Set up translation process in Crowdin

[fagiothree]

The Mexico team decided to go ahead with DeepL as the MT Engine. We need to set up the workflow in Crowdin as follows:

1. Machine Translation
2. Translation by Human (if there is any untranslated string from MT)
3. Proofread by Human No. 1
4. Proofread by Human No. 2 (Human No. 2 will approve the translations).

Since there are 2 proofreading steps (need to check if this possible), we might need to use a different project in crowdin (associated with the same github repo?) or to edit the current workflow to add the extra step just for Spanish as well as the machine pre-translation

[fagiothree]

We were able to add 2 proofreading steps just for Spanish and we edited the workflow as follows We created a different branch for Spanish (and any other language for which we may use machine translation). Note that DeepL is only available for Spanish and not for Spanish-Mexico so we had to choose Spanish as a language. In settings > export, for Spanish, we set that translations will be exported only if they have at least one approval (for other languages without pre-translation they are exported as soon as there is a translation).

Note from Crowdin support on having 2 proofreading steps: Different users need to be assigned for each proofreading step, because if a user is assigned to both steps, then approving the translation on the first step will transfer the string directly to END. Former Assignees (at any proofreading step in the workflow) should not be the same as current and former assignees on the other step too because their approvals would still be counted as well.