fix(azure-speech): bound generated speech downloads · openclaw/openclaw@e098fd4

Recent Commits to openclaw:main

vincentkoc · 2026-05-29 · via Recent Commits to openclaw:main

Original file line number	Diff line number	Diff line change
`@@ -185,6 +185,7 @@ describe("buildAzureSpeechProvider", () => {`
`185`	`185`	`lang: "en-US",`
`186`	`186`	`outputFormat: "ogg-24khz-16bit-mono-opus",`
`187`	`187`	`timeoutMs: 30_000,`
	`188`	`+maxBytes: 16 * 1024 * 1024,`
`188`	`189`	`});`
`189`	`190`	`expect(result).toEqual({`
`190`	`191`	`audioBuffer: Buffer.from("audio-bytes"),`
`@@ -222,6 +223,7 @@ describe("buildAzureSpeechProvider", () => {`
`222`	`223`	`lang: "es-US",`
`223`	`224`	`outputFormat: "raw-8khz-8bit-mono-mulaw",`
`224`	`225`	`timeoutMs: 30_000,`
	`226`	`+maxBytes: 16 * 1024 * 1024,`
`225`	`227`	`});`
`226`	`228`	`expect(result).toEqual({`
`227`	`229`	`audioBuffer: Buffer.from("audio-bytes"),`
`@@ -230,6 +232,34 @@ describe("buildAzureSpeechProvider", () => {`
`230`	`232`	`});`
`231`	`233`	`});`
`232`	`234`
	`235`	`+it("applies the configured media byte cap to synthesis requests", async () => {`
	`236`	`+const provider = buildAzureSpeechProvider();`
	`237`	`+`
	`238`	`+await provider.synthesize({`
	`239`	`+text: "hello",`
	`240`	`+cfg: {`
	`241`	`+agents: {`
	`242`	`+defaults: {`
	`243`	`+mediaMaxMb: 2,`
	`244`	`+},`
	`245`	`+},`
	`246`	`+} as never,`
	`247`	`+providerConfig: {`
	`248`	`+apiKey: "key",`
	`249`	`+region: "eastus",`
	`250`	`+voice: "en-US-JennyNeural",`
	`251`	`+},`
	`252`	`+target: "audio-file",`
	`253`	`+timeoutMs: 30_000,`
	`254`	`+});`
	`255`	`+`
	`256`	`+expect(azureSpeechTTSMock).toHaveBeenCalledWith(`
	`257`	`+expect.objectContaining({`
	`258`	`+maxBytes: 2 * 1024 * 1024,`
	`259`	`+}),`
	`260`	`+);`
	`261`	`+});`
	`262`	`+`
`233`	`263`	`it("lists voices through config or explicit request auth", async () => {`
`234`	`264`	`const provider = buildAzureSpeechProvider();`
`235`	`265`	`const voices = await provider.listVoices?.({`

Original file line number	Diff line number	Diff line change
`@@ -19,6 +19,8 @@ import {`
`19`	`19`	`normalizeAzureSpeechBaseUrl,`
`20`	`20`	`} from "./tts.js";`
`21`	`21`
	`22`	`+const DEFAULT_GENERATED_AUDIO_MAX_BYTES = 16 * 1024 * 1024;`
	`23`	`+`
`22`	`24`	`type AzureSpeechProviderConfig = {`
`23`	`25`	`apiKey?: string;`
`24`	`26`	`region?: string;`
`@@ -177,6 +179,16 @@ function resolveTimeoutMs(config: AzureSpeechProviderConfig, timeoutMs: number):`
`177`	`179`	`return config.timeoutMs ?? timeoutMs;`
`178`	`180`	`}`
`179`	`181`
	`182`	`+function resolveGeneratedAudioMaxBytes(req: {`
	`183`	`+cfg: { agents?: { defaults?: { mediaMaxMb?: number } } };`
	`184`	`+}): number {`
	`185`	`+const configured = req.cfg.agents?.defaults?.mediaMaxMb;`
	`186`	`+if (typeof configured === "number" && Number.isFinite(configured) && configured > 0) {`
	`187`	`+return Math.floor(configured * 1024 * 1024);`
	`188`	`+}`
	`189`	`+return DEFAULT_GENERATED_AUDIO_MAX_BYTES;`
	`190`	`+}`
	`191`	`+`
`180`	`192`	`export function buildAzureSpeechProvider(): SpeechProviderPlugin {`
`181`	`193`	`return {`
`182`	`194`	`id: "azure-speech",`
`@@ -269,6 +281,7 @@ export function buildAzureSpeechProvider(): SpeechProviderPlugin {`
`269`	`281`	`lang: overrides.lang ?? config.lang,`
`270`	`282`	`outputFormat,`
`271`	`283`	`timeoutMs: resolveTimeoutMs(config, req.timeoutMs),`
	`284`	`+maxBytes: resolveGeneratedAudioMaxBytes(req),`
`272`	`285`	`});`
`273`	`286`	`return {`
`274`	`287`	`audioBuffer,`
`@@ -295,6 +308,7 @@ export function buildAzureSpeechProvider(): SpeechProviderPlugin {`
`295`	`308`	`lang: overrides.lang ?? config.lang,`
`296`	`309`	`outputFormat: DEFAULT_AZURE_SPEECH_TELEPHONY_FORMAT,`
`297`	`310`	`timeoutMs: resolveTimeoutMs(config, req.timeoutMs),`
	`311`	`+maxBytes: resolveGeneratedAudioMaxBytes(req),`
`298`	`312`	`});`
`299`	`313`	`return {`
`300`	`314`	`audioBuffer,`

Original file line number	Diff line number	Diff line change
`@@ -12,6 +12,31 @@ import {`
`12`	`12`	`describe("azure speech tts", () => {`
`13`	`13`	`installPinnedHostnameTestHooks();`
`14`	`14`
	`15`	`+function createStreamingAudioResponse(params: {`
	`16`	`+chunkCount: number;`
	`17`	`+chunkSize: number;`
	`18`	`+byte: number;`
	`19`	`+}): { response: Response; getReadCount: () => number } {`
	`20`	`+let reads = 0;`
	`21`	`+const stream = new ReadableStream<Uint8Array>({`
	`22`	`+pull(controller) {`
	`23`	`+if (reads >= params.chunkCount) {`
	`24`	`+controller.close();`
	`25`	`+return;`
	`26`	`+}`
	`27`	`+reads += 1;`
	`28`	`+controller.enqueue(new Uint8Array(params.chunkSize).fill(params.byte));`
	`29`	`+},`
	`30`	`+});`
	`31`	`+return {`
	`32`	`+response: new Response(stream, {`
	`33`	`+status: 200,`
	`34`	`+headers: { "Content-Type": "audio/mpeg" },`
	`35`	`+}),`
	`36`	`+getReadCount: () => reads,`
	`37`	`+};`
	`38`	`+}`
	`39`	`+`
`15`	`40`	`afterEach(() => {`
`16`	`41`	`vi.unstubAllGlobals();`
`17`	`42`	`vi.restoreAllMocks();`
`@@ -82,6 +107,30 @@ describe("azure speech tts", () => {`
`82`	`107`	`expect(init.signal).toBeInstanceOf(AbortSignal);`
`83`	`108`	`});`
`84`	`109`
	`110`	`+it("caps streamed audio responses instead of buffering oversized TTS output", async () => {`
	`111`	`+const streamed = createStreamingAudioResponse({`
	`112`	`+chunkCount: 20,`
	`113`	`+chunkSize: 1024,`
	`114`	`+byte: 121,`
	`115`	`+});`
	`116`	`+vi.stubGlobal("fetch", vi.fn().mockResolvedValue(streamed.response));`
	`117`	`+`
	`118`	`+await expect(`
	`119`	`+azureSpeechTTS({`
	`120`	`+text: "hello",`
	`121`	`+apiKey: "speech-key",`
	`122`	`+region: "eastus",`
	`123`	`+voice: "en-US-JennyNeural",`
	`124`	`+lang: "en-US",`
	`125`	`+outputFormat: "audio-24khz-48kbitrate-mono-mp3",`
	`126`	`+timeoutMs: 1234,`
	`127`	`+maxBytes: 2048,`
	`128`	`+}),`
	`129`	`+).rejects.toThrow("Azure Speech TTS audio response exceeds 2048 bytes");`
	`130`	`+`
	`131`	`+expect(streamed.getReadCount()).toBeLessThan(20);`
	`132`	`+});`
	`133`	`+`
`85`	`134`	`it("lists voices with timeout and filters deprecated entries", async () => {`
`86`	`135`	`const fetchMock = vi.fn().mockResolvedValue(`
`87`	`136`	`new Response(`

Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,5 @@`
`1`	`1`	`import { assertOkOrThrowProviderError } from "openclaw/plugin-sdk/provider-http";`
	`2`	`+import { readResponseWithLimit } from "openclaw/plugin-sdk/response-limit-runtime";`
`2`	`3`	`import type { SpeechVoiceOption } from "openclaw/plugin-sdk/speech-core";`
`3`	`4`	`import { trimToUndefined } from "openclaw/plugin-sdk/speech-core";`
`4`	`5`	`import {`
`@@ -11,6 +12,7 @@ export const DEFAULT_AZURE_SPEECH_LANG = "en-US";`
`11`	`12`	`export const DEFAULT_AZURE_SPEECH_AUDIO_FORMAT = "audio-24khz-48kbitrate-mono-mp3";`
`12`	`13`	`export const DEFAULT_AZURE_SPEECH_VOICE_NOTE_FORMAT = "ogg-24khz-16bit-mono-opus";`
`13`	`14`	`export const DEFAULT_AZURE_SPEECH_TELEPHONY_FORMAT = "raw-8khz-8bit-mono-mulaw";`
	`15`	`+const DEFAULT_AZURE_SPEECH_MAX_BYTES = 16 * 1024 * 1024;`
`14`	`16`
`15`	`17`	`type AzureSpeechVoiceEntry = {`
`16`	`18`	`ShortName?: string;`
`@@ -175,6 +177,7 @@ export async function azureSpeechTTS(params: {`
`175`	`177`	`lang?: string;`
`176`	`178`	`outputFormat?: string;`
`177`	`179`	`timeoutMs?: number;`
	`180`	`+maxBytes?: number;`
`178`	`181`	`}): Promise<Buffer> {`
`179`	`182`	`const voice = trimToUndefined(params.voice) ?? DEFAULT_AZURE_SPEECH_VOICE;`
`180`	`183`	`const outputFormat = trimToUndefined(params.outputFormat) ?? DEFAULT_AZURE_SPEECH_AUDIO_FORMAT;`
`@@ -202,7 +205,14 @@ export async function azureSpeechTTS(params: {`
`202`	`205`
`203`	`206`	`try {`
`204`	`207`	`await assertOkOrThrowProviderError(response, "Azure Speech TTS API error");`
`205`		`-return Buffer.from(await response.arrayBuffer());`
	`208`	`+return await readResponseWithLimit(`
	`209`	`+response,`
	`210`	`+params.maxBytes ?? DEFAULT_AZURE_SPEECH_MAX_BYTES,`
	`211`	`+{`
	`212`	`+onOverflow: ({ maxBytes }) =>`
	`213`	+new Error(`Azure Speech TTS audio response exceeds ${maxBytes} bytes`),
	`214`	`+},`
	`215`	`+);`
`206`	`216`	`} finally {`
`207`	`217`	`await release();`
`208`	`218`	`}`

此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。原文来自 — 版权归原作者所有。

推荐订阅源

Recent Commits to openclaw:main