scribear / ScribeAR.github.io

Live Transcription for Augmented Reality Glasses
11 stars 16 forks source link

Add transcript download features #154

Open yuxuanjerrychen01 opened 10 months ago

yuxuanjerrychen01 commented 10 months ago

We should add a feature to the application that allows the user to download the transcript locally. This would be helpful in scenarios where saving the previous transcript is beneficial, e.g. classroom lectures.

ammpr commented 9 months ago

Hi Yuxuan, this is an idea I support and was also considering. Would love to discuss the details and formatting with you.

angrave commented 9 months ago

webvtt is a common fileformat for captions. It would be useful to be able to set when the '00:00:00' time offset is, so it can be sync'd up with a video. Would also be useful to download paragraphs just plain text. (With optional start-end timepoints for each paragraph)

ammpr commented 9 months ago

Doing research on this. Reminded me of the subrip format that pirates used.

Why not .txt, webvtt, and maybe later some BLOB options? Especially as we move towards multiplatform and mobile support, the AR needs to be put into ScribeAR.

From: Lawrence Angrave @.> Sent: Thursday, September 14, 2023 9:26 PM To: scribear/ScribeAR.github.io @.> Cc: Mehdipour, Arman Michael @.>; Comment @.> Subject: Re: [scribear/ScribeAR.github.io] Add transcript download features (Issue #154)

webvtt is a common fileformat for captions. It would be useful to be able to set when the '00:00:00' time offset is, so it can be sync'd up with a video. Would also be useful to download paragraphs just plain text. (With optional start-end timepoints for each paragraph)

— Reply to this email directly, view it on GitHubhttps://urldefense.com/v3/__https:/github.com/scribear/ScribeAR.github.io/issues/154*issuecomment-1720397331__;Iw!!DZ3fjg!6rXrrbcD-ax_EMrWjlIznKgVH5hHYNOgqW0XgvTTK9Y99PNXjV1GJxAycqRIiudPt-oO1JovvoWoWiORRUsOfUIu1iwi$, or unsubscribehttps://urldefense.com/v3/__https:/github.com/notifications/unsubscribe-auth/AUCFMWNW27YRTAAF4L6JLX3X2O4DRANCNFSM6AAAAAA4R4V3QY__;!!DZ3fjg!6rXrrbcD-ax_EMrWjlIznKgVH5hHYNOgqW0XgvTTK9Y99PNXjV1GJxAycqRIiudPt-oO1JovvoWoWiORRUsOfRzeWM0C$. You are receiving this because you commented.Message ID: @.**@.>>

ammpr commented 9 months ago

I've been going through the repo and trying to learn react. It looks like react-native-fs can be used to create txt files and react-native-blob-util in conjunction with the former can be used to generate blob files which are essentially multimedia files of the kind Notability employs.

I wanted to create an optional line numbering system to help people keep track of their position, but I think that is better served as an option for an exported transcript. Perhaps we could have an export options menu with a spring, that would specify the type and characteristics of the file exported? I'm not sure if export to clipboard should be rolled in because I think having quick access to that for people doing classwork is important.

There is a package for a webvtt compiler https://www.npmjs.com/package/node-webvtt However, there also seems to be an existing project that employs react which is somewhat similar to our objectives. https://github.com/umd-mith/webvtt-player

angrave commented 9 months ago

Creating simple webvtt captions files is easy; you don't need a full library for that.

WEBVTT

00:11.000 --> 00:13.000
We are in New York City

00:13.000 --> 00:16.000
We’re actually at the Lucern Hotel & just down the street
ammpr commented 9 months ago

https://developer.mozilla.org/en-US/docs/Web/API/WebVTT_API

You're right, thanks.

yuxuanjerrychen01 commented 9 months ago

Current implementation: added a download button to the top bar; users can click the download button to download the current transcript into a txt file with no format. Possible improvements: paragraph formatting, punctuation marks, time stamps. @ammpr Sure! We can meet and talk about some more details