@pr0gramm/fluester - npm Package Compare versions

Comparing version 0.3.9 to 0.3.10

dist/index.d.ts

 import { ModelName } from "./model.js";
-import { TranscriptLine } from "./transcript.js";
-export interface WhisperClientOptions {
+import { LanguageDetectionResult, TranscriptLine } from "./transcript.js";
+export interface WhisperClientOptionsBase {
     /** Path to the whisper executable */
     executablePath?: string;
 }
-export interface WhisperOptionsBase {
-    whisperOptions?: FlagTypes;
-}
-export interface WhisperOptionsWithModelPath extends WhisperOptionsBase {
-    modelPath: string;
-}
-export interface WhisperOptionsWithModelName extends WhisperOptionsBase {
+export interface WhisperClientOptionsWithModelName extends WhisperClientOptionsBase {
     /**
@@ -21,3 +15,10 @@ * Name of model stored in `node_modules/@pr0gramm/fluester/lib/whisper.cpp/models`
 }
-export type WhisperOptions = WhisperOptionsWithModelPath | WhisperOptionsWithModelName;
+export interface WhisperClientOptionsWithModelPath extends WhisperClientOptionsBase {
+    modelPath: string;
+}
+export type WhisperClientOptions = WhisperClientOptionsWithModelName | WhisperClientOptionsWithModelPath;
+export interface WhisperOptions {
+    whisperOptions?: FlagTypes;
+    sourceLanguage?: string;
+}
 export interface WhisperClient {
@@ -30,2 +31,3 @@ /**
     translate: (filePath: string, options: WhisperOptions) => Promise<TranscriptLine[]>;
+    detectLanguage: (filePath: string) => Promise<LanguageDetectionResult | undefined>;
 }
@@ -32,0 +34,0 @@ export declare function createWhisperClient(options: WhisperClientOptions): WhisperClient;

dist/index.js

@@ -6,3 +6,3 @@ import * as fs from "node:fs/promises";
 import { modelFileNames } from "./model.js";
-import transcriptToArray from "./transcript.js";
+import transcriptToArray, { parseDetectedLanguage, } from "./transcript.js";
 export function createWhisperClient(options) {
@@ -12,16 +12,13 @@ const effectiveOptions = {
         ...options,
+        modelPath: getModelPath(options),
     };
+    async function ensureModel() {
+        if (!(await fs.stat(effectiveOptions.modelPath))) {
+            throw new Error(`Model not found at "${effectiveOptions.modelPath}".`);
+        }
+    }
     return {
-        // TODO
-        /**
-         * @param filePath The audio file to translate.
-         * @param options
-         * @returns English translation of the audio file. If it's already english, it will be a transcription.
-         */
         translate: async (filePath, options) => {
+            await ensureModel();
             try {
-                const modelPath = getModelPath(options);
-                if (!(await fs.stat(modelPath))) {
-                    throw new Error(`Model not found at "${modelPath}".`);
-                }
                 // 1. create command string for whisper.cpp
@@ -31,3 +28,9 @@ const flags = options.whisperOptions
                     : [];
-                const args = [...flags, "-m", modelPath, "-f", filePath];
+                const args = [
+                    ...flags,
+                    "-m",
+                    effectiveOptions.modelPath,
+                    "-f",
+                    filePath,
+                ];
                 // 2. run command in whisper.cpp directory
@@ -43,2 +46,13 @@ // TODO: add return for continually updated progress value
         },
+        detectLanguage: async (filePath) => {
+            await ensureModel();
+            const result = await execute(effectiveOptions.executablePath, [
+                "--detect-language",
+                "-m",
+                effectiveOptions.modelPath,
+                filePath,
+            ]);
+            // TODO: Check for probability threshold
+            return parseDetectedLanguage(result.stderr.toString());
+        },
     };
@@ -45,0 +59,0 @@ }

dist/test.js

 import { createWhisperClient } from "./index.js";
 (async function run() {
-    const client = createWhisperClient({});
+    const client = createWhisperClient({
+        modelName: "base.en",
+    });
     try {
         const transcript = await client.translate("/Users/Shared/twospeak_clip.wav", {
             // modelPath: "/Users/Shared/custom-models/ggml-base.en.bin",
-            modelName: "base.en",
             whisperOptions: { wordTimestamps: true },
@@ -9,0 +10,0 @@ });

dist/transcript.d.ts

@@ -7,1 +7,6 @@ export interface TranscriptLine {
 export default function parseTranscript(vtt: string): TranscriptLine[];
+export interface LanguageDetectionResult {
+    language: string;
+    probability: number;
+}
+export declare function parseDetectedLanguage(output: string): LanguageDetectionResult | undefined;

dist/transcript.js

@@ -17,1 +17,15 @@ export default function parseTranscript(vtt) {
 }
+export function parseDetectedLanguage(output) {
+    // whisper.cpp appears to use two-letter-country-code:
+    // https://github.com/ggerganov/whisper.cpp/blob/940cdb13964a563d86c7dc6e160a43ec89b8bb2e/whisper.cpp#L195-L295
+    // Example line:
+    // whisper_full_with_state: auto-detected language: en (p = 0.958819)
+    const res = /auto-detected language: (\w\w)\s*\(p\s*=\s*(\d+\.\d+)\)/.exec(output);
+    if (!res) {
+        return undefined;
+    }
+    return {
+        language: res[1].toLowerCase(),
+        probability: Number(res[2].trim()),
+    };
+}

package.json

 {
   "name": "@pr0gramm/fluester",
-  "version": "0.3.9",
+  "version": "0.3.10",
   "license": "MIT",
@@ -5,0 +5,0 @@ "description": "Node.js bindings for OpenAI's Whisper. Optimized for CPU.",

README.md

@@ -1,3 +0,2 @@
-# fluester – [ˈflʏstɐ] [![CI](https://github.com/pr0gramm-com/fluester/actions/workflows/CI.yml/badge.svg)](https://github.com/pr0gramm-com/fluester/actions/workflows/CI.yml) [![CD](https://github.com/pr0gramm-com/fluester/actions/workflows/CD.yml/badge.svg)](https://github.com/pr0gramm-com/fluester/actions/workflows/CD.yml) ![version](https://img.shields.io/npm/v/%40pr0gramm/fluester) ![downloads](https://img.shields.io/npm/dm/%40pr0gramm/fluester)
- ![License](https://img.shields.io/npm/l/%40pr0gramm%2Ffluester)
+# fluester – [ˈflʏstɐ] [![CI](https://github.com/pr0gramm-com/fluester/actions/workflows/CI.yml/badge.svg)](https://github.com/pr0gramm-com/fluester/actions/workflows/CI.yml) [![CD](https://github.com/pr0gramm-com/fluester/actions/workflows/CD.yml/badge.svg)](https://github.com/pr0gramm-com/fluester/actions/workflows/CD.yml) ![version](https://img.shields.io/npm/v/%40pr0gramm/fluester) ![downloads](https://img.shields.io/npm/dm/%40pr0gramm/fluester) ![License](https://img.shields.io/npm/l/%40pr0gramm%2Ffluester)
 
@@ -32,6 +31,9 @@ Node.js bindings for OpenAI's Whisper. Hard-fork of [whisper-node](https://github.com/ariym/whisper-node).
 ## Usage
+### Translation
 ```js
 import { createWhisperClient } from "@pr0gramm/fluester";
 
-const client = createWhisperClient();
+const client = createWhisperClient({
+  modelName: "base",
+});
 
@@ -43,3 +45,3 @@ const transcript = await client.translate("example/sample.wav");
 
-### Output (JSON)
+#### Output (JSON)
 ```js
@@ -55,21 +57,16 @@ [
 
-### Usage with Additional Options
+### Language Detection
 ```js
-import whisper from "@pr0gramm/fluester";
+import { createWhisperClient } from "@pr0gramm/fluester";
 
-const filePath = "example/sample.wav", // required
+const client = createWhisperClient({
+  modelName: "base",
+});
 
-const options = {
-  modelName: "tiny.en", // default
-  modelPath: "/custom/path/to/model.bin", // use model in a custom directory
-  whisperOptions: {
-    generateTxt: false, // outputs .txt file
-    generateSubtitles: false, // outputs .srt file
-    generateVtt: false, // outputs .vtt file
-    timestampSize: 10, // amount of dialogue per timestamp pair
-    wordTimestamps: true // timestamp for every word
-  }
+const result = await client.detectLanguage("example/sample.wav");
+if(!result) {
+  console.log(`Detected: ${result.language} with probability ${result.probability}`);
+} else {
+  console.log("Did not detect anything :(");
 }
-
-const transcript = await whisper(filePath, options);
 ```
@@ -76,0 +73,0 @@

New alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Fixed alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Improved metrics

Total package byte prevSize

5314012

increased by0.02%

Lines of code

4178

increased by0.87%

Worsened metrics

Number of lines in readme file

99

decreased by-2.94%

No dependency changes