added support for audio and captions

[pietrop]

• Added support to upload audio with animated wave form
• added optional support for burning captions to both audio and video.

See example below

Note that, text of tweet, captions, and media content are randomly paired, for the purpose of testing only might come up with a better example use case later if needed. For now only focusing on functionality.

The code base is quite modular, but here are some things I'd like to improve before merging, with some refactoring

• In the individual modules, inside /lib, (eg audio-to-video-waveform, burn-captions, trim,convert-to-twitter-video-specs etc..) that use fluent-ffmpeg, there has to be a better way to wrap a pipe into a function (?) perhaps with promises instead of using callbacks?
• The logic in main index.js could be considerably simplified, to aid readability and future proofing.

Welcome, Ideas, suggestions, PRs etc...

[pietrop]

For context, this was made with the aim to use it in this autoEdit integration OpenNewsLabs/autoEdit_2#74

[pietrop]

Trying wrapping fluent-ffmpeg modules in promises, eg from fluent-ffmpeg/node-fluent-ffmpeg#710 as illustrated in example

  new Promise((resolve, reject) => {
    ffmpeg('...')
      .on('progress', (progress) => {
        log.verbose(`[ffmpeg] ${JSON.stringify(progress)}`);
      })
      .on('error', (err) => {
        log.info(`[ffmpeg] error: ${err.message}`);
        reject(err);
      })
      .on('end', () => {
        log.verbose('[ffmpeg] finished');
        resolve();
      })
      .save('...');
  });

---

example - trim module refactor, wrapping in promises.

/**
 * Trims video and match to the specs from the Twitter video documentation.
 * See Twitter Docs for for more info
 * https://developer.twitter.com/en/docs/media/upload-media/uploading-media/media-best-practices
 */
const ffmpeg = require('fluent-ffmpeg');

function trimVideo(opts, callback) {
  if(opts.ffmpegPath !== undefined){
    ffmpeg.setFfmpegPath(opts.ffmpegPath)
  }
  return new Promise((resolve, reject)=>{
  ffmpeg(opts.inputFile)
    .seekInput(opts.inputSeconds)
    .setDuration(opts.durationSeconds)
    .output(opts.outputFile)
    .audioCodec('aac')
    .videoCodec('libx264')
    .size('1280x720')
    .aspect('16:9')
    .format('mp4')
    .on('end', ()=>{
        resolve(opts.outputFile);
      }
    )
    .on('error',(error)=>{
      reject(error);
    })
    .run();
  })
}

module.exports = trimVideo;

example usage

const trimVideo = require('./index.js');
const path = require('path');
//  The `time` argument may be a number (in seconds) or a timestamp string (with format `[[hh:]mm:]ss[.xxx]`).
let opts = {
    inputFile: path.join(__dirname, '../../assets/test.mp4'),
    inputSeconds: 10,
    durationSeconds: 20,
    outputFile: path.join(__dirname, '../../assets/test_clipped.mp4')
}

trimVideo(opts)
    .then((res)=>{
        console.log('in example usage')
        console.log(res);
    })
    .catch(error => console.log(error));

Next, need to look at how changing all the modules like that effects the refactor of main index.js and hopefully simplifies that code.

[pietrop]

Still a bit of refactoring to wrap up, but had a first go at adding promises to both individual modules and main index.js with some good results.

[pietrop]

Added promises to both individual modules and refactored main index.js.

Also updated README.