bbc / subtitles-generator Goto Github PK

View Code? Open in Web Editor NEW

47.0 21.0 6.0 633 KB

A node module to generate subtitles by segmenting a list of time-coded text - BBC News Labs

JavaScript 100.00%

newslabs news-labs subtitles stt srt ttml vtt itt json premiere

subtitles-generator's Introduction

Subtitles Generator - draft

A node module to generate subtitles by segmenting a list of time-coded text.

Exports to

It can also provide pre-segmented lines if the input is plain text.

Setup

git clone, cd into folder, npm install

Usage

const subtitlesComposer = require('./src/index.js');
// const sampleWords = // some word json 
const subtitlesJson = subtitlesComposer({words: sampleWords, type: 'json'})
const ttmlPremiere = subtitlesComposer({words: sampleWords, type: 'premiere'})
const ittData = subtitlesComposer({words: sampleWords, type: 'itt'})
const ttmlData = subtitlesComposer({words: sampleWords, type: 'ttml'})
const srtData = subtitlesComposer({words: sampleWords, type: 'srt'})
const vttData = subtitlesComposer({words: sampleWords, type: 'vtt'})

see example-usage.js for more comprehensive example.

To try locally

npx babel-node example-usage.js

`words` Input

either an array list of words objects
example

const sampleWords =[ 
      {
        "id": 0,
        "start": 13.02,
        "end": 13.17,
        "text": "There"
      },
      {
        "id": 1,
        "start": 13.17,
        "end": 13.38,
        "text": "is"
      },
      {
        "id": 2,
        "start": 13.38,
        "end": 13.44,
        "text": "a"
      },
      {
        "id": 3,
        "start": 13.44,
        "end": 13.86,
        "text": "day."
      },
...

or a string of text
Example

const sampleWords = "There is a day. ..."

If input words is plain text only (and not a list of words with timecodes) then can only use pre-segment-txt option. (see test-presegment.txt for example)

Output:

see example-output folder for examples.

System Architecture

In pseudo code, at a high level

// expecting array list of words OR plain text string

  // if array list of words, convert text into string

  // presegment the text 
     using pre segmentation algorithm to break into line of x char - default 35

// generate subtitles 
   use subtitles generators for various format to convert presegemented json into subtitles

// return trsult

Segmentation algorithm refactored from pietrop/subtitlesComposer originally by @polizoto. And subtitles generation in various originally format by @laurian and @maboaas part of BBC Subtitlelizer project.