wind-agent-cli 1.37.2 → 1.37.3

package/server/modules/voice/voice.routes.ts

@@ -1,113 +1,113 @@
-import { Readable } from 'node:stream';
-
-import express from 'express';
-
-import type { VoiceRequestOverrides, VoiceService, VoiceServiceResult } from '@/shared/types.js';
-import { asyncHandler } from '@/shared/utils.js';
-
-type VoiceRouterDependencies = {
-  voiceService: VoiceService;
-  parseAudioUpload: express.RequestHandler;
-};
-
-function readHeaderValue(value: string | string[] | undefined): string | undefined {
-  const normalizedValue = Array.isArray(value) ? value[0] : value;
-  const trimmedValue = normalizedValue?.trim();
-  return trimmedValue || undefined;
-}
-
-function parseVoiceOverrides(request: express.Request): VoiceRequestOverrides {
-  return {
-    apiKey: readHeaderValue(request.headers['x-voice-api-key']),
-    sttModel: readHeaderValue(request.headers['x-voice-stt-model']),
-    ttsModel: readHeaderValue(request.headers['x-voice-tts-model']),
-    ttsVoice: readHeaderValue(request.headers['x-voice-tts-voice']),
-    ttsFormat: readHeaderValue(request.headers['x-voice-tts-format']),
-  };
-}
-
-function sendFailure<TValue>(
-  response: express.Response,
-  result: VoiceServiceResult<TValue>,
-): result is Extract<VoiceServiceResult<TValue>, { ok: false }> {
-  if (result.ok) {
-    return false;
-  }
-
-  response.status(result.status).json({ error: result.error });
-  return true;
-}
-
-/**
- * Creates the transport-only router used by the Voice composition root. It is
- * exported for Voice route tests; other modules consume only the composed
- * router exposed from the Voice barrel.
- */
-export function createVoiceRouter(dependencies: VoiceRouterDependencies): express.Router {
-  const router = express.Router();
-
-  router.get('/health', (_request, response) => {
-    response.json(dependencies.voiceService.getHealth());
-  });
-
-  router.post('/transcribe', (request, response, next) => {
-    dependencies.parseAudioUpload(request, response, (uploadError?: unknown) => {
-      if (uploadError) {
-        const message = uploadError instanceof Error ? uploadError.message : String(uploadError);
-        response.status(400).json({ error: message });
-        return;
-      }
-
-      // Multer uses a callback API, so bridge its parsed request into the async
-      // service call and forward unexpected rejections to Express middleware.
-      void (async () => {
-        if (!request.file) {
-          response.status(400).json({ error: 'No audio uploaded' });
-          return;
-        }
-
-        const result = await dependencies.voiceService.transcribe({
-          audio: {
-            bytes: request.file.buffer,
-            mimeType: request.file.mimetype || 'audio/webm',
-            fileName: request.file.originalname || 'recording.webm',
-          },
-          overrides: parseVoiceOverrides(request),
-        });
-
-        if (sendFailure(response, result)) {
-          return;
-        }
-
-        response.json(result.value);
-      })().catch(next);
-    });
-  });
-
-  router.post('/tts', asyncHandler(async (request, response) => {
-    const text = request.body?.text;
-    if (typeof text !== 'string' || !text.trim()) {
-      response.status(400).json({ error: 'text required' });
-      return;
-    }
-
-    const result = await dependencies.voiceService.synthesizeSpeech({
-      text,
-      overrides: parseVoiceOverrides(request),
-    });
-    if (sendFailure(response, result)) {
-      return;
-    }
-
-    response.setHeader('Content-Type', result.value.contentType);
-    response.setHeader('Cache-Control', 'no-store');
-    if (!result.value.body) {
-      response.end();
-      return;
-    }
-
-    Readable.fromWeb(result.value.body).on('error', (error) => response.destroy(error)).pipe(response);
-  }));
-
-  return router;
-}
+import { Readable } from 'node:stream';
+
+import express from 'express';
+
+import type { VoiceRequestOverrides, VoiceService, VoiceServiceResult } from '@/shared/types.js';
+import { asyncHandler } from '@/shared/utils.js';
+
+type VoiceRouterDependencies = {
+  voiceService: VoiceService;
+  parseAudioUpload: express.RequestHandler;
+};
+
+function readHeaderValue(value: string | string[] | undefined): string | undefined {
+  const normalizedValue = Array.isArray(value) ? value[0] : value;
+  const trimmedValue = normalizedValue?.trim();
+  return trimmedValue || undefined;
+}
+
+function parseVoiceOverrides(request: express.Request): VoiceRequestOverrides {
+  return {
+    apiKey: readHeaderValue(request.headers['x-voice-api-key']),
+    sttModel: readHeaderValue(request.headers['x-voice-stt-model']),
+    ttsModel: readHeaderValue(request.headers['x-voice-tts-model']),
+    ttsVoice: readHeaderValue(request.headers['x-voice-tts-voice']),
+    ttsFormat: readHeaderValue(request.headers['x-voice-tts-format']),
+  };
+}
+
+function sendFailure<TValue>(
+  response: express.Response,
+  result: VoiceServiceResult<TValue>,
+): result is Extract<VoiceServiceResult<TValue>, { ok: false }> {
+  if (result.ok) {
+    return false;
+  }
+
+  response.status(result.status).json({ error: result.error });
+  return true;
+}
+
+/**
+ * Creates the transport-only router used by the Voice composition root. It is
+ * exported for Voice route tests; other modules consume only the composed
+ * router exposed from the Voice barrel.
+ */
+export function createVoiceRouter(dependencies: VoiceRouterDependencies): express.Router {
+  const router = express.Router();
+
+  router.get('/health', (_request, response) => {
+    response.json(dependencies.voiceService.getHealth());
+  });
+
+  router.post('/transcribe', (request, response, next) => {
+    dependencies.parseAudioUpload(request, response, (uploadError?: unknown) => {
+      if (uploadError) {
+        const message = uploadError instanceof Error ? uploadError.message : String(uploadError);
+        response.status(400).json({ error: message });
+        return;
+      }
+
+      // Multer uses a callback API, so bridge its parsed request into the async
+      // service call and forward unexpected rejections to Express middleware.
+      void (async () => {
+        if (!request.file) {
+          response.status(400).json({ error: 'No audio uploaded' });
+          return;
+        }
+
+        const result = await dependencies.voiceService.transcribe({
+          audio: {
+            bytes: request.file.buffer,
+            mimeType: request.file.mimetype || 'audio/webm',
+            fileName: request.file.originalname || 'recording.webm',
+          },
+          overrides: parseVoiceOverrides(request),
+        });
+
+        if (sendFailure(response, result)) {
+          return;
+        }
+
+        response.json(result.value);
+      })().catch(next);
+    });
+  });
+
+  router.post('/tts', asyncHandler(async (request, response) => {
+    const text = request.body?.text;
+    if (typeof text !== 'string' || !text.trim()) {
+      response.status(400).json({ error: 'text required' });
+      return;
+    }
+
+    const result = await dependencies.voiceService.synthesizeSpeech({
+      text,
+      overrides: parseVoiceOverrides(request),
+    });
+    if (sendFailure(response, result)) {
+      return;
+    }
+
+    response.setHeader('Content-Type', result.value.contentType);
+    response.setHeader('Cache-Control', 'no-store');
+    if (!result.value.body) {
+      response.end();
+      return;
+    }
+
+    Readable.fromWeb(result.value.body).on('error', (error) => response.destroy(error)).pipe(response);
+  }));
+
+  return router;
+}

package/server/modules/voice/voice.service.ts

@@ -1,192 +1,192 @@
-import type {
-  VoiceAudioUpload,
-  VoiceRequestOverrides,
-  VoiceService,
-  VoiceServiceResult,
-  VoiceSpeechPayload,
-} from '@/shared/types.js';
-
-type VoiceServiceDependencies = {
-  defaults: {
-    baseUrl: string;
-    apiKey: string;
-    sttModel: string;
-    ttsModel: string;
-    ttsVoice: string;
-  };
-  timeoutMs: number;
-  fetchBackend(url: string, options: RequestInit): Promise<Response>;
-};
-
-type ResolvedVoiceConfig = VoiceServiceDependencies['defaults'] & {
-  ttsFormat: string;
-};
-
-function resolveVoiceConfig(
-  defaults: VoiceServiceDependencies['defaults'],
-  overrides: VoiceRequestOverrides,
-): ResolvedVoiceConfig {
-  return {
-    baseUrl: defaults.baseUrl,
-    apiKey: overrides.apiKey || defaults.apiKey,
-    sttModel: overrides.sttModel || defaults.sttModel,
-    ttsModel: overrides.ttsModel || defaults.ttsModel,
-    ttsVoice: overrides.ttsVoice || defaults.ttsVoice,
-    ttsFormat: overrides.ttsFormat?.trim() ?? '',
-  };
-}
-
-function validateBackendBaseUrl(baseUrl: string): boolean {
-  try {
-    const parsedUrl = new URL(baseUrl);
-    if (parsedUrl.protocol !== 'http:' && parsedUrl.protocol !== 'https:') {
-      return false;
-    }
-
-    // Local and private backends are supported intentionally. Only link-local
-    // metadata addresses remain blocked as a defense in depth measure.
-    return parsedUrl.hostname !== '169.254.169.254'
-      && !parsedUrl.hostname.startsWith('169.254.');
-  } catch {
-    return false;
-  }
-}
-
-function authorizationHeader(apiKey: string): Record<string, string> {
-  return apiKey ? { Authorization: `Bearer ${apiKey}` } : {};
-}
-
-function backendFailure(status: number, responseText?: string): VoiceServiceResult<never> {
-  if (status === 401 || status === 403) {
-    return {
-      ok: false,
-      status: 502,
-      error: 'Voice backend rejected the request (check the API key).',
-    };
-  }
-
-  return {
-    ok: false,
-    status,
-    error: responseText || 'voice backend error',
-  };
-}
-
-function unreachableBackendFailure(error: unknown, timeoutMs: number): VoiceServiceResult<never> {
-  if (error instanceof Error && error.name === 'AbortError') {
-    return {
-      ok: false,
-      status: 504,
-      error: `Voice backend timed out after ${Math.round(timeoutMs / 1000)}s. Check your voice backend.`,
-    };
-  }
-
-  const message = error instanceof Error ? error.message : String(error);
-  return {
-    ok: false,
-    status: 502,
-    error: `Voice backend unreachable: ${message}`,
-  };
-}
-
-function validateConfiguredBackend(config: ResolvedVoiceConfig): VoiceServiceResult<never> | null {
-  if (!config.baseUrl) {
-    return { ok: false, status: 503, error: 'No voice backend configured' };
-  }
-
-  if (!validateBackendBaseUrl(config.baseUrl)) {
-    return { ok: false, status: 400, error: 'Invalid voice backend URL.' };
-  }
-
-  return null;
-}
-
-function createTranscriptionFormData(audio: VoiceAudioUpload, sttModel: string): FormData {
-  const formData = new FormData();
-  formData.append('file', new Blob([audio.bytes], { type: audio.mimeType }), audio.fileName);
-  formData.append('model', sttModel);
-  return formData;
-}
-
-/**
- * Creates the Voice application service used by the Voice composition root and
- * its unit tests. The outbound request function and server configuration are
- * required so the service never reads globals or creates production defaults.
- */
-export function createVoiceService(dependencies: VoiceServiceDependencies): VoiceService {
-  return {
-    getHealth: () => ({ configured: Boolean(dependencies.defaults.baseUrl) }),
-
-    async transcribe(input) {
-      const config = resolveVoiceConfig(dependencies.defaults, input.overrides);
-      const configurationFailure = validateConfiguredBackend(config);
-      if (configurationFailure) {
-        return configurationFailure;
-      }
-
-      try {
-        const response = await dependencies.fetchBackend(
-          `${config.baseUrl}/audio/transcriptions`,
-          {
-            method: 'POST',
-            headers: authorizationHeader(config.apiKey),
-            body: createTranscriptionFormData(input.audio, config.sttModel),
-          },
-        );
-        const responseText = await response.text();
-        if (!response.ok) {
-          return backendFailure(response.status, responseText);
-        }
-
-        try {
-          const parsed = JSON.parse(responseText) as { text?: unknown };
-          return {
-            ok: true,
-            value: { text: typeof parsed.text === 'string' ? parsed.text : '' },
-          };
-        } catch {
-          return { ok: true, value: { text: responseText } };
-        }
-      } catch (error) {
-        return unreachableBackendFailure(error, dependencies.timeoutMs);
-      }
-    },
-
-    async synthesizeSpeech(input) {
-      const config = resolveVoiceConfig(dependencies.defaults, input.overrides);
-      const configurationFailure = validateConfiguredBackend(config);
-      if (configurationFailure) {
-        return configurationFailure;
-      }
-
-      try {
-        const response = await dependencies.fetchBackend(`${config.baseUrl}/audio/speech`, {
-          method: 'POST',
-          headers: {
-            'Content-Type': 'application/json',
-            ...authorizationHeader(config.apiKey),
-          },
-          body: JSON.stringify({
-            model: config.ttsModel,
-            voice: config.ttsVoice,
-            input: input.text,
-            ...(config.ttsFormat ? { response_format: config.ttsFormat } : {}),
-          }),
-        });
-
-        if (!response.ok) {
-          const responseText = await response.text().catch(() => 'tts failed');
-          return backendFailure(response.status, responseText);
-        }
-
-        const value: VoiceSpeechPayload = {
-          contentType: response.headers.get('content-type') || 'audio/mpeg',
-          body: response.body,
-        };
-        return { ok: true, value };
-      } catch (error) {
-        return unreachableBackendFailure(error, dependencies.timeoutMs);
-      }
-    },
-  };
-}
+import type {
+  VoiceAudioUpload,
+  VoiceRequestOverrides,
+  VoiceService,
+  VoiceServiceResult,
+  VoiceSpeechPayload,
+} from '@/shared/types.js';
+
+type VoiceServiceDependencies = {
+  defaults: {
+    baseUrl: string;
+    apiKey: string;
+    sttModel: string;
+    ttsModel: string;
+    ttsVoice: string;
+  };
+  timeoutMs: number;
+  fetchBackend(url: string, options: RequestInit): Promise<Response>;
+};
+
+type ResolvedVoiceConfig = VoiceServiceDependencies['defaults'] & {
+  ttsFormat: string;
+};
+
+function resolveVoiceConfig(
+  defaults: VoiceServiceDependencies['defaults'],
+  overrides: VoiceRequestOverrides,
+): ResolvedVoiceConfig {
+  return {
+    baseUrl: defaults.baseUrl,
+    apiKey: overrides.apiKey || defaults.apiKey,
+    sttModel: overrides.sttModel || defaults.sttModel,
+    ttsModel: overrides.ttsModel || defaults.ttsModel,
+    ttsVoice: overrides.ttsVoice || defaults.ttsVoice,
+    ttsFormat: overrides.ttsFormat?.trim() ?? '',
+  };
+}
+
+function validateBackendBaseUrl(baseUrl: string): boolean {
+  try {
+    const parsedUrl = new URL(baseUrl);
+    if (parsedUrl.protocol !== 'http:' && parsedUrl.protocol !== 'https:') {
+      return false;
+    }
+
+    // Local and private backends are supported intentionally. Only link-local
+    // metadata addresses remain blocked as a defense in depth measure.
+    return parsedUrl.hostname !== '169.254.169.254'
+      && !parsedUrl.hostname.startsWith('169.254.');
+  } catch {
+    return false;
+  }
+}
+
+function authorizationHeader(apiKey: string): Record<string, string> {
+  return apiKey ? { Authorization: `Bearer ${apiKey}` } : {};
+}
+
+function backendFailure(status: number, responseText?: string): VoiceServiceResult<never> {
+  if (status === 401 || status === 403) {
+    return {
+      ok: false,
+      status: 502,
+      error: 'Voice backend rejected the request (check the API key).',
+    };
+  }
+
+  return {
+    ok: false,
+    status,
+    error: responseText || 'voice backend error',
+  };
+}
+
+function unreachableBackendFailure(error: unknown, timeoutMs: number): VoiceServiceResult<never> {
+  if (error instanceof Error && error.name === 'AbortError') {
+    return {
+      ok: false,
+      status: 504,
+      error: `Voice backend timed out after ${Math.round(timeoutMs / 1000)}s. Check your voice backend.`,
+    };
+  }
+
+  const message = error instanceof Error ? error.message : String(error);
+  return {
+    ok: false,
+    status: 502,
+    error: `Voice backend unreachable: ${message}`,
+  };
+}
+
+function validateConfiguredBackend(config: ResolvedVoiceConfig): VoiceServiceResult<never> | null {
+  if (!config.baseUrl) {
+    return { ok: false, status: 503, error: 'No voice backend configured' };
+  }
+
+  if (!validateBackendBaseUrl(config.baseUrl)) {
+    return { ok: false, status: 400, error: 'Invalid voice backend URL.' };
+  }
+
+  return null;
+}
+
+function createTranscriptionFormData(audio: VoiceAudioUpload, sttModel: string): FormData {
+  const formData = new FormData();
+  formData.append('file', new Blob([audio.bytes], { type: audio.mimeType }), audio.fileName);
+  formData.append('model', sttModel);
+  return formData;
+}
+
+/**
+ * Creates the Voice application service used by the Voice composition root and
+ * its unit tests. The outbound request function and server configuration are
+ * required so the service never reads globals or creates production defaults.
+ */
+export function createVoiceService(dependencies: VoiceServiceDependencies): VoiceService {
+  return {
+    getHealth: () => ({ configured: Boolean(dependencies.defaults.baseUrl) }),
+
+    async transcribe(input) {
+      const config = resolveVoiceConfig(dependencies.defaults, input.overrides);
+      const configurationFailure = validateConfiguredBackend(config);
+      if (configurationFailure) {
+        return configurationFailure;
+      }
+
+      try {
+        const response = await dependencies.fetchBackend(
+          `${config.baseUrl}/audio/transcriptions`,
+          {
+            method: 'POST',
+            headers: authorizationHeader(config.apiKey),
+            body: createTranscriptionFormData(input.audio, config.sttModel),
+          },
+        );
+        const responseText = await response.text();
+        if (!response.ok) {
+          return backendFailure(response.status, responseText);
+        }
+
+        try {
+          const parsed = JSON.parse(responseText) as { text?: unknown };
+          return {
+            ok: true,
+            value: { text: typeof parsed.text === 'string' ? parsed.text : '' },
+          };
+        } catch {
+          return { ok: true, value: { text: responseText } };
+        }
+      } catch (error) {
+        return unreachableBackendFailure(error, dependencies.timeoutMs);
+      }
+    },
+
+    async synthesizeSpeech(input) {
+      const config = resolveVoiceConfig(dependencies.defaults, input.overrides);
+      const configurationFailure = validateConfiguredBackend(config);
+      if (configurationFailure) {
+        return configurationFailure;
+      }
+
+      try {
+        const response = await dependencies.fetchBackend(`${config.baseUrl}/audio/speech`, {
+          method: 'POST',
+          headers: {
+            'Content-Type': 'application/json',
+            ...authorizationHeader(config.apiKey),
+          },
+          body: JSON.stringify({
+            model: config.ttsModel,
+            voice: config.ttsVoice,
+            input: input.text,
+            ...(config.ttsFormat ? { response_format: config.ttsFormat } : {}),
+          }),
+        });
+
+        if (!response.ok) {
+          const responseText = await response.text().catch(() => 'tts failed');
+          return backendFailure(response.status, responseText);
+        }
+
+        const value: VoiceSpeechPayload = {
+          contentType: response.headers.get('content-type') || 'audio/mpeg',
+          body: response.body,
+        };
+        return { ok: true, value };
+      } catch (error) {
+        return unreachableBackendFailure(error, dependencies.timeoutMs);
+      }
+    },
+  };
+}