@wovin/tranz-cli 0.0.26

package/dist/commands/split/index.js

@@ -0,0 +1,95 @@
+import { Command, Flags } from '@oclif/core';
+// import { Logger } from '@ztax/logger/src'
+import { getName } from "../../utils/file-utils.js";
+const splitOnSil = (silenceInfo, output, sourceFileName) => {
+    const splitSeeks = [];
+    let report = `
+        Exclusions and inclusions:
+        `;
+    let segmentDuration = 0;
+    for (const [inx, eachSil] of silenceInfo.entries()) {
+        const { stSec, endSec, duration } = eachSil;
+        segmentDuration = duration;
+        const nextSil = silenceInfo[inx + 1];
+        if (!nextSil)
+            break;
+        const seekExclude = (duration - (Split.DEFAULTS.SILBUF * 2)).toFixed(3);
+        const segmentSt = +(endSec - Split.DEFAULTS.SILBUF).toFixed(3);
+        const segmentDur = +((nextSil?.stSec ?? 0) + Split.DEFAULTS.SILBUF - segmentSt).toFixed(3);
+        report += ` ex:${seekExclude} - in:${segmentDur}`;
+        const ts = 'todo';
+        const outputPath = `${output}/${inx}-${ts}-from-${sourceFileName}.wav`;
+        splitSeeks.push([segmentSt, segmentDur, ts, outputPath]);
+    }
+    report += ` ex:${segmentDuration.toFixed(3)} `;
+};
+const doSplit = () => {
+    const { WARN, LOG, DEBUG, VERBOSE, ERROR } = { WARN: console.warn, LOG: console.log, DEBUG: console.log, VERBOSE: console.debug, ERROR: (...Any) => new Error(JSON.stringify(Any)) }; // Logger.setup(Logger.INFO); // eslint-disable-line no-unused-vars
+    LOG('//TODO');
+    // return new Promise<SplitResult>(async (resolveFx, rejectFx) => {
+    //   const transResults = await getWhisperTranscript(phase1OutputPath, output)
+    // if (!transResults?.trans?.transcription) return rejectFx(ERROR('bum transcription', transResults))
+    // const { code, trace, trans } = transResults
+    // const transTrace = ['Transcription results: ', { code, trace, trans }]
+    // info.traces.push(transTrace); DEBUG(transTrace)
+    // const transcription = trans.transcription as TranscriptionArray
+    // const isCreateInterimArtifacts = true // TODO implement chaining if this is false and add param
+    // const splitSeeks = []
+    // for (const [inx, eachTrans] of transcription.entries()) {
+    //   const { offsets: { from: segSt, to: segEnd },
+    //     timestamps: { from: tsFrom, to: tsTo },
+    //     text, speaker_turn_next: isNewSpeaker } = eachTrans
+    //   const {
+    //     offsets: { from: nxSegSt, to: nxSegEnd },
+    //     // timestamps: { from: tsFromNx, to: tsToNx },
+    //     text: textNx, speaker_turn_next: isNewSpeakerNx } = transcription[inx + 1] ?? { offsets: {} }
+    //   const gap = (nxSegSt ?? segEnd) - segEnd
+    //   const duration = (segEnd + gap * 0.5) - segSt
+    //   const ts = tsFrom // TODO absolute uts ts if startTimestamp is given or
+    //   const outputPath = `${output}/${sourceFileName}-${inx}-${ts}-${text.substring(0, 42)}.wav` as string
+    //   splitSeeks.push([segSt, duration, ts, outputPath])
+    // }
+    // console.log('starting segment splitting' /*, { splitSeeks,  leadingSil, leadingSilEnd, inputParams  }*/);
+    // for (const [inx, [sSt, sDur, ts, outputPath]] of splitSeeks.entries()) {
+    //   console.log({ inx, sSt, sDur, ts })
+    //   ffmpeg(phase2OutputPath, { logger: console }) //.inputOptions(inputParams)
+    //     .seekInput(sSt)
+    //     .duration(sDur)
+    //     .save(outputPath as string)
+    //     .on('end', function (stdout, stderr) {
+    //     })
+    // }
+    // console.log('wrote', splitSeeks.length, 'files to', output)
+    // })
+};
+class Split extends Command {
+    async run() {
+        const { args: { input }, flags, flags: { output, silthr, sildur } } = await this.parse(Split);
+        const sourceFileName = getName(input);
+        this.log(`\n   Preparing ${flags.input} into ${output}, maybe \n`);
+        const report = await Split.split();
+        console.log(report);
+    }
+}
+Split.description = 'Split audio file -  on silence, or split info from transcription';
+Split.DEFAULTS = {
+    OUTDIR: './out',
+    SILDUR: '1.3',
+    SILBUF: 0.2,
+    SILTHR: '-35dB',
+};
+Split.flags = {
+    output: Flags.string({ char: 'o', description: 'output directory', required: false, default: Split.DEFAULTS.OUTDIR }),
+    silthr: Flags.string({ char: 's', description: 'silence threshold', required: false, default: Split.DEFAULTS.SILTHR }),
+    sildur: Flags.string({ char: 'd', description: 'silence duration', required: false, default: Split.DEFAULTS.SILDUR }),
+};
+Split.args = [
+    { name: 'input', description: 'input file', required: true },
+];
+Split.split = doSplit;
+Split.examples = [
+    `$ tranz prep split /pathTo/audio.wav
+...TODO
+`,
+];
+export default Split;

package/dist/index.d.ts

@@ -0,0 +1,2 @@
+export { run } from '@oclif/core';
+export * from '@wovin/tranz';

package/dist/index.js

@@ -0,0 +1,3 @@
+export { run } from '@oclif/core';
+// Re-export library for convenience
+export * from '@wovin/tranz';

package/dist/utils/file-utils.d.ts

@@ -0,0 +1,5 @@
+import path from 'node:path';
+export declare const getExt: (filePath: string) => string;
+export declare const getName: (filePath: string) => string;
+export declare const getNameWithExt: (filePath: string) => string;
+export declare const getFileInfo: (filePath: string) => path.ParsedPath;

package/dist/utils/file-utils.js

@@ -0,0 +1,15 @@
+import path from 'node:path';
+export const getExt = (filePath) => {
+    return getFileInfo(filePath).ext;
+};
+export const getName = (filePath) => {
+    return getFileInfo(filePath).name;
+};
+export const getNameWithExt = (filePath) => {
+    const { name, ext } = getFileInfo(filePath);
+    return `${name}.${ext}`;
+};
+export const getFileInfo = (filePath) => {
+    const normed = path.normalize(filePath);
+    return path.parse(normed);
+};

package/dist/utils/formatBytes.d.ts

@@ -0,0 +1,14 @@
+/**
+ * Format bytes into human-readable size string
+ *
+ * @param bytes - Number of bytes
+ * @param decimals - Number of decimal places (default: 1)
+ * @param binary - Use binary (1024) vs decimal (1000) units (default: true)
+ * @returns Formatted string like "36.9 MiB" or "38.7 MB"
+ *
+ * @example
+ * formatBytes(38746650) // "36.9 MiB"
+ * formatBytes(38746650, 2) // "36.95 MiB"
+ * formatBytes(38746650, 1, false) // "38.7 MB"
+ */
+export declare function formatBytes(bytes: number, decimals?: number, binary?: boolean): string;

package/dist/utils/formatBytes.js

@@ -0,0 +1,22 @@
+/**
+ * Format bytes into human-readable size string
+ *
+ * @param bytes - Number of bytes
+ * @param decimals - Number of decimal places (default: 1)
+ * @param binary - Use binary (1024) vs decimal (1000) units (default: true)
+ * @returns Formatted string like "36.9 MiB" or "38.7 MB"
+ *
+ * @example
+ * formatBytes(38746650) // "36.9 MiB"
+ * formatBytes(38746650, 2) // "36.95 MiB"
+ * formatBytes(38746650, 1, false) // "38.7 MB"
+ */
+export function formatBytes(bytes, decimals = 1, binary = true) {
+    if (bytes === 0)
+        return '0 B';
+    const k = binary ? 1024 : 1000;
+    const sizes = binary ? ['B', 'KiB', 'MiB', 'GiB', 'TiB', 'PiB'] : ['B', 'KB', 'MB', 'GB', 'TB', 'PB'];
+    const i = Math.floor(Math.log(bytes) / Math.log(k));
+    const value = bytes / Math.pow(k, i);
+    return `${value.toFixed(decimals)} ${sizes[i]}`;
+}

package/dist/utils/secrets.d.ts

@@ -0,0 +1,18 @@
+/**
+ * Load an API key from environment variable or secret file
+ * @param envVar - Environment variable name (e.g., 'MISTRAL_API_KEY')
+ * @param secretFile - Path to secret file (e.g., 'secret/mistral')
+ * @returns The API key, or undefined if not found
+ */
+export declare function loadSecret(envVar: string, secretFile: string): Promise<string | undefined>;
+/**
+ * Load an API key from environment variable or secret file, with error handling
+ * @param envVar - Environment variable name
+ * @param secretFile - Path to secret file
+ * @param displayName - Human-readable name for error messages (e.g., 'Mistral API key')
+ * @returns The API key
+ * @throws If the key is not found
+ */
+export declare function loadSecretOrThrow(envVar: string, secretFile: string, displayName: string): Promise<string>;
+export declare const loadApiKey: typeof loadSecret;
+export declare const loadApiKeyOrThrow: typeof loadSecretOrThrow;

package/dist/utils/secrets.js

@@ -0,0 +1,40 @@
+import * as fs from 'node:fs';
+/**
+ * Load an API key from environment variable or secret file
+ * @param envVar - Environment variable name (e.g., 'MISTRAL_API_KEY')
+ * @param secretFile - Path to secret file (e.g., 'secret/mistral')
+ * @returns The API key, or undefined if not found
+ */
+export async function loadSecret(envVar, secretFile) {
+    // Try environment variable first
+    const envValue = process.env[envVar];
+    if (envValue)
+        return envValue;
+    // Fall back to secret file
+    try {
+        const fileContent = fs.readFileSync(secretFile, 'utf-8');
+        return fileContent.trim() || undefined;
+    }
+    catch {
+        // File doesn't exist or can't be read
+        return undefined;
+    }
+}
+/**
+ * Load an API key from environment variable or secret file, with error handling
+ * @param envVar - Environment variable name
+ * @param secretFile - Path to secret file
+ * @param displayName - Human-readable name for error messages (e.g., 'Mistral API key')
+ * @returns The API key
+ * @throws If the key is not found
+ */
+export async function loadSecretOrThrow(envVar, secretFile, displayName) {
+    const secret = await loadSecret(envVar, secretFile);
+    if (!secret) {
+        throw new Error(`${displayName} not found. Set ${envVar} environment variable or create ${secretFile} file`);
+    }
+    return secret;
+}
+// Keep backwards compatibility aliases
+export const loadApiKey = loadSecret;
+export const loadApiKeyOrThrow = loadSecretOrThrow;

package/oclif.manifest.json

@@ -0,0 +1,284 @@
+{
+  "version": "0.0.26",
+  "commands": {
+    "prep": {
+      "id": "prep",
+      "description": "Prepare audio file - normalize, noise reduce, split on silence",
+      "strict": true,
+      "pluginName": "@wovin/tranz-cli",
+      "pluginAlias": "@wovin/tranz-cli",
+      "pluginType": "core",
+      "aliases": [],
+      "hiddenAliases": [],
+      "examples": [
+        "$ tranz prep AI_Could_Be_The_End_Of_Democracy.m4a' -s -35dB -d 1.5\n--> sets silence threshold to -35dB and min silence duration to 1.5s\n--> runs normalization and noise removal\n"
+      ],
+      "flags": {
+        "norm": {
+          "name": "norm",
+          "type": "boolean",
+          "char": "n",
+          "description": "do normalization?",
+          "required": false,
+          "allowNo": false
+        },
+        "verbose": {
+          "name": "verbose",
+          "type": "boolean",
+          "char": "v",
+          "description": "trace more?",
+          "required": false,
+          "allowNo": false
+        },
+        "outdir": {
+          "name": "outdir",
+          "type": "option",
+          "char": "o",
+          "description": "output directory",
+          "required": false,
+          "multiple": false,
+          "default": "./out"
+        },
+        "silthr": {
+          "name": "silthr",
+          "type": "option",
+          "char": "s",
+          "description": "silence threshold",
+          "required": false,
+          "multiple": false,
+          "default": "-35dB"
+        },
+        "sildur": {
+          "name": "sildur",
+          "type": "option",
+          "char": "d",
+          "description": "silence duration",
+          "required": false,
+          "multiple": false,
+          "default": "1.3"
+        }
+      },
+      "args": {
+        "input": {
+          "name": "input",
+          "description": "input file",
+          "required": true
+        }
+      },
+      "DEFAULTS": {
+        "SILDUR": "1.3",
+        "SILBUF": 0.2,
+        "SILTHR": "-35dB",
+        "RNNN": "https://raw.githubusercontent.com/GregorR/rnnoise-models/master/somnolent-hogwash-2018-09-01/sh.rnnn"
+      }
+    },
+    "scribe": {
+      "id": "scribe",
+      "description": "Transcribe audio file - optionally prepare first",
+      "strict": true,
+      "pluginName": "@wovin/tranz-cli",
+      "pluginAlias": "@wovin/tranz-cli",
+      "pluginType": "core",
+      "aliases": [],
+      "hiddenAliases": [],
+      "examples": [
+        "$ tranz scribe 'path/to/16khz-audiofile.wav'\nruns whisper and outputs and saves a transcription json\n\n$ tranz scribe 'path/to/whatever-audiofile.mp3' -p\nfirst prepares and then runs whisper and outputs and saves a transcription json\n\n$ tranz scribe 'path/to/audiofile.mp3' -r mistral -m voxtral-mini-latest\ntranscribe using Mistral provider with specified model\n\n$ tranz scribe 'path/to/audiofile.mp3' -r mistral --timestamps segment\ntranscribe using Mistral with segment-level timestamps\n\n$ tranz scribe 'path/to/audiofile.mp3' -r mistral --diarization --timestamps word\ntranscribe using Mistral with speaker diarization and word-level timestamps\n\n$ tranz scribe 'path/to/audiofile.mp3' -r greenpt -l en\ntranscribe using GreenPT provider with language specification\n\n$ tranz scribe 'path/to/long-audio.mp3' --auto-split --max-segment 300\nautomatically split long audio at silence boundaries (max 5 min segments)\n"
+      ],
+      "flags": {
+        "prep": {
+          "name": "prep",
+          "type": "boolean",
+          "char": "p",
+          "description": "do prep?",
+          "required": false,
+          "allowNo": false
+        },
+        "withGPU": {
+          "name": "withGPU",
+          "type": "boolean",
+          "char": "g",
+          "description": "use gpu?",
+          "required": false,
+          "allowNo": false
+        },
+        "norm": {
+          "name": "norm",
+          "type": "boolean",
+          "char": "n",
+          "description": "do normalization?",
+          "required": false,
+          "allowNo": false
+        },
+        "output": {
+          "name": "output",
+          "type": "option",
+          "char": "o",
+          "description": "output directory",
+          "required": false,
+          "multiple": false,
+          "default": "./out"
+        },
+        "diarization": {
+          "name": "diarization",
+          "type": "boolean",
+          "char": "d",
+          "description": "enable speaker diarization",
+          "required": false,
+          "allowNo": false
+        },
+        "separate_speakers": {
+          "name": "separate_speakers",
+          "type": "boolean",
+          "description": "separate via diarization (deprecated: use --diarization)",
+          "required": false,
+          "allowNo": false,
+          "deprecated": {
+            "message": "use --diarization instead"
+          }
+        },
+        "provider": {
+          "name": "provider",
+          "type": "option",
+          "char": "r",
+          "description": "transcription provider (whisper, mistral, greenpt)",
+          "multiple": false,
+          "options": [
+            "whisper",
+            "mistral",
+            "voxtral",
+            "greenpt"
+          ],
+          "default": "whisper"
+        },
+        "model": {
+          "name": "model",
+          "type": "option",
+          "char": "m",
+          "description": "model name (provider-specific)",
+          "multiple": false
+        },
+        "language": {
+          "name": "language",
+          "type": "option",
+          "char": "l",
+          "description": "language code (e.g., en, de, fr)",
+          "multiple": false
+        },
+        "timestamps": {
+          "name": "timestamps",
+          "type": "option",
+          "description": "enable timestamp granularity (segment, word) - Mistral only",
+          "multiple": false,
+          "options": [
+            "segment",
+            "word"
+          ]
+        },
+        "no-pauses": {
+          "name": "no-pauses",
+          "type": "boolean",
+          "description": "disable pause-based line break formatting in transcript preview",
+          "allowNo": false
+        },
+        "auto-split": {
+          "name": "auto-split",
+          "type": "boolean",
+          "description": "automatically split long audio files at silence boundaries",
+          "allowNo": false
+        },
+        "max-segment": {
+          "name": "max-segment",
+          "type": "option",
+          "description": "maximum segment duration in seconds for auto-split (default: 600)",
+          "multiple": false,
+          "default": 600
+        },
+        "split-silence-dur": {
+          "name": "split-silence-dur",
+          "type": "option",
+          "description": "minimum silence duration for split points (default: 1.0)",
+          "multiple": false,
+          "default": "1.0"
+        }
+      },
+      "args": {
+        "input": {
+          "name": "input",
+          "description": "input file",
+          "required": true
+        }
+      },
+      "DEFAULTS": {
+        "DIARIZE": false,
+        "SILDUR": "1.3",
+        "SILBUF": 0.2,
+        "SILTHR": "-35dB",
+        "MODEL_KEYS": {
+          "tinyd": "ggml-small.en-tdrz.bin",
+          "small": "ggml-small.bin",
+          "medium": "ggml-medium.bin"
+        },
+        "MODELS": {
+          "tinyd": "https://huggingface.co/akashmjn/tinydiarize-whisper.cpp/resolve/main/ggml-small.en-tdrz.bin",
+          "small": "https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-small.bin",
+          "medium": "https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-medium.bin"
+        }
+      }
+    },
+    "split": {
+      "id": "split",
+      "description": "Split audio file -  on silence, or split info from transcription",
+      "strict": true,
+      "pluginName": "@wovin/tranz-cli",
+      "pluginAlias": "@wovin/tranz-cli",
+      "pluginType": "core",
+      "aliases": [],
+      "hiddenAliases": [],
+      "examples": [
+        "$ tranz prep split /pathTo/audio.wav\n...TODO\n"
+      ],
+      "flags": {
+        "output": {
+          "name": "output",
+          "type": "option",
+          "char": "o",
+          "description": "output directory",
+          "required": false,
+          "multiple": false,
+          "default": "./out"
+        },
+        "silthr": {
+          "name": "silthr",
+          "type": "option",
+          "char": "s",
+          "description": "silence threshold",
+          "required": false,
+          "multiple": false,
+          "default": "-35dB"
+        },
+        "sildur": {
+          "name": "sildur",
+          "type": "option",
+          "char": "d",
+          "description": "silence duration",
+          "required": false,
+          "multiple": false,
+          "default": "1.3"
+        }
+      },
+      "args": {
+        "input": {
+          "name": "input",
+          "description": "input file",
+          "required": true
+        }
+      },
+      "DEFAULTS": {
+        "OUTDIR": "./out",
+        "SILDUR": "1.3",
+        "SILBUF": 0.2,
+        "SILTHR": "-35dB"
+      }
+    }
+  }
+}

package/package.json

@@ -0,0 +1,90 @@
+{
+  "type": "module",
+  "name": "@wovin/tranz-cli",
+  "version": "0.0.26",
+  "description": "CLI for audio and video transcription",
+  "author": "gotjoshua @gotjoshua",
+  "bin": {
+    "tranz": "./bin/run.js",
+    "tranzz": "./bin/dev.js"
+  },
+  "homepage": "https://gitlab.com/onezoomin/ztax/tranz",
+  "license": "ISC",
+  "main": "",
+  "repository": {
+    "type": "git",
+    "url": "git+https://gitlab.com/wovin/wovin.git",
+    "directory": "packages/@wovin/tranz-cli"
+  },
+  "files": [
+    "/bin",
+    "/dist",
+    "/npm-shrinkwrap.json",
+    "/oclif.manifest.json"
+  ],
+  "publishConfig": {
+    "access": "public"
+  },
+  "dependencies": {
+    "@dharmendrasha/ffmpeg-normalize": "^1.8.2",
+    "@oclif/core": "^2",
+    "@oclif/plugin-help": "^5",
+    "@oclif/plugin-plugins": "^3.1.8",
+    "@ztax/logger": "^0.2.1",
+    "fluent-ffmpeg": "^2.1.2",
+    "@wovin/tranz": "^0.0.26"
+  },
+  "devDependencies": {
+    "@oclif/test": "^2.4.2",
+    "@types/chai": "^5.2.3",
+    "@types/fluent-ffmpeg": "^2.1.21",
+    "@types/mocha": "^10.0.10",
+    "@types/node": "^24.10.1",
+    "chai": "^4",
+    "eslint": "^7.32.0",
+    "eslint-config-oclif": "^4",
+    "eslint-config-oclif-typescript": "^1.0.3",
+    "mocha": "^9",
+    "oclif": "^3.11.3",
+    "shx": "^0.4.0",
+    "ts-node": "^10.9.1",
+    "tslib": "^2.6.1",
+    "typescript": "^5.9.3",
+    "whisper-node": "^1.1.1"
+  },
+  "oclif": {
+    "bin": "tranz",
+    "dirname": "tranz",
+    "commands": "./dist/commands",
+    "plugins": [
+      "@oclif/plugin-help",
+      "@oclif/plugin-plugins"
+    ],
+    "topicSeparator": " ",
+    "topics": {
+      "prep": {
+        "description": "Prepare audio for transcription"
+      }
+    }
+  },
+  "engines": {
+    "node": ">=12.0.0"
+  },
+  "bugs": "https://gitlab.com/onezoomin/ztax/tranz/-/issues",
+  "keywords": [
+    "oclif",
+    "cli",
+    "transcription"
+  ],
+  "types": "dist/index.d.ts",
+  "scripts": {
+    "tranz": "./bin/dev.js",
+    "oclif": "oclif",
+    "manifest": "ts-node oclif manifest",
+    "build": "shx rm -rf dist && tsc -b",
+    "lint": "eslint . --ext .ts --config .eslintrc",
+    "posttest": "npm run lint",
+    "test": "mocha --forbid-only \"test/**/*.test.ts\"",
+    "version": "oclif readme && git add README.md"
+  }
+}