zardbot-telegram 1.0.0

package/dist/stt/client.js

@@ -0,0 +1,97 @@
+import { config } from "../config.js";
+import { logger } from "../utils/logger.js";
+import { execSync } from "node:child_process";
+import { tmpdir } from "node:os";
+import { join } from "node:path";
+import { unlinkSync, writeFileSync, readFileSync } from "node:fs";
+const STT_REQUEST_TIMEOUT_MS = 60_000;
+function convertToWav(audioBuffer, filename) {
+    const ext = filename.split(".").pop()?.toLowerCase();
+    if (ext === "wav" || ext === "wave") {
+        return audioBuffer;
+    }
+    const inputPath = join(tmpdir(), `stt_input_${Date.now()}.${ext}`);
+    const outputPath = join(tmpdir(), `stt_output_${Date.now()}.wav`);
+    try {
+        writeFileSync(inputPath, audioBuffer);
+        execSync(`ffmpeg -y -i "${inputPath}" -ar 16000 -ac 1 -f wav "${outputPath}"`, {
+            stdio: "pipe",
+        });
+        return readFileSync(outputPath);
+    }
+    finally {
+        try {
+            unlinkSync(inputPath);
+        }
+        catch { }
+        try {
+            unlinkSync(outputPath);
+        }
+        catch { }
+    }
+}
+/**
+ * Returns true if STT is configured (API URL and API key are set).
+ */
+export function isSttConfigured() {
+    return Boolean(config.stt.apiUrl && config.stt.apiKey);
+}
+/**
+ * Transcribes an audio buffer using a Whisper-compatible API (OpenAI / Groq / etc.).
+ *
+ * Sends a multipart/form-data POST to `{STT_API_URL}/audio/transcriptions`.
+ *
+ * @param audioBuffer - Raw audio file bytes (ogg, mp3, wav, m4a, webm, etc.)
+ * @param filename    - Original filename with extension (used by the API to detect format)
+ * @returns Transcribed text
+ * @throws Error if STT is not configured, the request fails, or the response is invalid
+ */
+export async function transcribeAudio(audioBufferInput, filename) {
+    if (!isSttConfigured()) {
+        throw new Error("STT is not configured: STT_API_URL and STT_API_KEY are required");
+    }
+    const url = `${config.stt.apiUrl}/audio/transcriptions`;
+    const audioBuffer = convertToWav(audioBufferInput, filename);
+    const wavFilename = filename.replace(/\.[^.]+$/, ".wav");
+    const formData = new FormData();
+    formData.append("file", new Blob([new Uint8Array(audioBuffer)]), wavFilename);
+    if (config.stt.model) {
+        formData.append("model", config.stt.model);
+    }
+    formData.append("response_format", "json");
+    if (config.stt.language) {
+        formData.append("language", config.stt.language);
+    }
+    logger.debug(`[STT] Sending transcription request: url=${url}, model=${config.stt.model}, filename=${filename}, size=${audioBuffer.length} bytes`);
+    const controller = new AbortController();
+    const timeout = setTimeout(() => controller.abort(), STT_REQUEST_TIMEOUT_MS);
+    try {
+        const response = await fetch(url, {
+            method: "POST",
+            headers: {
+                Authorization: `Bearer ${config.stt.apiKey}`,
+            },
+            body: formData,
+            signal: controller.signal,
+        });
+        if (!response.ok) {
+            const errorBody = await response.text().catch(() => "");
+            throw new Error(`STT API returned HTTP ${response.status}: ${errorBody || response.statusText}`);
+        }
+        const data = (await response.json());
+        if (typeof data.text !== "string") {
+            throw new Error("STT API response does not contain a text field");
+        }
+        logger.debug(`[STT] Transcription result: ${data.text.length} chars`);
+        return { text: data.text };
+    }
+    catch (err) {
+        if (err instanceof DOMException && err.name === "AbortError") {
+            throw new Error(`STT request timed out after ${STT_REQUEST_TIMEOUT_MS}ms`);
+        }
+        throw err;
+    }
+    finally {
+        clearTimeout(timeout);
+    }
+}