Page MenuHomePhabricator

Issues with apostrophes when translating with MinT
Closed, ResolvedPublic

Description

When translating the "Avocado toast" article from English to French (quick link) and Catalan (quick link) the apostrophes usually used in those languages (e.g., d'avocat or /l'avocat) are translated using different marks. For example, in the first paragraph translated to French the translation includes apart from the initial d'avocat (as expected), the following apostrophes using unexpected symbols: d"avocado, d'" épices, d " ingrédients, d\'olive and l\'hummus which use unexpected characters instead of the expected ' to connect the article and the noun without additional spaces.

Below, text included below and screenshot with cases underlined.

Original text in EnglishTranslation with MinT to FrenchTranslation with MinT to Catalan
Avocado toast is a type of open sandwich consisting of toast with mashed avocado, and any of a variety of spices and flavorful ingredients. The most popular are usually salt and black pepper, sometimes lemon juice or other citrus, while others include olive oil, hummus, vinegar, red pepper flakes, feta, duqqa and tomato.Le toast d'avocat est un type de sandwich ouvert composé de toasts avec de la pures d"avocado, et de toute une variété d'" épices " et d " ingrédients savoureux. Les plus populaires sont généralement le sel et le poivre noir, parfois le jus de citron ou d'autres agrumes, tandis que d’autres incluent l’huile d\'olive, l\'hummus, le vinaigre, les flocons de poipre rouge, la feta, duqqa et la tomate.La tost de l'avocat és un tipus de sandíqued obert que consisteix en tost amb puré d'avoucado, i qualsevol d 'una varietat d" especies i ingredients saborosos. Els més populars són generalment la sal i el piment negre, a vegades suc de llimona o altres cítrics, mentre que altres inclouen oli d'oliva, humus, vinagre, flocs de piments vermells, feta, duqqa i tomà.

fr.wikipedia.org_wiki_Special_ContentTranslation_from=en&to=fr&page=Avocado+toast&targettitle=Toast+%C3%A0+l%27avocat(iPad Air).png (1×2 px, 757 KB)

Event Timeline

Pginer-WMF triaged this task as Medium priority.May 29 2023, 1:57 PM
Pginer-WMF updated the task description. (Show Details)
Pginer-WMF moved this task from Backlog to General translation functionality on the MinT board.

Change 927520 had a related patch set uploaded (by Santhosh; author: Santhosh):

[mediawiki/services/machinetranslation@master] Increase no_repeat_ngram_size to 4

https://gerrit.wikimedia.org/r/927520

Change 927520 merged by jenkins-bot:

[mediawiki/services/machinetranslation@master] Increase no_repeat_ngram_size to 4

https://gerrit.wikimedia.org/r/927520

Change 927160 had a related patch set uploaded (by KartikMistry; author: KartikMistry):

[operations/deployment-charts@master] Update MinT to 2023-06-06-111852-production

https://gerrit.wikimedia.org/r/927160

Change 927160 merged by jenkins-bot:

[operations/deployment-charts@master] Update MinT to 2023-06-06-120533-production

https://gerrit.wikimedia.org/r/927160