Creating a course from another course's data

(Moved here from → this thread)

Does that mean you have found an automated way to download a course including all audio and create a new one from that data? That does sound quite interesting and might actually help to create courses with a different source language, but based on an existing one!

Hi Olaf,

Not exactly.

I would simply save the text of an existing course using the Memrise Scraper, then manually copy-and-paste the text level-by-level into a new course. I’d then add audio using the Chrome extension “Memrise Audio Uploader”.

For a course with a few dozen levels, the process would take about half an hour. Note that any additional columns (part of speech etc.) are not transferred.

*Memrise Scraper - tech189's Website
*Memrise Audio Uploader - Chrome Web Store

You can do it with Anki.

So you wouldn’t actually download/upload the course’s original audio but rather what the Memrise Audio Uploader provides. However, the original audio is actually what I’d be after. Even if text-to-speech is getting better all the time, recordings by actual humans are still preferable by far.

Do what, import to Anki? Yes, I know. With all the problems Memrise has, I’d still like to keep it in Memrise though, I simply can’t stand Anki. :person_shrugging:t4:

You can import your course including audio files to Anki and then export it in text format and have audio files extracted. I know, it’s not easy, but that’s the way to do it if you want original audio files. :slight_smile:

Yes, thanks - I knew that, but, hey, I for one won’t upload 5k+ audio files one by one, I’m just not that crazy/passionate. :slight_smile:

I just had the impression that someone found a way to automate the whole process. If I was creating a tool, I’d download all items, create an Excel workbook from them (thus allowing to insert a different source or target language) and simply assign an ID to each entry. All downloaded audio files would then use that ID for their filename. A corresponding uploader tool would then have to deal with uploading the audio files for each item by finding an audio file via the corresponding ID. Well, that’s just how I’d theoretically approach this, but I suppose you get the idea.

Memrise obviously has no API or other means to allow a 3rd party to manage/access a course’s database, and my time is both too limited and precious to play around until it works (or until I find out that it simply won’t work), and it seemed that @ian_mn had found a way or tool that would do something like that.

Sorry for the confusion.

The course we were looking at had only a small percentage of items with audio and part-of-speech etc attached. (Memrise appears to automatically and randomly add these when a course is created.)

So, I would have generated a good replica of the original course, but Memrise made @electroblastz2 a contributor so this was not needed.

On a whim, I decided to create a new course based on the current syllabus, in case anyone is interested.
*Edexcel GCSE German (2018) - by ian_mn - Memrise