6.0
|
39 |
|
40 | - Accidentally published version, but never tagged as `latest`.
|
41 |
|
42 | ### v0.5.1
|
43 |
|
44 | - Fix broken 0.5.0 publish.
|
45 |
|
46 | ### v0.5.0
|
47 |
|
48 | - Removed `browser` field from `package.json`. Consumers can use the UMD directly from unpkg.
|
49 | - Removed `actIndex` as it is no longer supported.
|
50 |
|
51 | ### v0.4.2
|
52 |
|
53 | - Buffer `set-memory` events until `status: 'ready'` is received.
|
54 |
|
55 | ### v0.4.1
|
56 |
|
57 | - `actIndex` and `sceneIndex` can now be set on the `start` event to start from a specific story scene.
|
58 |
|
59 | ### v0.4.0
|
60 |
|
61 | - **BREAKING**: UMD name changed from `Charisma` to `CharismaSDK`. The ES/CJS builds now don't bundle their dependencies.
|
62 | - Added `setMemory` method to directly set a memory.
|
63 | - Fixed all ID types to be `number`, not `string`.
|
64 |
|
65 | ### v0.3.1
|
66 |
|
67 | - Passing no `version` to the `connect` method now results in using the latest published version, rather than the draft version.
|
68 |
|
69 | ### v0.3.0
|
70 |
|
71 | - Package renamed (rescoped) to `@charisma-ai/sdk`.
|
72 |
|
73 | ### v0.2.0
|
74 |
|
75 | - The `debug` option has been replaced with the `version` option, which defaults to `undefined` (the latest published story version).
|
76 |
|
77 | ### v0.1.2
|
78 |
|
79 | - The microphone now stops listening when a reply with `endStory` set to `true` is emitted.
|
80 |
|
81 | ### v0.1.1
|
82 |
|
83 | - `AudioContext` is now created on-demand rather than on initialisation.
|
84 |
|
85 | ### v0.1.0
|
86 |
|
87 | - Socket.io now forces websockets, skipping the long-polling check.
|
88 |
|
89 | ### v0.0.4
|
90 |
|
91 | - Fixed issue where audio was not working on Safari.
|
92 |
|
93 | ### v0.0.3
|
94 |
|
95 | - Microphone keeps better track of whether to resume speech recognition after speaking.
|
96 |
|
97 | ### v0.0.2
|
98 |
|
99 | - Support for recording speech-to-text via Chrome SpeechRecognition.
|
100 | - New events `recognise-interim` and `recognise`.
|
101 | - Speech recognition is paused while the audio is played.
|
102 |
|
103 | ### v0.0.1
|
104 |
|
105 | - Initial release.
|
106 | - Support for `reply` and `start` client events, and `reply`, `start-typing` and `stop-typing` server events.
|
107 | - Support for playing text-to-speech audio.
|