@speech-sdk/core 0.7.0 → 0.8.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (245) hide show
  1. package/README.md +227 -108
  2. package/dist/__tests__/e2e/_save-audio.d.ts +0 -42
  3. package/dist/__tests__/e2e/_save-audio.d.ts.map +1 -1
  4. package/dist/__tests__/e2e/_save-audio.js +0 -59
  5. package/dist/__tests__/e2e/_save-audio.js.map +1 -1
  6. package/dist/audio-decode.d.ts +7 -0
  7. package/dist/audio-decode.d.ts.map +1 -0
  8. package/dist/audio-decode.js +109 -0
  9. package/dist/audio-decode.js.map +1 -0
  10. package/dist/audio-duration.d.ts +0 -5
  11. package/dist/audio-duration.d.ts.map +1 -1
  12. package/dist/audio-duration.js +5 -21
  13. package/dist/audio-duration.js.map +1 -1
  14. package/dist/audio-output.d.ts +39 -0
  15. package/dist/audio-output.d.ts.map +1 -0
  16. package/dist/audio-output.js +111 -0
  17. package/dist/audio-output.js.map +1 -0
  18. package/dist/audio-utils.d.ts +2 -10
  19. package/dist/audio-utils.d.ts.map +1 -1
  20. package/dist/audio-utils.js +57 -15
  21. package/dist/audio-utils.js.map +1 -1
  22. package/dist/captions.d.ts +0 -108
  23. package/dist/captions.d.ts.map +1 -1
  24. package/dist/captions.js +8 -98
  25. package/dist/captions.js.map +1 -1
  26. package/dist/conversation/attribute-timestamps.d.ts +26 -0
  27. package/dist/conversation/attribute-timestamps.d.ts.map +1 -0
  28. package/dist/conversation/attribute-timestamps.js +276 -0
  29. package/dist/conversation/attribute-timestamps.js.map +1 -0
  30. package/dist/conversation/dispatch.d.ts +5 -5
  31. package/dist/conversation/dispatch.d.ts.map +1 -1
  32. package/dist/conversation/dispatch.js +18 -8
  33. package/dist/conversation/dispatch.js.map +1 -1
  34. package/dist/conversation/errors.d.ts +3 -0
  35. package/dist/conversation/errors.d.ts.map +1 -1
  36. package/dist/conversation/errors.js +6 -0
  37. package/dist/conversation/errors.js.map +1 -1
  38. package/dist/conversation/pcm-concat.d.ts +0 -24
  39. package/dist/conversation/pcm-concat.d.ts.map +1 -1
  40. package/dist/conversation/pcm-concat.js +8 -183
  41. package/dist/conversation/pcm-concat.js.map +1 -1
  42. package/dist/conversation/proportional-fill.d.ts +10 -0
  43. package/dist/conversation/proportional-fill.d.ts.map +1 -0
  44. package/dist/conversation/proportional-fill.js +64 -0
  45. package/dist/conversation/proportional-fill.js.map +1 -0
  46. package/dist/conversation/silence-detection.d.ts +14 -0
  47. package/dist/conversation/silence-detection.d.ts.map +1 -0
  48. package/dist/conversation/silence-detection.js +52 -0
  49. package/dist/conversation/silence-detection.js.map +1 -0
  50. package/dist/conversation/stitch.d.ts +9 -6
  51. package/dist/conversation/stitch.d.ts.map +1 -1
  52. package/dist/conversation/stitch.js +72 -51
  53. package/dist/conversation/stitch.js.map +1 -1
  54. package/dist/conversation/types.d.ts +7 -37
  55. package/dist/conversation/types.d.ts.map +1 -1
  56. package/dist/conversation/validate.d.ts +1 -16
  57. package/dist/conversation/validate.d.ts.map +1 -1
  58. package/dist/conversation/validate.js +29 -29
  59. package/dist/conversation/validate.js.map +1 -1
  60. package/dist/default-stt-fallback.d.ts +3 -0
  61. package/dist/default-stt-fallback.d.ts.map +1 -0
  62. package/dist/default-stt-fallback.js +11 -0
  63. package/dist/default-stt-fallback.js.map +1 -0
  64. package/dist/derive-timestamps.d.ts +1 -5
  65. package/dist/derive-timestamps.d.ts.map +1 -1
  66. package/dist/derive-timestamps.js +1 -15
  67. package/dist/derive-timestamps.js.map +1 -1
  68. package/dist/encoders/mp3.d.ts +6 -0
  69. package/dist/encoders/mp3.d.ts.map +1 -0
  70. package/dist/encoders/mp3.js +54 -0
  71. package/dist/encoders/mp3.js.map +1 -0
  72. package/dist/errors.d.ts +20 -13
  73. package/dist/errors.d.ts.map +1 -1
  74. package/dist/errors.js +49 -15
  75. package/dist/errors.js.map +1 -1
  76. package/dist/generate-conversation.d.ts +5 -4
  77. package/dist/generate-conversation.d.ts.map +1 -1
  78. package/dist/generate-conversation.js +250 -93
  79. package/dist/generate-conversation.js.map +1 -1
  80. package/dist/generate-speech.d.ts +7 -28
  81. package/dist/generate-speech.d.ts.map +1 -1
  82. package/dist/generate-speech.js +185 -94
  83. package/dist/generate-speech.js.map +1 -1
  84. package/dist/index.d.ts +7 -11
  85. package/dist/index.d.ts.map +1 -1
  86. package/dist/index.js +6 -4
  87. package/dist/index.js.map +1 -1
  88. package/dist/logger.d.ts.map +1 -1
  89. package/dist/logger.js +2 -13
  90. package/dist/logger.js.map +1 -1
  91. package/dist/metadata.d.ts +0 -22
  92. package/dist/metadata.d.ts.map +1 -1
  93. package/dist/pronunciations/errors.d.ts +5 -0
  94. package/dist/pronunciations/errors.d.ts.map +1 -0
  95. package/dist/pronunciations/errors.js +8 -0
  96. package/dist/pronunciations/errors.js.map +1 -0
  97. package/dist/pronunciations/inverse-align.d.ts +4 -0
  98. package/dist/pronunciations/inverse-align.d.ts.map +1 -0
  99. package/dist/pronunciations/inverse-align.js +54 -0
  100. package/dist/pronunciations/inverse-align.js.map +1 -0
  101. package/dist/pronunciations/merge.d.ts +4 -0
  102. package/dist/pronunciations/merge.d.ts.map +1 -0
  103. package/dist/pronunciations/merge.js +13 -0
  104. package/dist/pronunciations/merge.js.map +1 -0
  105. package/dist/pronunciations/substitute.d.ts +6 -0
  106. package/dist/pronunciations/substitute.d.ts.map +1 -0
  107. package/dist/pronunciations/substitute.js +67 -0
  108. package/dist/pronunciations/substitute.js.map +1 -0
  109. package/dist/pronunciations/types.d.ts +18 -0
  110. package/dist/pronunciations/types.d.ts.map +1 -0
  111. package/dist/pronunciations/types.js +2 -0
  112. package/dist/pronunciations/types.js.map +1 -0
  113. package/dist/pronunciations/validate.d.ts +3 -0
  114. package/dist/pronunciations/validate.d.ts.map +1 -0
  115. package/dist/pronunciations/validate.js +26 -0
  116. package/dist/pronunciations/validate.js.map +1 -0
  117. package/dist/provider-utils.d.ts +4 -9
  118. package/dist/provider-utils.d.ts.map +1 -1
  119. package/dist/provider-utils.js +60 -51
  120. package/dist/provider-utils.js.map +1 -1
  121. package/dist/providers/cartesia/alignment.d.ts +0 -16
  122. package/dist/providers/cartesia/alignment.d.ts.map +1 -1
  123. package/dist/providers/cartesia/alignment.js +1 -6
  124. package/dist/providers/cartesia/alignment.js.map +1 -1
  125. package/dist/providers/cartesia/index.d.ts +29 -19
  126. package/dist/providers/cartesia/index.d.ts.map +1 -1
  127. package/dist/providers/cartesia/index.js +116 -80
  128. package/dist/providers/cartesia/index.js.map +1 -1
  129. package/dist/providers/deepgram/index.d.ts +23 -8
  130. package/dist/providers/deepgram/index.d.ts.map +1 -1
  131. package/dist/providers/deepgram/index.js +51 -18
  132. package/dist/providers/deepgram/index.js.map +1 -1
  133. package/dist/providers/elevenlabs/alignment.d.ts +7 -21
  134. package/dist/providers/elevenlabs/alignment.d.ts.map +1 -1
  135. package/dist/providers/elevenlabs/alignment.js +8 -9
  136. package/dist/providers/elevenlabs/alignment.js.map +1 -1
  137. package/dist/providers/elevenlabs/index.d.ts +14 -38
  138. package/dist/providers/elevenlabs/index.d.ts.map +1 -1
  139. package/dist/providers/elevenlabs/index.js +186 -169
  140. package/dist/providers/elevenlabs/index.js.map +1 -1
  141. package/dist/providers/fal/index.d.ts +11 -20
  142. package/dist/providers/fal/index.d.ts.map +1 -1
  143. package/dist/providers/fal/index.js +49 -37
  144. package/dist/providers/fal/index.js.map +1 -1
  145. package/dist/providers/fish-audio/index.d.ts +14 -8
  146. package/dist/providers/fish-audio/index.d.ts.map +1 -1
  147. package/dist/providers/fish-audio/index.js +47 -19
  148. package/dist/providers/fish-audio/index.js.map +1 -1
  149. package/dist/providers/gateway/index.d.ts +76 -0
  150. package/dist/providers/gateway/index.d.ts.map +1 -0
  151. package/dist/providers/gateway/index.js +251 -0
  152. package/dist/providers/gateway/index.js.map +1 -0
  153. package/dist/providers/google/index.d.ts +12 -20
  154. package/dist/providers/google/index.d.ts.map +1 -1
  155. package/dist/providers/google/index.js +180 -162
  156. package/dist/providers/google/index.js.map +1 -1
  157. package/dist/providers/hume/alignment.d.ts +30 -35
  158. package/dist/providers/hume/alignment.d.ts.map +1 -1
  159. package/dist/providers/hume/alignment.js +14 -8
  160. package/dist/providers/hume/alignment.js.map +1 -1
  161. package/dist/providers/hume/index.d.ts +16 -16
  162. package/dist/providers/hume/index.d.ts.map +1 -1
  163. package/dist/providers/hume/index.js +79 -65
  164. package/dist/providers/hume/index.js.map +1 -1
  165. package/dist/providers/inworld/alignment.d.ts +8 -22
  166. package/dist/providers/inworld/alignment.d.ts.map +1 -1
  167. package/dist/providers/inworld/alignment.js +9 -8
  168. package/dist/providers/inworld/alignment.js.map +1 -1
  169. package/dist/providers/inworld/index.d.ts +17 -20
  170. package/dist/providers/inworld/index.d.ts.map +1 -1
  171. package/dist/providers/inworld/index.js +79 -47
  172. package/dist/providers/inworld/index.js.map +1 -1
  173. package/dist/providers/mistral/index.d.ts +14 -8
  174. package/dist/providers/mistral/index.d.ts.map +1 -1
  175. package/dist/providers/mistral/index.js +63 -48
  176. package/dist/providers/mistral/index.js.map +1 -1
  177. package/dist/providers/murf/alignment.d.ts +10 -19
  178. package/dist/providers/murf/alignment.d.ts.map +1 -1
  179. package/dist/providers/murf/alignment.js +10 -5
  180. package/dist/providers/murf/alignment.js.map +1 -1
  181. package/dist/providers/murf/index.d.ts +15 -16
  182. package/dist/providers/murf/index.d.ts.map +1 -1
  183. package/dist/providers/murf/index.js +105 -58
  184. package/dist/providers/murf/index.js.map +1 -1
  185. package/dist/providers/openai/index.d.ts +43 -29
  186. package/dist/providers/openai/index.d.ts.map +1 -1
  187. package/dist/providers/openai/index.js +294 -106
  188. package/dist/providers/openai/index.js.map +1 -1
  189. package/dist/providers/resemble/alignment.d.ts +8 -29
  190. package/dist/providers/resemble/alignment.d.ts.map +1 -1
  191. package/dist/providers/resemble/alignment.js +9 -12
  192. package/dist/providers/resemble/alignment.js.map +1 -1
  193. package/dist/providers/resemble/index.d.ts +21 -11
  194. package/dist/providers/resemble/index.d.ts.map +1 -1
  195. package/dist/providers/resemble/index.js +89 -49
  196. package/dist/providers/resemble/index.js.map +1 -1
  197. package/dist/providers/smallest-ai/index.d.ts +47 -0
  198. package/dist/providers/smallest-ai/index.d.ts.map +1 -0
  199. package/dist/providers/smallest-ai/index.js +107 -0
  200. package/dist/providers/smallest-ai/index.js.map +1 -0
  201. package/dist/providers/xai/index.d.ts +25 -9
  202. package/dist/providers/xai/index.d.ts.map +1 -1
  203. package/dist/providers/xai/index.js +63 -40
  204. package/dist/providers/xai/index.js.map +1 -1
  205. package/dist/providers.d.ts +31 -0
  206. package/dist/providers.d.ts.map +1 -0
  207. package/dist/providers.js +16 -0
  208. package/dist/providers.js.map +1 -0
  209. package/dist/resolve-provider.d.ts.map +1 -1
  210. package/dist/resolve-provider.js +8 -51
  211. package/dist/resolve-provider.js.map +1 -1
  212. package/dist/retry-options.d.ts +6 -0
  213. package/dist/retry-options.d.ts.map +1 -0
  214. package/dist/retry-options.js +48 -0
  215. package/dist/retry-options.js.map +1 -0
  216. package/dist/speech-provider.d.ts +28 -53
  217. package/dist/speech-provider.d.ts.map +1 -1
  218. package/dist/speech-provider.js +5 -26
  219. package/dist/speech-provider.js.map +1 -1
  220. package/dist/speech-result.d.ts +8 -9
  221. package/dist/speech-result.d.ts.map +1 -1
  222. package/dist/speech-result.js.map +1 -1
  223. package/dist/speech-to-text-provider.d.ts +0 -12
  224. package/dist/speech-to-text-provider.d.ts.map +1 -1
  225. package/dist/stream-speech.d.ts +4 -2
  226. package/dist/stream-speech.d.ts.map +1 -1
  227. package/dist/stream-speech.js +36 -22
  228. package/dist/stream-speech.js.map +1 -1
  229. package/dist/timestamps.d.ts +3 -17
  230. package/dist/timestamps.d.ts.map +1 -1
  231. package/dist/turns.d.ts +9 -0
  232. package/dist/turns.d.ts.map +1 -0
  233. package/dist/turns.js +21 -0
  234. package/dist/turns.js.map +1 -0
  235. package/dist/types.d.ts +31 -0
  236. package/dist/types.d.ts.map +1 -1
  237. package/dist/volume-adjust.d.ts +0 -6
  238. package/dist/volume-adjust.d.ts.map +1 -1
  239. package/dist/volume-adjust.js +4 -16
  240. package/dist/volume-adjust.js.map +1 -1
  241. package/package.json +13 -66
  242. package/dist/stt-providers/openai/index.d.ts +0 -42
  243. package/dist/stt-providers/openai/index.d.ts.map +0 -1
  244. package/dist/stt-providers/openai/index.js +0 -184
  245. package/dist/stt-providers/openai/index.js.map +0 -1
@@ -1,122 +1,226 @@
1
1
  import pRetry from "p-retry";
2
2
  import { computeAudioDuration } from "./audio-duration.js";
3
+ import { applyOptionalOutputConversion, validateOutput, } from "./audio-output.js";
3
4
  import { detectAudioTags, stripAudioTags } from "./audio-tags.js";
5
+ import { getDefaultSTTFallback } from "./default-stt-fallback.js";
4
6
  import { deriveTimestampsViaSTT } from "./derive-timestamps.js";
5
- import { ApiError, NoSpeechGeneratedError, VolumeAdjustmentUnsupportedError, } from "./errors.js";
7
+ import { NoSpeechGeneratedError, OutputConversionUnsupportedError, VolumeAdjustmentUnsupportedError, } from "./errors.js";
6
8
  import { debug } from "./logger.js";
9
+ import { inverseAlign } from "./pronunciations/inverse-align.js";
10
+ import { mergeRules } from "./pronunciations/merge.js";
11
+ import { substitute } from "./pronunciations/substitute.js";
12
+ import { validatePronunciationsInput } from "./pronunciations/validate.js";
7
13
  import { resolveModel } from "./resolve-provider.js";
8
- import { modelDeclaresNativeTimestamps, } from "./speech-provider.js";
14
+ import { buildRetryOptions } from "./retry-options.js";
15
+ import { isSpeechGatewayModel, modelDeclaresNativeTimestamps, } from "./speech-provider.js";
9
16
  import { DefaultGeneratedAudioFile } from "./speech-result.js";
10
17
  export async function generateSpeech(options) {
11
- const { model, voice, abortSignal, headers, volumeDbfs, timestamps: timestampMode = "auto", timestampProvider, } = options;
18
+ const { model, voice, abortSignal, headers, volumeDbfs, timestamps = false, } = options;
12
19
  const maxRetries = options.maxRetries ?? 2;
20
+ validateOutput(options.output);
13
21
  const resolved = resolveModel(model, { apiKey: options.apiKey });
14
22
  const modelIdentifier = `${resolved.provider.id}/${resolved.modelId}`;
15
- let providerOptions = options.providerOptions;
16
- if (volumeDbfs != null) {
17
- const stitchOpts = resolved.provider.getStitchOptions?.(resolved.modelId);
18
- if (!stitchOpts) {
19
- throw new VolumeAdjustmentUnsupportedError(modelIdentifier);
20
- }
21
- // Stitch-mode options are applied last so they win over user-supplied
22
- // providerOptions — otherwise a caller could silently break the decoder
23
- // by e.g. passing `response_format: "mp3"` alongside `volumeDbfs`.
24
- providerOptions = {
25
- ...options.providerOptions,
26
- ...stitchOpts.providerOptions,
27
- };
28
- }
29
- const { text: processedText, warnings } = preprocessText(resolved, options.text, modelIdentifier);
30
- if (processedText.trim().length === 0) {
23
+ const isGateway = isSpeechGatewayModel(resolved);
24
+ validatePronunciationsInput(options.pronunciations, isGateway);
25
+ const { text: strippedText, warnings } = preprocessText(resolved, options.text, modelIdentifier);
26
+ if (strippedText.trim().length === 0) {
31
27
  throw new NoSpeechGeneratedError(warnings.length > 0
32
28
  ? `Text is empty after removing unsupported audio tags for ${modelIdentifier}.`
33
29
  : "Text must not be empty.");
34
30
  }
31
+ let textToSend = strippedText;
32
+ let pronunciationEdits = [];
33
+ if (!isGateway && options.pronunciations?.rules?.length) {
34
+ const ruleMap = mergeRules(options.pronunciations.rules);
35
+ const subbed = substitute(strippedText, ruleMap);
36
+ textToSend = subbed.text;
37
+ pronunciationEdits = subbed.edits;
38
+ }
39
+ const providerOptions = resolveProviderOptionsForLocalDecoding({
40
+ resolved,
41
+ isGateway,
42
+ modelIdentifier,
43
+ volumeDbfs,
44
+ output: options.output,
45
+ callerOptions: options.providerOptions,
46
+ });
35
47
  const hasNativeTimestamps = modelDeclaresNativeTimestamps(resolved);
36
- // For "on" we still ask the provider natively first — if it has native
37
- // alignment, we skip the STT round-trip.
38
- const shouldRequestNative = (timestampMode === "on" || timestampMode === "auto") && hasNativeTimestamps;
48
+ const shouldRequestNative = timestamps && (hasNativeTimestamps || isGateway);
49
+ const effectiveFallback = !timestamps || shouldRequestNative
50
+ ? undefined
51
+ : (resolved.fallbackSTT ?? (await getDefaultSTTFallback()));
39
52
  logTimestampDecision({
40
53
  modelIdentifier,
41
- mode: timestampMode,
54
+ enabled: timestamps,
42
55
  hasNative: hasNativeTimestamps,
43
56
  willRequestNative: shouldRequestNative,
44
- timestampProvider,
57
+ effectiveFallback,
45
58
  });
46
59
  const startTime = performance.now();
47
- const result = await pRetry(() => resolved.provider.generate({
48
- modelId: resolved.modelId,
49
- text: processedText,
50
- voice,
51
- providerOptions,
52
- abortSignal,
53
- headers,
54
- includeTimestamps: shouldRequestNative,
55
- }), {
56
- retries: maxRetries,
57
- signal: abortSignal,
58
- shouldRetry: ({ error }) => {
59
- if (error instanceof ApiError && error.statusCode < 500) {
60
- return false;
61
- }
62
- return true;
63
- },
64
- });
60
+ const result = await pRetry(() => {
61
+ if (isGateway) {
62
+ const gatewayProvider = resolved.provider;
63
+ return gatewayProvider.generate({
64
+ modelId: resolved.modelId,
65
+ text: textToSend,
66
+ voice: voice,
67
+ providerOptions,
68
+ abortSignal,
69
+ headers,
70
+ includeTimestamps: shouldRequestNative,
71
+ volumeDbfs,
72
+ output: options.output,
73
+ pronunciations: options.pronunciations,
74
+ });
75
+ }
76
+ return resolved.provider.generate({
77
+ modelId: resolved.modelId,
78
+ text: textToSend,
79
+ voice,
80
+ providerOptions,
81
+ abortSignal,
82
+ headers,
83
+ includeTimestamps: shouldRequestNative,
84
+ });
85
+ }, buildRetryOptions({ maxRetries, abortSignal }));
65
86
  const latencyMs = Math.round(performance.now() - startTime);
66
87
  const audioData = result.audio;
67
88
  if (audioData.length === 0) {
68
89
  throw new NoSpeechGeneratedError();
69
90
  }
70
- let outputBytes = audioData;
71
- let outputMediaType = result.mediaType;
72
- if (volumeDbfs != null) {
73
- const { adjustVolume } = await import("./volume-adjust.js");
74
- outputBytes = await adjustVolume({
91
+ const { bytes: outputBytes, mediaType: outputMediaType } = isGateway
92
+ ? { bytes: audioData, mediaType: result.mediaType }
93
+ : await applyLocalAudioPostProcessing({
75
94
  audio: audioData,
76
95
  mediaType: result.mediaType,
77
96
  volumeDbfs,
97
+ output: options.output,
78
98
  });
79
- outputMediaType = "audio/wav";
80
- }
81
99
  const audio = new DefaultGeneratedAudioFile({
82
100
  data: outputBytes,
83
101
  mediaType: outputMediaType,
84
102
  });
85
- const audioDurationMs = (await computeAudioDuration(audio.uint8Array, outputMediaType)) ??
86
- result.audioDurationMs;
87
- let timestamps;
88
- if (timestampMode !== "off") {
89
- if (result.timestamps && result.timestamps.length > 0) {
90
- debug(`${modelIdentifier}: returned ${result.timestamps.length} native word timestamps.`);
91
- timestamps = result.timestamps;
92
- }
93
- else if (timestampMode === "on") {
94
- timestamps = await deriveTimestampsViaSTT({
95
- ttsModel: modelIdentifier,
96
- audio: audio.uint8Array,
97
- mediaType: outputMediaType,
98
- timestampProvider,
99
- abortSignal,
100
- });
101
- debug(`${modelIdentifier}: derived ${timestamps.length} word timestamps via STT fallback.`);
102
- }
103
- }
103
+ const [computedDuration, resolvedTimestamps] = await Promise.all([
104
+ computeAudioDuration(audio.uint8Array, outputMediaType),
105
+ resolveTimestamps({
106
+ timestamps,
107
+ modelIdentifier,
108
+ resolved,
109
+ resultTimestamps: result.timestamps,
110
+ audio: audio.uint8Array,
111
+ mediaType: outputMediaType,
112
+ abortSignal,
113
+ }),
114
+ ]);
115
+ const audioDurationMs = computedDuration ?? result.audioDurationMs;
116
+ const finalTimestamps = resolvedTimestamps && pronunciationEdits.length > 0
117
+ ? inverseAlign(resolvedTimestamps, textToSend, pronunciationEdits)
118
+ : resolvedTimestamps;
104
119
  const metadata = {
105
120
  latencyMs,
106
- inputChars: processedText.length,
107
- provider: resolved.provider.id,
108
- model: resolved.modelId,
121
+ inputChars: options.text.length,
109
122
  ...(audioDurationMs != null && { audioDurationMs }),
110
123
  };
111
124
  return {
112
125
  audio,
113
126
  metadata,
114
127
  providerMetadata: result.providerMetadata,
115
- warnings: warnings.length > 0 ? warnings : undefined,
116
- timestamps,
128
+ warnings: mergeWarnings(warnings, result.warnings),
129
+ timestamps: finalTimestamps,
117
130
  };
118
131
  }
132
+ function mergeWarnings(preprocessingWarnings, providerWarnings) {
133
+ const merged = [...preprocessingWarnings, ...(providerWarnings ?? [])];
134
+ return merged.length > 0 ? merged : undefined;
135
+ }
136
+ async function resolveTimestamps(args) {
137
+ if (!args.timestamps) {
138
+ return;
139
+ }
140
+ if (args.resultTimestamps?.length) {
141
+ debug(`${args.modelIdentifier}: returned ${args.resultTimestamps.length} native word timestamps.`);
142
+ return args.resultTimestamps;
143
+ }
144
+ if (isSpeechGatewayModel(args.resolved)) {
145
+ return;
146
+ }
147
+ const fallback = args.resolved.fallbackSTT ?? (await getDefaultSTTFallback());
148
+ const timestamps = await deriveTimestampsViaSTT({
149
+ ttsModel: args.modelIdentifier,
150
+ audio: args.audio,
151
+ mediaType: args.mediaType,
152
+ timestampFallback: fallback,
153
+ abortSignal: args.abortSignal,
154
+ });
155
+ debug(`${args.modelIdentifier}: derived ${timestamps.length} word timestamps via STT fallback.`);
156
+ return timestamps;
157
+ }
158
+ function resolveProviderOptionsForLocalDecoding(args) {
159
+ if (args.isGateway) {
160
+ return args.callerOptions;
161
+ }
162
+ // volumeDbfs needs a known-decodable wire format to decode→re-level→re-encode,
163
+ // so SDK-required keys win over caller overrides — otherwise a stray
164
+ // override would silently violate the normalization contract.
165
+ if (args.volumeDbfs != null) {
166
+ const stitchOpts = args.resolved.provider.getStitchOptions?.(args.resolved.modelId);
167
+ if (!stitchOpts) {
168
+ throw new VolumeAdjustmentUnsupportedError(args.modelIdentifier);
169
+ }
170
+ return { ...args.callerOptions, ...stitchOpts.providerOptions };
171
+ }
172
+ if (args.output != null) {
173
+ // Native path: provider produces the requested format directly, so caller's
174
+ // providerOptions are an explicit escape hatch and win over our defaults
175
+ // (e.g. tweaking bitrate/sample_rate). Post-processing reads the actual
176
+ // response mediaType and adapts.
177
+ const native = args.resolved.provider.resolveOutputFormat?.(args.resolved.modelId, args.output);
178
+ if (native) {
179
+ return { ...native.providerOptions, ...args.callerOptions };
180
+ }
181
+ // Stitch fallback: the SDK MUST decode locally to convert into the
182
+ // requested format, so SDK-required keys win — same rationale as the
183
+ // volumeDbfs path above.
184
+ const stitchOpts = args.resolved.provider.getStitchOptions?.(args.resolved.modelId);
185
+ if (!stitchOpts) {
186
+ throw new OutputConversionUnsupportedError(args.modelIdentifier);
187
+ }
188
+ return { ...args.callerOptions, ...stitchOpts.providerOptions };
189
+ }
190
+ return args.callerOptions;
191
+ }
192
+ async function applyLocalAudioPostProcessing(args) {
193
+ let bytes = args.audio;
194
+ let mediaType = args.mediaType;
195
+ if (args.volumeDbfs != null) {
196
+ const { adjustVolume } = await import("./volume-adjust.js");
197
+ bytes = await adjustVolume({
198
+ audio: args.audio,
199
+ mediaType: args.mediaType,
200
+ volumeDbfs: args.volumeDbfs,
201
+ });
202
+ mediaType = "audio/wav";
203
+ }
204
+ if (args.output != null) {
205
+ const decoded = new DefaultGeneratedAudioFile({
206
+ data: bytes,
207
+ mediaType,
208
+ }).uint8Array;
209
+ const converted = await applyOptionalOutputConversion({
210
+ audio: decoded,
211
+ mediaType,
212
+ output: args.output,
213
+ });
214
+ bytes = converted.audio;
215
+ mediaType = converted.mediaType;
216
+ }
217
+ return { bytes, mediaType };
218
+ }
119
219
  function preprocessText(resolved, rawText, modelIdentifier) {
220
+ // Gateway server handles audio-tag normalization itself — pass raw text through.
221
+ if (isSpeechGatewayModel(resolved)) {
222
+ return { text: rawText, warnings: [] };
223
+ }
120
224
  if (resolved.provider.processAudioTags) {
121
225
  return resolved.provider.processAudioTags(rawText, resolved.modelId);
122
226
  }
@@ -126,32 +230,19 @@ function preprocessText(resolved, rawText, modelIdentifier) {
126
230
  }
127
231
  return { text: rawText, warnings: [] };
128
232
  }
129
- /**
130
- * Logs the timestamp routing decision at debug level so developers can see
131
- * why they are / aren't getting alignment data. Silent unless `DEBUG`
132
- * includes `speech-sdk` (or `*`).
133
- */
134
233
  function logTimestampDecision(args) {
135
- const { modelIdentifier, mode, willRequestNative } = args;
136
- if (mode === "off") {
137
- debug(`${modelIdentifier}: timestamps: "off" — skipping alignment.`);
234
+ const { modelIdentifier, enabled, willRequestNative } = args;
235
+ if (!enabled) {
236
+ debug(`${modelIdentifier}: timestamps: false — skipping alignment.`);
138
237
  return;
139
238
  }
140
239
  if (willRequestNative) {
141
- debug(`${modelIdentifier}: timestamps: "${mode}" — requesting native alignment from the provider.`);
142
- return;
143
- }
144
- if (mode === "auto") {
145
- debug(`${modelIdentifier}: timestamps: "auto" — model has no native alignment; skipping. Pass timestamps: "on" to derive via STT (adds a round-trip of the synthesized audio through Whisper by default).`);
240
+ debug(`${modelIdentifier}: timestamps: true — requesting native alignment from the provider.`);
146
241
  return;
147
242
  }
148
- // mode === "on" and no native support → will fall back to STT
149
- debug(`${modelIdentifier}: timestamps: "on" but no native alignment available — will pipe synthesized audio through ${describeSTTTarget(args.timestampProvider)} for word timestamps (adds a round-trip).`);
150
- }
151
- function describeSTTTarget(provider) {
152
- if (provider) {
153
- return `${provider.provider.id}/${provider.modelId}`;
154
- }
155
- return "openai/whisper-1 (default)";
243
+ const target = args.effectiveFallback
244
+ ? `${args.effectiveFallback.provider.id}/${args.effectiveFallback.modelId}`
245
+ : "unconfigured STT fallback";
246
+ debug(`${modelIdentifier}: timestamps: true but no native alignment available — will pipe synthesized audio through ${target} for word timestamps (adds a round-trip).`);
156
247
  }
157
248
  //# sourceMappingURL=generate-speech.js.map
@@ -1 +1 @@
1
- {"version":3,"file":"generate-speech.js","sourceRoot":"","sources":["../src/generate-speech.ts"],"names":[],"mappings":"AAAA,OAAO,MAAM,MAAM,SAAS,CAAC;AAC7B,OAAO,EAAE,oBAAoB,EAAE,MAAM,qBAAqB,CAAC;AAC3D,OAAO,EAAE,eAAe,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AAClE,OAAO,EAAE,sBAAsB,EAAE,MAAM,wBAAwB,CAAC;AAChE,OAAO,EACL,QAAQ,EACR,sBAAsB,EACtB,gCAAgC,GACjC,MAAM,aAAa,CAAC;AACrB,OAAO,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AAEpC,OAAO,EAAE,YAAY,EAAE,MAAM,uBAAuB,CAAC;AACrD,OAAO,EACL,6BAA6B,GAG9B,MAAM,sBAAsB,CAAC;AAE9B,OAAO,EAAE,yBAAyB,EAAE,MAAM,oBAAoB,CAAC;AAI/D,MAAM,CAAC,KAAK,UAAU,cAAc,CAA0B,OAkC7D;IACC,MAAM,EACJ,KAAK,EACL,KAAK,EACL,WAAW,EACX,OAAO,EACP,UAAU,EACV,UAAU,EAAE,aAAa,GAAG,MAAM,EAClC,iBAAiB,GAClB,GAAG,OAAO,CAAC;IACZ,MAAM,UAAU,GAAG,OAAO,CAAC,UAAU,IAAI,CAAC,CAAC;IAE3C,MAAM,QAAQ,GAAG,YAAY,CAAC,KAAK,EAAE,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;IACjE,MAAM,eAAe,GAAG,GAAG,QAAQ,CAAC,QAAQ,CAAC,EAAE,IAAI,QAAQ,CAAC,OAAO,EAAE,CAAC;IAEtE,IAAI,eAAe,GAAG,OAAO,CAAC,eAAe,CAAC;IAE9C,IAAI,UAAU,IAAI,IAAI,EAAE,CAAC;QACvB,MAAM,UAAU,GAAG,QAAQ,CAAC,QAAQ,CAAC,gBAAgB,EAAE,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;QAC1E,IAAI,CAAC,UAAU,EAAE,CAAC;YAChB,MAAM,IAAI,gCAAgC,CAAC,eAAe,CAAC,CAAC;QAC9D,CAAC;QACD,sEAAsE;QACtE,wEAAwE;QACxE,mEAAmE;QACnE,eAAe,GAAG;YAChB,GAAG,OAAO,CAAC,eAAe;YAC1B,GAAG,UAAU,CAAC,eAAe;SAC9B,CAAC;IACJ,CAAC;IAED,MAAM,EAAE,IAAI,EAAE,aAAa,EAAE,QAAQ,EAAE,GAAG,cAAc,CACtD,QAAQ,EACR,OAAO,CAAC,IAAI,EACZ,eAAe,CAChB,CAAC;IAEF,IAAI,aAAa,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACtC,MAAM,IAAI,sBAAsB,CAC9B,QAAQ,CAAC,MAAM,GAAG,CAAC;YACjB,CAAC,CAAC,2DAA2D,eAAe,GAAG;YAC/E,CAAC,CAAC,yBAAyB,CAC9B,CAAC;IACJ,CAAC;IAED,MAAM,mBAAmB,GAAG,6BAA6B,CAAC,QAAQ,CAAC,CAAC;IAEpE,uEAAuE;IACvE,yCAAyC;IACzC,MAAM,mBAAmB,GACvB,CAAC,aAAa,KAAK,IAAI,IAAI,aAAa,KAAK,MAAM,CAAC,IAAI,mBAAmB,CAAC;IAE9E,oBAAoB,CAAC;QACnB,eAAe;QACf,IAAI,EAAE,aAAa;QACnB,SAAS,EAAE,mBAAmB;QAC9B,iBAAiB,EAAE,mBAAmB;QACtC,iBAAiB;KAClB,CAAC,CAAC;IAEH,MAAM,SAAS,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC;IAEpC,MAAM,MAAM,GAAG,MAAM,MAAM,CACzB,GAAG,EAAE,CACH,QAAQ,CAAC,QAAQ,CAAC,QAAQ,CAAC;QACzB,OAAO,EAAE,QAAQ,CAAC,OAAO;QACzB,IAAI,EAAE,aAAa;QACnB,KAAK;QACL,eAAe;QACf,WAAW;QACX,OAAO;QACP,iBAAiB,EAAE,mBAAmB;KACvC,CAAC,EACJ;QACE,OAAO,EAAE,UAAU;QACnB,MAAM,EAAE,WAAW;QACnB,WAAW,EAAE,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE;YACzB,IAAI,KAAK,YAAY,QAAQ,IAAI,KAAK,CAAC,UAAU,GAAG,GAAG,EAAE,CAAC;gBACxD,OAAO,KAAK,CAAC;YACf,CAAC;YACD,OAAO,IAAI,CAAC;QACd,CAAC;KACF,CACF,CAAC;IAEF,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,GAAG,EAAE,GAAG,SAAS,CAAC,CAAC;IAE5D,MAAM,SAAS,GAAG,MAAM,CAAC,KAAK,CAAC;IAE/B,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC3B,MAAM,IAAI,sBAAsB,EAAE,CAAC;IACrC,CAAC;IAED,IAAI,WAAW,GAAwB,SAAS,CAAC;IACjD,IAAI,eAAe,GAAG,MAAM,CAAC,SAAS,CAAC;IAEvC,IAAI,UAAU,IAAI,IAAI,EAAE,CAAC;QACvB,MAAM,EAAE,YAAY,EAAE,GAAG,MAAM,MAAM,CAAC,oBAAoB,CAAC,CAAC;QAC5D,WAAW,GAAG,MAAM,YAAY,CAAC;YAC/B,KAAK,EAAE,SAAS;YAChB,SAAS,EAAE,MAAM,CAAC,SAAS;YAC3B,UAAU;SACX,CAAC,CAAC;QACH,eAAe,GAAG,WAAW,CAAC;IAChC,CAAC;IAED,MAAM,KAAK,GAAG,IAAI,yBAAyB,CAAC;QAC1C,IAAI,EAAE,WAAW;QACjB,SAAS,EAAE,eAAe;KAC3B,CAAC,CAAC;IAEH,MAAM,eAAe,GACnB,CAAC,MAAM,oBAAoB,CAAC,KAAK,CAAC,UAAU,EAAE,eAAe,CAAC,CAAC;QAC/D,MAAM,CAAC,eAAe,CAAC;IAEzB,IAAI,UAAgD,CAAC;IACrD,IAAI,aAAa,KAAK,KAAK,EAAE,CAAC;QAC5B,IAAI,MAAM,CAAC,UAAU,IAAI,MAAM,CAAC,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACtD,KAAK,CACH,GAAG,eAAe,cAAc,MAAM,CAAC,UAAU,CAAC,MAAM,0BAA0B,CACnF,CAAC;YACF,UAAU,GAAG,MAAM,CAAC,UAAU,CAAC;QACjC,CAAC;aAAM,IAAI,aAAa,KAAK,IAAI,EAAE,CAAC;YAClC,UAAU,GAAG,MAAM,sBAAsB,CAAC;gBACxC,QAAQ,EAAE,eAAe;gBACzB,KAAK,EAAE,KAAK,CAAC,UAAU;gBACvB,SAAS,EAAE,eAAe;gBAC1B,iBAAiB;gBACjB,WAAW;aACZ,CAAC,CAAC;YACH,KAAK,CACH,GAAG,eAAe,aAAa,UAAU,CAAC,MAAM,oCAAoC,CACrF,CAAC;QACJ,CAAC;IACH,CAAC;IAED,MAAM,QAAQ,GAAmB;QAC/B,SAAS;QACT,UAAU,EAAE,aAAa,CAAC,MAAM;QAChC,QAAQ,EAAE,QAAQ,CAAC,QAAQ,CAAC,EAAE;QAC9B,KAAK,EAAE,QAAQ,CAAC,OAAO;QACvB,GAAG,CAAC,eAAe,IAAI,IAAI,IAAI,EAAE,eAAe,EAAE,CAAC;KACpD,CAAC;IAEF,OAAO;QACL,KAAK;QACL,QAAQ;QACR,gBAAgB,EAAE,MAAM,CAAC,gBAAgB;QACzC,QAAQ,EAAE,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,SAAS;QACpD,UAAU;KACX,CAAC;AACJ,CAAC;AAED,SAAS,cAAc,CACrB,QAAuB,EACvB,OAAe,EACf,eAAuB;IAEvB,IAAI,QAAQ,CAAC,QAAQ,CAAC,gBAAgB,EAAE,CAAC;QACvC,OAAO,QAAQ,CAAC,QAAQ,CAAC,gBAAgB,CAAC,OAAO,EAAE,QAAQ,CAAC,OAAO,CAAC,CAAC;IACvE,CAAC;IACD,MAAM,IAAI,GAAG,eAAe,CAAC,OAAO,CAAC,CAAC;IACtC,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACpB,OAAO,cAAc,CAAC,OAAO,EAAE,eAAe,CAAC,CAAC;IAClD,CAAC;IACD,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;AACzC,CAAC;AAED;;;;GAIG;AACH,SAAS,oBAAoB,CAAC,IAM7B;IACC,MAAM,EAAE,eAAe,EAAE,IAAI,EAAE,iBAAiB,EAAE,GAAG,IAAI,CAAC;IAC1D,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;QACnB,KAAK,CAAC,GAAG,eAAe,2CAA2C,CAAC,CAAC;QACrE,OAAO;IACT,CAAC;IACD,IAAI,iBAAiB,EAAE,CAAC;QACtB,KAAK,CACH,GAAG,eAAe,kBAAkB,IAAI,oDAAoD,CAC7F,CAAC;QACF,OAAO;IACT,CAAC;IACD,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;QACpB,KAAK,CACH,GAAG,eAAe,kLAAkL,CACrM,CAAC;QACF,OAAO;IACT,CAAC;IACD,8DAA8D;IAC9D,KAAK,CACH,GAAG,eAAe,8FAA8F,iBAAiB,CAAC,IAAI,CAAC,iBAAiB,CAAC,2CAA2C,CACrM,CAAC;AACJ,CAAC;AAED,SAAS,iBAAiB,CAAC,QAAsC;IAC/D,IAAI,QAAQ,EAAE,CAAC;QACb,OAAO,GAAG,QAAQ,CAAC,QAAQ,CAAC,EAAE,IAAI,QAAQ,CAAC,OAAO,EAAE,CAAC;IACvD,CAAC;IACD,OAAO,4BAA4B,CAAC;AACtC,CAAC"}
1
+ {"version":3,"file":"generate-speech.js","sourceRoot":"","sources":["../src/generate-speech.ts"],"names":[],"mappings":"AAAA,OAAO,MAAM,MAAM,SAAS,CAAC;AAC7B,OAAO,EAAE,oBAAoB,EAAE,MAAM,qBAAqB,CAAC;AAC3D,OAAO,EAEL,6BAA6B,EAC7B,cAAc,GACf,MAAM,mBAAmB,CAAC;AAC3B,OAAO,EAAE,eAAe,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AAClE,OAAO,EAAE,qBAAqB,EAAE,MAAM,2BAA2B,CAAC;AAClE,OAAO,EAAE,sBAAsB,EAAE,MAAM,wBAAwB,CAAC;AAChE,OAAO,EACL,sBAAsB,EACtB,gCAAgC,EAChC,gCAAgC,GACjC,MAAM,aAAa,CAAC;AACrB,OAAO,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AAEpC,OAAO,EAAE,YAAY,EAAE,MAAM,mCAAmC,CAAC;AACjE,OAAO,EAAE,UAAU,EAAE,MAAM,2BAA2B,CAAC;AACvD,OAAO,EAAE,UAAU,EAAE,MAAM,gCAAgC,CAAC;AAE5D,OAAO,EAAE,2BAA2B,EAAE,MAAM,8BAA8B,CAAC;AAE3E,OAAO,EAAE,YAAY,EAAE,MAAM,uBAAuB,CAAC;AACrD,OAAO,EAAE,iBAAiB,EAAE,MAAM,oBAAoB,CAAC;AACvD,OAAO,EACL,oBAAoB,EACpB,6BAA6B,GAG9B,MAAM,sBAAsB,CAAC;AAE9B,OAAO,EAAE,yBAAyB,EAAE,MAAM,oBAAoB,CAAC;AAI/D,MAAM,CAAC,KAAK,UAAU,cAAc,CAGlC,OAcD;IACC,MAAM,EACJ,KAAK,EACL,KAAK,EACL,WAAW,EACX,OAAO,EACP,UAAU,EACV,UAAU,GAAG,KAAK,GACnB,GAAG,OAAO,CAAC;IACZ,MAAM,UAAU,GAAG,OAAO,CAAC,UAAU,IAAI,CAAC,CAAC;IAE3C,cAAc,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;IAE/B,MAAM,QAAQ,GAAG,YAAY,CAAC,KAAK,EAAE,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;IACjE,MAAM,eAAe,GAAG,GAAG,QAAQ,CAAC,QAAQ,CAAC,EAAE,IAAI,QAAQ,CAAC,OAAO,EAAE,CAAC;IACtE,MAAM,SAAS,GAAG,oBAAoB,CAAC,QAAQ,CAAC,CAAC;IAEjD,2BAA2B,CAAC,OAAO,CAAC,cAAc,EAAE,SAAS,CAAC,CAAC;IAE/D,MAAM,EAAE,IAAI,EAAE,YAAY,EAAE,QAAQ,EAAE,GAAG,cAAc,CACrD,QAAQ,EACR,OAAO,CAAC,IAAI,EACZ,eAAe,CAChB,CAAC;IAEF,IAAI,YAAY,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACrC,MAAM,IAAI,sBAAsB,CAC9B,QAAQ,CAAC,MAAM,GAAG,CAAC;YACjB,CAAC,CAAC,2DAA2D,eAAe,GAAG;YAC/E,CAAC,CAAC,yBAAyB,CAC9B,CAAC;IACJ,CAAC;IAED,IAAI,UAAU,GAAG,YAAY,CAAC;IAC9B,IAAI,kBAAkB,GAAoB,EAAE,CAAC;IAC7C,IAAI,CAAC,SAAS,IAAI,OAAO,CAAC,cAAc,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;QACxD,MAAM,OAAO,GAAG,UAAU,CAAC,OAAO,CAAC,cAAc,CAAC,KAAK,CAAC,CAAC;QACzD,MAAM,MAAM,GAAG,UAAU,CAAC,YAAY,EAAE,OAAO,CAAC,CAAC;QACjD,UAAU,GAAG,MAAM,CAAC,IAAI,CAAC;QACzB,kBAAkB,GAAG,MAAM,CAAC,KAAK,CAAC;IACpC,CAAC;IAED,MAAM,eAAe,GAAG,sCAAsC,CAAC;QAC7D,QAAQ;QACR,SAAS;QACT,eAAe;QACf,UAAU;QACV,MAAM,EAAE,OAAO,CAAC,MAAM;QACtB,aAAa,EAAE,OAAO,CAAC,eAAe;KACvC,CAAC,CAAC;IAEH,MAAM,mBAAmB,GAAG,6BAA6B,CAAC,QAAQ,CAAC,CAAC;IACpE,MAAM,mBAAmB,GAAG,UAAU,IAAI,CAAC,mBAAmB,IAAI,SAAS,CAAC,CAAC;IAE7E,MAAM,iBAAiB,GACrB,CAAC,UAAU,IAAI,mBAAmB;QAChC,CAAC,CAAC,SAAS;QACX,CAAC,CAAC,CAAC,QAAQ,CAAC,WAAW,IAAI,CAAC,MAAM,qBAAqB,EAAE,CAAC,CAAC,CAAC;IAChE,oBAAoB,CAAC;QACnB,eAAe;QACf,OAAO,EAAE,UAAU;QACnB,SAAS,EAAE,mBAAmB;QAC9B,iBAAiB,EAAE,mBAAmB;QACtC,iBAAiB;KAClB,CAAC,CAAC;IAEH,MAAM,SAAS,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC;IAEpC,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,GAAG,EAAE;QAC/B,IAAI,SAAS,EAAE,CAAC;YACd,MAAM,eAAe,GAAG,QAAQ,CAAC,QAAiC,CAAC;YACnE,OAAO,eAAe,CAAC,QAAQ,CAAC;gBAC9B,OAAO,EAAE,QAAQ,CAAC,OAAO;gBACzB,IAAI,EAAE,UAAU;gBAChB,KAAK,EAAE,KAA0B;gBACjC,eAAe;gBACf,WAAW;gBACX,OAAO;gBACP,iBAAiB,EAAE,mBAAmB;gBACtC,UAAU;gBACV,MAAM,EAAE,OAAO,CAAC,MAAM;gBACtB,cAAc,EAAE,OAAO,CAAC,cAAc;aACvC,CAAC,CAAC;QACL,CAAC;QACD,OAAO,QAAQ,CAAC,QAAQ,CAAC,QAAQ,CAAC;YAChC,OAAO,EAAE,QAAQ,CAAC,OAAO;YACzB,IAAI,EAAE,UAAU;YAChB,KAAK;YACL,eAAe;YACf,WAAW;YACX,OAAO;YACP,iBAAiB,EAAE,mBAAmB;SACvC,CAAC,CAAC;IACL,CAAC,EAAE,iBAAiB,CAAC,EAAE,UAAU,EAAE,WAAW,EAAE,CAAC,CAAC,CAAC;IAEnD,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,GAAG,EAAE,GAAG,SAAS,CAAC,CAAC;IAE5D,MAAM,SAAS,GAAG,MAAM,CAAC,KAAK,CAAC;IAE/B,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC3B,MAAM,IAAI,sBAAsB,EAAE,CAAC;IACrC,CAAC;IAED,MAAM,EAAE,KAAK,EAAE,WAAW,EAAE,SAAS,EAAE,eAAe,EAAE,GAAG,SAAS;QAClE,CAAC,CAAC,EAAE,KAAK,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,CAAC,SAAS,EAAE;QACnD,CAAC,CAAC,MAAM,6BAA6B,CAAC;YAClC,KAAK,EAAE,SAAS;YAChB,SAAS,EAAE,MAAM,CAAC,SAAS;YAC3B,UAAU;YACV,MAAM,EAAE,OAAO,CAAC,MAAM;SACvB,CAAC,CAAC;IAEP,MAAM,KAAK,GAAG,IAAI,yBAAyB,CAAC;QAC1C,IAAI,EAAE,WAAW;QACjB,SAAS,EAAE,eAAe;KAC3B,CAAC,CAAC;IAEH,MAAM,CAAC,gBAAgB,EAAE,kBAAkB,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;QAC/D,oBAAoB,CAAC,KAAK,CAAC,UAAU,EAAE,eAAe,CAAC;QACvD,iBAAiB,CAAC;YAChB,UAAU;YACV,eAAe;YACf,QAAQ;YACR,gBAAgB,EAAE,MAAM,CAAC,UAAU;YACnC,KAAK,EAAE,KAAK,CAAC,UAAU;YACvB,SAAS,EAAE,eAAe;YAC1B,WAAW;SACZ,CAAC;KACH,CAAC,CAAC;IACH,MAAM,eAAe,GAAG,gBAAgB,IAAI,MAAM,CAAC,eAAe,CAAC;IAEnE,MAAM,eAAe,GACnB,kBAAkB,IAAI,kBAAkB,CAAC,MAAM,GAAG,CAAC;QACjD,CAAC,CAAC,YAAY,CAAC,kBAAkB,EAAE,UAAU,EAAE,kBAAkB,CAAC;QAClE,CAAC,CAAC,kBAAkB,CAAC;IAEzB,MAAM,QAAQ,GAAmB;QAC/B,SAAS;QACT,UAAU,EAAE,OAAO,CAAC,IAAI,CAAC,MAAM;QAC/B,GAAG,CAAC,eAAe,IAAI,IAAI,IAAI,EAAE,eAAe,EAAE,CAAC;KACpD,CAAC;IAEF,OAAO;QACL,KAAK;QACL,QAAQ;QACR,gBAAgB,EAAE,MAAM,CAAC,gBAAgB;QACzC,QAAQ,EAAE,aAAa,CAAC,QAAQ,EAAE,MAAM,CAAC,QAAQ,CAAC;QAClD,UAAU,EAAE,eAAe;KAC5B,CAAC;AACJ,CAAC;AAED,SAAS,aAAa,CACpB,qBAA+B,EAC/B,gBAAsC;IAEtC,MAAM,MAAM,GAAG,CAAC,GAAG,qBAAqB,EAAE,GAAG,CAAC,gBAAgB,IAAI,EAAE,CAAC,CAAC,CAAC;IACvE,OAAO,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,SAAS,CAAC;AAChD,CAAC;AAED,KAAK,UAAU,iBAAiB,CAAC,IAQhC;IACC,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,CAAC;QACrB,OAAO;IACT,CAAC;IACD,IAAI,IAAI,CAAC,gBAAgB,EAAE,MAAM,EAAE,CAAC;QAClC,KAAK,CACH,GAAG,IAAI,CAAC,eAAe,cAAc,IAAI,CAAC,gBAAgB,CAAC,MAAM,0BAA0B,CAC5F,CAAC;QACF,OAAO,IAAI,CAAC,gBAAgB,CAAC;IAC/B,CAAC;IACD,IAAI,oBAAoB,CAAC,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAC;QACxC,OAAO;IACT,CAAC;IACD,MAAM,QAAQ,GAAG,IAAI,CAAC,QAAQ,CAAC,WAAW,IAAI,CAAC,MAAM,qBAAqB,EAAE,CAAC,CAAC;IAC9E,MAAM,UAAU,GAAG,MAAM,sBAAsB,CAAC;QAC9C,QAAQ,EAAE,IAAI,CAAC,eAAe;QAC9B,KAAK,EAAE,IAAI,CAAC,KAAK;QACjB,SAAS,EAAE,IAAI,CAAC,SAAS;QACzB,iBAAiB,EAAE,QAAQ;QAC3B,WAAW,EAAE,IAAI,CAAC,WAAW;KAC9B,CAAC,CAAC;IACH,KAAK,CACH,GAAG,IAAI,CAAC,eAAe,aAAa,UAAU,CAAC,MAAM,oCAAoC,CAC1F,CAAC;IACF,OAAO,UAAU,CAAC;AACpB,CAAC;AAED,SAAS,sCAAsC,CAAC,IAO/C;IACC,IAAI,IAAI,CAAC,SAAS,EAAE,CAAC;QACnB,OAAO,IAAI,CAAC,aAAa,CAAC;IAC5B,CAAC;IAED,+EAA+E;IAC/E,qEAAqE;IACrE,8DAA8D;IAC9D,IAAI,IAAI,CAAC,UAAU,IAAI,IAAI,EAAE,CAAC;QAC5B,MAAM,UAAU,GAAG,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,gBAAgB,EAAE,CAC1D,IAAI,CAAC,QAAQ,CAAC,OAAO,CACtB,CAAC;QACF,IAAI,CAAC,UAAU,EAAE,CAAC;YAChB,MAAM,IAAI,gCAAgC,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC;QACnE,CAAC;QACD,OAAO,EAAE,GAAG,IAAI,CAAC,aAAa,EAAE,GAAG,UAAU,CAAC,eAAe,EAAE,CAAC;IAClE,CAAC;IAED,IAAI,IAAI,CAAC,MAAM,IAAI,IAAI,EAAE,CAAC;QACxB,4EAA4E;QAC5E,yEAAyE;QACzE,wEAAwE;QACxE,iCAAiC;QACjC,MAAM,MAAM,GAAG,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,mBAAmB,EAAE,CACzD,IAAI,CAAC,QAAQ,CAAC,OAAO,EACrB,IAAI,CAAC,MAAM,CACZ,CAAC;QACF,IAAI,MAAM,EAAE,CAAC;YACX,OAAO,EAAE,GAAG,MAAM,CAAC,eAAe,EAAE,GAAG,IAAI,CAAC,aAAa,EAAE,CAAC;QAC9D,CAAC;QACD,mEAAmE;QACnE,qEAAqE;QACrE,yBAAyB;QACzB,MAAM,UAAU,GAAG,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,gBAAgB,EAAE,CAC1D,IAAI,CAAC,QAAQ,CAAC,OAAO,CACtB,CAAC;QACF,IAAI,CAAC,UAAU,EAAE,CAAC;YAChB,MAAM,IAAI,gCAAgC,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC;QACnE,CAAC;QACD,OAAO,EAAE,GAAG,IAAI,CAAC,aAAa,EAAE,GAAG,UAAU,CAAC,eAAe,EAAE,CAAC;IAClE,CAAC;IAED,OAAO,IAAI,CAAC,aAAa,CAAC;AAC5B,CAAC;AAED,KAAK,UAAU,6BAA6B,CAAC,IAK5C;IACC,IAAI,KAAK,GAAwB,IAAI,CAAC,KAAK,CAAC;IAC5C,IAAI,SAAS,GAAG,IAAI,CAAC,SAAS,CAAC;IAE/B,IAAI,IAAI,CAAC,UAAU,IAAI,IAAI,EAAE,CAAC;QAC5B,MAAM,EAAE,YAAY,EAAE,GAAG,MAAM,MAAM,CAAC,oBAAoB,CAAC,CAAC;QAC5D,KAAK,GAAG,MAAM,YAAY,CAAC;YACzB,KAAK,EAAE,IAAI,CAAC,KAAK;YACjB,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,UAAU,EAAE,IAAI,CAAC,UAAU;SAC5B,CAAC,CAAC;QACH,SAAS,GAAG,WAAW,CAAC;IAC1B,CAAC;IAED,IAAI,IAAI,CAAC,MAAM,IAAI,IAAI,EAAE,CAAC;QACxB,MAAM,OAAO,GAAG,IAAI,yBAAyB,CAAC;YAC5C,IAAI,EAAE,KAAK;YACX,SAAS;SACV,CAAC,CAAC,UAAU,CAAC;QACd,MAAM,SAAS,GAAG,MAAM,6BAA6B,CAAC;YACpD,KAAK,EAAE,OAAO;YACd,SAAS;YACT,MAAM,EAAE,IAAI,CAAC,MAAM;SACpB,CAAC,CAAC;QACH,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC;QACxB,SAAS,GAAG,SAAS,CAAC,SAAS,CAAC;IAClC,CAAC;IAED,OAAO,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC;AAC9B,CAAC;AAED,SAAS,cAAc,CACrB,QAAuB,EACvB,OAAe,EACf,eAAuB;IAEvB,iFAAiF;IACjF,IAAI,oBAAoB,CAAC,QAAQ,CAAC,EAAE,CAAC;QACnC,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;IACzC,CAAC;IACD,IAAI,QAAQ,CAAC,QAAQ,CAAC,gBAAgB,EAAE,CAAC;QACvC,OAAO,QAAQ,CAAC,QAAQ,CAAC,gBAAgB,CAAC,OAAO,EAAE,QAAQ,CAAC,OAAO,CAAC,CAAC;IACvE,CAAC;IACD,MAAM,IAAI,GAAG,eAAe,CAAC,OAAO,CAAC,CAAC;IACtC,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACpB,OAAO,cAAc,CAAC,OAAO,EAAE,eAAe,CAAC,CAAC;IAClD,CAAC;IACD,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;AACzC,CAAC;AAED,SAAS,oBAAoB,CAAC,IAM7B;IACC,MAAM,EAAE,eAAe,EAAE,OAAO,EAAE,iBAAiB,EAAE,GAAG,IAAI,CAAC;IAC7D,IAAI,CAAC,OAAO,EAAE,CAAC;QACb,KAAK,CAAC,GAAG,eAAe,2CAA2C,CAAC,CAAC;QACrE,OAAO;IACT,CAAC;IACD,IAAI,iBAAiB,EAAE,CAAC;QACtB,KAAK,CACH,GAAG,eAAe,qEAAqE,CACxF,CAAC;QACF,OAAO;IACT,CAAC;IACD,MAAM,MAAM,GAAG,IAAI,CAAC,iBAAiB;QACnC,CAAC,CAAC,GAAG,IAAI,CAAC,iBAAiB,CAAC,QAAQ,CAAC,EAAE,IAAI,IAAI,CAAC,iBAAiB,CAAC,OAAO,EAAE;QAC3E,CAAC,CAAC,2BAA2B,CAAC;IAChC,KAAK,CACH,GAAG,eAAe,8FAA8F,MAAM,2CAA2C,CAClK,CAAC;AACJ,CAAC"}
package/dist/index.d.ts CHANGED
@@ -1,15 +1,11 @@
1
- export { detectAudioTags, stripAudioTags } from "./audio-tags.js";
2
- export type { CaptionFormat, CaptionsOptions } from "./captions.js";
1
+ export type { AudioOutput, AudioOutputFormat } from "./audio-output.js";
3
2
  export { timestampsToCaptions } from "./captions.js";
4
- export { ApiError, NoSpeechGeneratedError, SpeechSDKError, StreamingNotSupportedError, TimestampKeyMissingError, } from "./errors.js";
3
+ export { ConversationInputError, DialogueConstraintError, StitchUnsupportedError, } from "./conversation/errors.js";
4
+ export { ApiError, AudioOutputInputError, GatewayInputError, MissingApiKeyError, NoSpeechGeneratedError, OutputConversionUnsupportedError, SpeechSDKError, StreamingNotSupportedError, TimestampKeyMissingError, VolumeAdjustmentUnsupportedError, } from "./errors.js";
5
+ export { generateConversation } from "./generate-conversation.js";
5
6
  export { generateSpeech } from "./generate-speech.js";
6
- export type { SpeechMetadata } from "./metadata.js";
7
- export type { Feature, ModelInfo, ResolvedModel, SpeechProvider, TimestampsFeature, Voice, } from "./speech-provider.js";
8
- export { FEATURES, getFeature, hasFeature, } from "./speech-provider.js";
9
- export type { GeneratedAudioFile, SpeechResult } from "./speech-result.js";
10
- export type { ResolvedSTTModel, SpeechToTextProvider, STTModelInfo, } from "./speech-to-text-provider.js";
7
+ export { DictionaryIdsRequireGatewayError } from "./pronunciations/errors.js";
8
+ export type { Pronunciation, PronunciationsInput, } from "./pronunciations/types.js";
11
9
  export { streamSpeech } from "./stream-speech.js";
12
- export type { StreamSpeechResult } from "./stream-speech-result.js";
13
- export type { TimestampMode, WordTimestamp } from "./timestamps.js";
14
- export type { GenerateSpeechOptions } from "./types.js";
10
+ export { timestampsToTurns } from "./turns.js";
15
11
  //# sourceMappingURL=index.d.ts.map
@@ -1 +1 @@
1
- {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,eAAe,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AAClE,YAAY,EAAE,aAAa,EAAE,eAAe,EAAE,MAAM,eAAe,CAAC;AACpE,OAAO,EAAE,oBAAoB,EAAE,MAAM,eAAe,CAAC;AACrD,OAAO,EACL,QAAQ,EACR,sBAAsB,EACtB,cAAc,EACd,0BAA0B,EAC1B,wBAAwB,GACzB,MAAM,aAAa,CAAC;AACrB,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AACtD,YAAY,EAAE,cAAc,EAAE,MAAM,eAAe,CAAC;AACpD,YAAY,EACV,OAAO,EACP,SAAS,EACT,aAAa,EACb,cAAc,EACd,iBAAiB,EACjB,KAAK,GACN,MAAM,sBAAsB,CAAC;AAC9B,OAAO,EACL,QAAQ,EACR,UAAU,EACV,UAAU,GACX,MAAM,sBAAsB,CAAC;AAC9B,YAAY,EAAE,kBAAkB,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAC3E,YAAY,EACV,gBAAgB,EAChB,oBAAoB,EACpB,YAAY,GACb,MAAM,8BAA8B,CAAC;AACtC,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAClD,YAAY,EAAE,kBAAkB,EAAE,MAAM,2BAA2B,CAAC;AACpE,YAAY,EAAE,aAAa,EAAE,aAAa,EAAE,MAAM,iBAAiB,CAAC;AACpE,YAAY,EAAE,qBAAqB,EAAE,MAAM,YAAY,CAAC"}
1
+ {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,YAAY,EAAE,WAAW,EAAE,iBAAiB,EAAE,MAAM,mBAAmB,CAAC;AAExE,OAAO,EAAE,oBAAoB,EAAE,MAAM,eAAe,CAAC;AACrD,OAAO,EACL,sBAAsB,EACtB,uBAAuB,EACvB,sBAAsB,GACvB,MAAM,0BAA0B,CAAC;AAClC,OAAO,EACL,QAAQ,EACR,qBAAqB,EACrB,iBAAiB,EACjB,kBAAkB,EAClB,sBAAsB,EACtB,gCAAgC,EAChC,cAAc,EACd,0BAA0B,EAC1B,wBAAwB,EACxB,gCAAgC,GACjC,MAAM,aAAa,CAAC;AACrB,OAAO,EAAE,oBAAoB,EAAE,MAAM,4BAA4B,CAAC;AAClE,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AACtD,OAAO,EAAE,gCAAgC,EAAE,MAAM,4BAA4B,CAAC;AAC9E,YAAY,EACV,aAAa,EACb,mBAAmB,GACpB,MAAM,2BAA2B,CAAC;AACnC,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAClD,OAAO,EAAE,iBAAiB,EAAE,MAAM,YAAY,CAAC"}
package/dist/index.js CHANGED
@@ -1,8 +1,10 @@
1
- // biome-ignore lint/performance/noBarrelFile: intentional public API barrel
2
- export { detectAudioTags, stripAudioTags } from "./audio-tags.js";
1
+ // biome-ignore lint/performance/noBarrelFile: public API entry point
3
2
  export { timestampsToCaptions } from "./captions.js";
4
- export { ApiError, NoSpeechGeneratedError, SpeechSDKError, StreamingNotSupportedError, TimestampKeyMissingError, } from "./errors.js";
3
+ export { ConversationInputError, DialogueConstraintError, StitchUnsupportedError, } from "./conversation/errors.js";
4
+ export { ApiError, AudioOutputInputError, GatewayInputError, MissingApiKeyError, NoSpeechGeneratedError, OutputConversionUnsupportedError, SpeechSDKError, StreamingNotSupportedError, TimestampKeyMissingError, VolumeAdjustmentUnsupportedError, } from "./errors.js";
5
+ export { generateConversation } from "./generate-conversation.js";
5
6
  export { generateSpeech } from "./generate-speech.js";
6
- export { FEATURES, getFeature, hasFeature, } from "./speech-provider.js";
7
+ export { DictionaryIdsRequireGatewayError } from "./pronunciations/errors.js";
7
8
  export { streamSpeech } from "./stream-speech.js";
9
+ export { timestampsToTurns } from "./turns.js";
8
10
  //# sourceMappingURL=index.js.map
package/dist/index.js.map CHANGED
@@ -1 +1 @@
1
- {"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,4EAA4E;AAC5E,OAAO,EAAE,eAAe,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AAElE,OAAO,EAAE,oBAAoB,EAAE,MAAM,eAAe,CAAC;AACrD,OAAO,EACL,QAAQ,EACR,sBAAsB,EACtB,cAAc,EACd,0BAA0B,EAC1B,wBAAwB,GACzB,MAAM,aAAa,CAAC;AACrB,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AAUtD,OAAO,EACL,QAAQ,EACR,UAAU,EACV,UAAU,GACX,MAAM,sBAAsB,CAAC;AAO9B,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC"}
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AACA,qEAAqE;AACrE,OAAO,EAAE,oBAAoB,EAAE,MAAM,eAAe,CAAC;AACrD,OAAO,EACL,sBAAsB,EACtB,uBAAuB,EACvB,sBAAsB,GACvB,MAAM,0BAA0B,CAAC;AAClC,OAAO,EACL,QAAQ,EACR,qBAAqB,EACrB,iBAAiB,EACjB,kBAAkB,EAClB,sBAAsB,EACtB,gCAAgC,EAChC,cAAc,EACd,0BAA0B,EAC1B,wBAAwB,EACxB,gCAAgC,GACjC,MAAM,aAAa,CAAC;AACrB,OAAO,EAAE,oBAAoB,EAAE,MAAM,4BAA4B,CAAC;AAClE,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AACtD,OAAO,EAAE,gCAAgC,EAAE,MAAM,4BAA4B,CAAC;AAK9E,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAClD,OAAO,EAAE,iBAAiB,EAAE,MAAM,YAAY,CAAC"}
@@ -1 +1 @@
1
- {"version":3,"file":"logger.d.ts","sourceRoot":"","sources":["../src/logger.ts"],"names":[],"mappings":"AAsCA,wBAAgB,KAAK,CAAC,OAAO,EAAE,MAAM,GAAG,IAAI,CAK3C"}
1
+ {"version":3,"file":"logger.d.ts","sourceRoot":"","sources":["../src/logger.ts"],"names":[],"mappings":"AA2BA,wBAAgB,KAAK,CAAC,OAAO,EAAE,MAAM,GAAG,IAAI,CAK3C"}
package/dist/logger.js CHANGED
@@ -1,12 +1,4 @@
1
- /**
2
- * Minimal debug-level logger. Emits a namespaced message only when the
3
- * `DEBUG` env var opts in (convention borrowed from the `debug` npm
4
- * package, without the dependency). Matches any of:
5
- * DEBUG=* enables everything
6
- * DEBUG=speech-sdk enables the SDK
7
- * DEBUG=speech-sdk:* same (wildcard namespace)
8
- * DEBUG=foo,speech-sdk comma list
9
- */
1
+ // Honors DEBUG env var: "*", "speech-sdk", "speech-sdk:*", or comma list.
10
2
  const NAMESPACE = "speech-sdk";
11
3
  function debugEnabled() {
12
4
  if (typeof process === "undefined" || !process.env?.DEBUG) {
@@ -26,10 +18,7 @@ function debugEnabled() {
26
18
  }
27
19
  return false;
28
20
  }
29
- // Evaluated once at module load; avoids reading env on every call in hot
30
- // paths. Developers toggling DEBUG mid-process would need to re-import —
31
- // acceptable trade-off since debug logging is an operator concern set at
32
- // startup, not a runtime setting.
21
+ // Evaluated once toggling DEBUG mid-process requires a re-import.
33
22
  const ENABLED = debugEnabled();
34
23
  export function debug(message) {
35
24
  if (!ENABLED) {
@@ -1 +1 @@
1
- {"version":3,"file":"logger.js","sourceRoot":"","sources":["../src/logger.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AACH,MAAM,SAAS,GAAG,YAAY,CAAC;AAE/B,SAAS,YAAY;IACnB,IAAI,OAAO,OAAO,KAAK,WAAW,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,KAAK,EAAE,CAAC;QAC1D,OAAO,KAAK,CAAC;IACf,CAAC;IACD,MAAM,GAAG,GAAG,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC;IAC9B,IAAI,GAAG,KAAK,GAAG,EAAE,CAAC;QAChB,OAAO,IAAI,CAAC;IACd,CAAC;IACD,KAAK,MAAM,KAAK,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;QACnC,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;QAC7B,IACE,OAAO,KAAK,SAAS;YACrB,OAAO,KAAK,GAAG,SAAS,IAAI;YAC5B,OAAO,CAAC,UAAU,CAAC,GAAG,SAAS,GAAG,CAAC,EACnC,CAAC;YACD,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,yEAAyE;AACzE,yEAAyE;AACzE,yEAAyE;AACzE,kCAAkC;AAClC,MAAM,OAAO,GAAG,YAAY,EAAE,CAAC;AAE/B,MAAM,UAAU,KAAK,CAAC,OAAe;IACnC,IAAI,CAAC,OAAO,EAAE,CAAC;QACb,OAAO;IACT,CAAC;IACD,OAAO,CAAC,KAAK,CAAC,IAAI,SAAS,KAAK,OAAO,EAAE,CAAC,CAAC;AAC7C,CAAC"}
1
+ {"version":3,"file":"logger.js","sourceRoot":"","sources":["../src/logger.ts"],"names":[],"mappings":"AAAA,0EAA0E;AAC1E,MAAM,SAAS,GAAG,YAAY,CAAC;AAE/B,SAAS,YAAY;IACnB,IAAI,OAAO,OAAO,KAAK,WAAW,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,KAAK,EAAE,CAAC;QAC1D,OAAO,KAAK,CAAC;IACf,CAAC;IACD,MAAM,GAAG,GAAG,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC;IAC9B,IAAI,GAAG,KAAK,GAAG,EAAE,CAAC;QAChB,OAAO,IAAI,CAAC;IACd,CAAC;IACD,KAAK,MAAM,KAAK,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;QACnC,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;QAC7B,IACE,OAAO,KAAK,SAAS;YACrB,OAAO,KAAK,GAAG,SAAS,IAAI;YAC5B,OAAO,CAAC,UAAU,CAAC,GAAG,SAAS,GAAG,CAAC,EACnC,CAAC;YACD,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,oEAAoE;AACpE,MAAM,OAAO,GAAG,YAAY,EAAE,CAAC;AAE/B,MAAM,UAAU,KAAK,CAAC,OAAe;IACnC,IAAI,CAAC,OAAO,EAAE,CAAC;QACb,OAAO;IACT,CAAC;IACD,OAAO,CAAC,KAAK,CAAC,IAAI,SAAS,KAAK,OAAO,EAAE,CAAC,CAAC;AAC7C,CAAC"}
@@ -1,29 +1,7 @@
1
1
  export interface SpeechMetadata {
2
- /**
3
- * Duration of the generated audio in milliseconds.
4
- *
5
- * For `generateSpeech()`, this is computed from the audio bytes via
6
- * mediabunny, falling back to the provider-reported value if parsing
7
- * fails. For `streamSpeech()`, it is only the provider-reported value
8
- * (since the full audio isn't available until the stream is consumed),
9
- * and is undefined when the provider does not report it.
10
- */
11
2
  readonly audioDurationMs?: number;
12
- /** Number of characters in the input text (after audio tag processing). */
13
3
  readonly inputChars: number;
14
- /**
15
- * Time from request start to the response being ready, in milliseconds.
16
- *
17
- * For `generateSpeech()`, this is the full round-trip latency (request
18
- * sent → full response received). For `streamSpeech()`, the SDK returns
19
- * as soon as the stream response is ready, so this equals `ttfbMs`.
20
- */
21
4
  readonly latencyMs: number;
22
- /** Model identifier (e.g. "tts-1", "eleven_multilingual_v2"). */
23
- readonly model: string;
24
- /** Provider identifier (e.g. "openai", "elevenlabs"). */
25
- readonly provider: string;
26
- /** Time from request start to first byte received, in milliseconds. Only set for streaming. */
27
5
  readonly ttfbMs?: number;
28
6
  }
29
7
  //# sourceMappingURL=metadata.d.ts.map
@@ -1 +1 @@
1
- {"version":3,"file":"metadata.d.ts","sourceRoot":"","sources":["../src/metadata.ts"],"names":[],"mappings":"AAAA,MAAM,WAAW,cAAc;IAC7B;;;;;;;;OAQG;IACH,QAAQ,CAAC,eAAe,CAAC,EAAE,MAAM,CAAC;IAClC,2EAA2E;IAC3E,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAC;IAC5B;;;;;;OAMG;IACH,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;IAC3B,iEAAiE;IACjE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IACvB,yDAAyD;IACzD,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,+FAA+F;IAC/F,QAAQ,CAAC,MAAM,CAAC,EAAE,MAAM,CAAC;CAC1B"}
1
+ {"version":3,"file":"metadata.d.ts","sourceRoot":"","sources":["../src/metadata.ts"],"names":[],"mappings":"AAAA,MAAM,WAAW,cAAc;IAE7B,QAAQ,CAAC,eAAe,CAAC,EAAE,MAAM,CAAC;IAClC,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAC;IAE5B,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;IAE3B,QAAQ,CAAC,MAAM,CAAC,EAAE,MAAM,CAAC;CAC1B"}
@@ -0,0 +1,5 @@
1
+ import { SpeechSDKError } from "../errors.js";
2
+ export declare class DictionaryIdsRequireGatewayError extends SpeechSDKError {
3
+ constructor();
4
+ }
5
+ //# sourceMappingURL=errors.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"errors.d.ts","sourceRoot":"","sources":["../../src/pronunciations/errors.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,cAAc,EAAE,MAAM,cAAc,CAAC;AAE9C,qBAAa,gCAAiC,SAAQ,cAAc;;CAOnE"}
@@ -0,0 +1,8 @@
1
+ import { SpeechSDKError } from "../errors.js";
2
+ export class DictionaryIdsRequireGatewayError extends SpeechSDKError {
3
+ constructor() {
4
+ super("dictionaryIds require the gateway path. Use the gateway model string (e.g., 'openai/tts-1') or pass inline rules with { rules: [...] }.");
5
+ this.name = "DictionaryIdsRequireGatewayError";
6
+ }
7
+ }
8
+ //# sourceMappingURL=errors.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"errors.js","sourceRoot":"","sources":["../../src/pronunciations/errors.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,cAAc,EAAE,MAAM,cAAc,CAAC;AAE9C,MAAM,OAAO,gCAAiC,SAAQ,cAAc;IAClE;QACE,KAAK,CACH,yIAAyI,CAC1I,CAAC;QACF,IAAI,CAAC,IAAI,GAAG,kCAAkC,CAAC;IACjD,CAAC;CACF"}
@@ -0,0 +1,4 @@
1
+ import type { WordTimestamp } from "../timestamps.js";
2
+ import type { Edit } from "./types.js";
3
+ export declare function inverseAlign<T extends WordTimestamp>(timestamps: readonly T[], substitutedText: string, edits: readonly Edit[]): T[];
4
+ //# sourceMappingURL=inverse-align.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"inverse-align.d.ts","sourceRoot":"","sources":["../../src/pronunciations/inverse-align.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAC;AACtD,OAAO,KAAK,EAAE,IAAI,EAAE,MAAM,YAAY,CAAC;AAmBvC,wBAAgB,YAAY,CAAC,CAAC,SAAS,aAAa,EAClD,UAAU,EAAE,SAAS,CAAC,EAAE,EACxB,eAAe,EAAE,MAAM,EACvB,KAAK,EAAE,SAAS,IAAI,EAAE,GACrB,CAAC,EAAE,CAiDL"}
@@ -0,0 +1,54 @@
1
+ function findEditAt(position, edits) {
2
+ return edits.find((e) => position >= e.replacementRange[0] && position < e.replacementRange[1]);
3
+ }
4
+ function findTokenStart(haystack, searchFrom, token) {
5
+ return haystack.indexOf(token.toLowerCase(), searchFrom);
6
+ }
7
+ export function inverseAlign(timestamps, substitutedText, edits) {
8
+ if (edits.length === 0) {
9
+ return [...timestamps];
10
+ }
11
+ const out = [];
12
+ let cursor = 0;
13
+ let pendingGroup = null;
14
+ const haystack = substitutedText.toLowerCase();
15
+ const flushPending = () => {
16
+ if (!pendingGroup) {
17
+ return;
18
+ }
19
+ const merged = {
20
+ ...pendingGroup.first,
21
+ text: pendingGroup.edit.originalWord,
22
+ start: pendingGroup.first.start,
23
+ end: pendingGroup.last.end,
24
+ };
25
+ out.push(merged);
26
+ pendingGroup = null;
27
+ };
28
+ for (const ts of timestamps) {
29
+ const pos = findTokenStart(haystack, cursor, ts.text);
30
+ if (pos === -1) {
31
+ flushPending();
32
+ out.push(ts);
33
+ continue;
34
+ }
35
+ cursor = pos + ts.text.length;
36
+ const edit = findEditAt(pos, edits);
37
+ if (edit) {
38
+ if (pendingGroup && pendingGroup.edit === edit) {
39
+ pendingGroup.last = ts;
40
+ }
41
+ else {
42
+ flushPending();
43
+ pendingGroup = { edit, first: ts, last: ts };
44
+ }
45
+ }
46
+ else {
47
+ flushPending();
48
+ out.push(ts);
49
+ }
50
+ }
51
+ flushPending();
52
+ return out;
53
+ }
54
+ //# sourceMappingURL=inverse-align.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"inverse-align.js","sourceRoot":"","sources":["../../src/pronunciations/inverse-align.ts"],"names":[],"mappings":"AAGA,SAAS,UAAU,CACjB,QAAgB,EAChB,KAAsB;IAEtB,OAAO,KAAK,CAAC,IAAI,CACf,CAAC,CAAC,EAAE,EAAE,CAAC,QAAQ,IAAI,CAAC,CAAC,gBAAgB,CAAC,CAAC,CAAC,IAAI,QAAQ,GAAG,CAAC,CAAC,gBAAgB,CAAC,CAAC,CAAC,CAC7E,CAAC;AACJ,CAAC;AAED,SAAS,cAAc,CACrB,QAAgB,EAChB,UAAkB,EAClB,KAAa;IAEb,OAAO,QAAQ,CAAC,OAAO,CAAC,KAAK,CAAC,WAAW,EAAE,EAAE,UAAU,CAAC,CAAC;AAC3D,CAAC;AAED,MAAM,UAAU,YAAY,CAC1B,UAAwB,EACxB,eAAuB,EACvB,KAAsB;IAEtB,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,CAAC,GAAG,UAAU,CAAC,CAAC;IACzB,CAAC;IAED,MAAM,GAAG,GAAQ,EAAE,CAAC;IACpB,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,IAAI,YAAY,GAA6C,IAAI,CAAC;IAClE,MAAM,QAAQ,GAAG,eAAe,CAAC,WAAW,EAAE,CAAC;IAE/C,MAAM,YAAY,GAAG,GAAG,EAAE;QACxB,IAAI,CAAC,YAAY,EAAE,CAAC;YAClB,OAAO;QACT,CAAC;QACD,MAAM,MAAM,GAAG;YACb,GAAG,YAAY,CAAC,KAAK;YACrB,IAAI,EAAE,YAAY,CAAC,IAAI,CAAC,YAAY;YACpC,KAAK,EAAE,YAAY,CAAC,KAAK,CAAC,KAAK;YAC/B,GAAG,EAAE,YAAY,CAAC,IAAI,CAAC,GAAG;SACtB,CAAC;QACP,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACjB,YAAY,GAAG,IAAI,CAAC;IACtB,CAAC,CAAC;IAEF,KAAK,MAAM,EAAE,IAAI,UAAU,EAAE,CAAC;QAC5B,MAAM,GAAG,GAAG,cAAc,CAAC,QAAQ,EAAE,MAAM,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC;QACtD,IAAI,GAAG,KAAK,CAAC,CAAC,EAAE,CAAC;YACf,YAAY,EAAE,CAAC;YACf,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YACb,SAAS;QACX,CAAC;QACD,MAAM,GAAG,GAAG,GAAG,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC;QAC9B,MAAM,IAAI,GAAG,UAAU,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;QAEpC,IAAI,IAAI,EAAE,CAAC;YACT,IAAI,YAAY,IAAI,YAAY,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC;gBAC/C,YAAY,CAAC,IAAI,GAAG,EAAE,CAAC;YACzB,CAAC;iBAAM,CAAC;gBACN,YAAY,EAAE,CAAC;gBACf,YAAY,GAAG,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,IAAI,EAAE,EAAE,EAAE,CAAC;YAC/C,CAAC;QACH,CAAC;aAAM,CAAC;YACN,YAAY,EAAE,CAAC;YACf,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACf,CAAC;IACH,CAAC;IAED,YAAY,EAAE,CAAC;IACf,OAAO,GAAG,CAAC;AACb,CAAC"}
@@ -0,0 +1,4 @@
1
+ import type { Pronunciation } from "./types.js";
2
+ export declare function ruleMapKey(word: string, caseSensitive: boolean): string;
3
+ export declare function mergeRules(rules: readonly Pronunciation[]): Map<string, Pronunciation>;
4
+ //# sourceMappingURL=merge.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"merge.d.ts","sourceRoot":"","sources":["../../src/pronunciations/merge.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAGhD,wBAAgB,UAAU,CAAC,IAAI,EAAE,MAAM,EAAE,aAAa,EAAE,OAAO,GAAG,MAAM,CAEvE;AAED,wBAAgB,UAAU,CACxB,KAAK,EAAE,SAAS,aAAa,EAAE,GAC9B,GAAG,CAAC,MAAM,EAAE,aAAa,CAAC,CAO5B"}