This version would probably make better use of space.

...since there's a new one.

Given an API key and a URL users should be able to transcribe content using the Deepgram API

The edited transcript should be editable as an Interactive Transcript:

• Markup
• Markup + Player

Not a big deal, but it looks like the playhead is further down the page.

In both SRT and WebVTT formats.

Not required in edit mode and has an impact on performance

Sometimes when editing a transcript text ends up being added between spans, whereas all text in a Hypertranscript should exist between spans.

parseData function actually:

1. takes the branches of the json and collects the useful data for the subsequent steps
2. processes deepgram data and transforms it into hyperTranscript
3. put the hyperTranscript in the DOM.

I recommend splitting the current function into three functions, so that the code is more easily testable and reusable

I would like to suggest using the notification() function in JavaScript (https://developer.mozilla.org/en-US/docs/Web/API/Notification) to notify the user that the transcription has been completed . This function can provide a more user-friendly and intuitive way for the user to be informed about the status of the transcription process: using the notification() function would also be a better solution than keeping the page open and constantly checking for updates. It would allow the user to continue using the application or navigate to a different page while the transcription is being processed, without the need to constantly monitor the status.

I believe that this small change can improve the user experience of the application and make it more convenient for the user to use.

Thank you for considering this suggestion.

https://api.deepgram.com/v1/listen?model=general&tier=enhanced&punctuate=true&language=en-GB 400 (Bad Request)https://api.deepgram.com/v1/listen?model=general&tier=enhanced&punctuate=true&language=en-GB 400 (Bad Request)

Probably because en-GB is not supported with the enhanced model https://developers.deepgram.com/api-reference/#transcription

There are probably a few things we can do here.

Currently we regenerate the contents of the download urls everytime we run the sanitise function. We could generate these on demand using a technique similar to...

const saveDataToFile = (data, filename) => {
  const a = document.createElement('a');
  a.style = 'display:none';

  document.body.appendChild(a);

  const blob = new Blob([data], {
    type: 'octet/stream',
  });

  const url = window.URL.createObjectURL(blob);

  // eslint-disable-next-line
  console.log(`saveDataTofile: ${blob.size} bytes to ${filename}`);

  a.href = url;
  a.download = filename;

  a.click();

  window.URL.revokeObjectURL(url);
  a.remove();
};

Or inserting values into the template using this...

   fetch('hyperaudio-template.html')
    .then(function(response) {
        // When the page is loaded convert it to text
        return response.text()
    })
    .then(function(html) {
        // Initialize the DOM parser
        var parser = new DOMParser();

        // Parse the text
        var doc = parser.parseFromString(html, "text/html");

        // You can now even select part of that html as you would in the regular DOM 
        // Example:
        var docArticle = doc.querySelector('article').innerHTML;

        console.log(doc);
    })

Smaller paragraphs mean better performance.

parseData function actually:

1. takes the branches of the json and collects the useful data for the subsequent steps
2. processes deepgram data and transforms it into hyperTranscript
3. put the hyperTranscript in the DOM.

I recommend splitting the current function into three functions, so that the code is more easily testable and usable

Joining by delete, etc. paragraph A and B, results in AB, highlight does not work on B part of AB.

Ideally we'd also change the mime type, maybe doing that is enough.

Recreating the word list should fix this.

Currently the "karaoke" is broken when a file is loaded.

When copying text from the transcript timecodes can be inadvertently pasted in, causing issues.

Other formatting from source documents could also cause issues so best to strip all formatting and paste plain text.