Cambios en "Machine translation enhancement for source language detection edge cases"

Is your feature request related to a problem? A clear and concise description of what the problem is.

The futureu.europa.eu platform is one of the first Decidim instances to be using the machine translation feature.

We are faced with the following challenge :

As a logged in user browsing the site in Spanish, I might come across a proposal in English. As I also speak English I’m prone to posting a comment to reply to the proposal in English.

The problem we face is that the machine translation takes as its source language the one I’m using to browse the website.

So in this case, If I’m browsing the website in Spanish but post in English something like “This is a great idea”. The machine translation will take the string thinking it’s Spanish and provide an erroneous translation for all the other languages.

Describe the solution you'd like

We propose to solve this problem by adding a language selection dropdown to the forms used by users to insert generated content (UGC) (meetings, proposals, comments)

Acceptance criteria

Given that I’m logged in
When I am writing a comment, a proposal, debate or a meeting.
Then I can select the language in which I’m posting my content.

This language dropdown will be shown on the UGC forms only if there is more than one language activated on the organization. Otherwise, an input type hidden having the default language should be present on the form for further use. This is meant to serve as a starting point in case multiple languages are being activated on the platform, or it is decided to have the machine translation activated on a later date. 

Here is the related Pull Request : https://github.com/tremend-cofe/decidim/pull/117

Describe alternatives you've considered

Additional context

This meta proposal doesn’t fix the problem entirely, as some users might not take time to select the right language in which they are posting.

We will post a second proposal that will leverage on the automatic language detection feature that some machine translation services offer so that we are able to offer a more complete solution.

Does this issue could impact on users private data?

No

Funded by
EU Commission