@Wang, Zheng By default, Azure AI Translator service API is case-sensitive, meaning that it matches terms in the source text based on case. This is documented here.
- Partial sentence application. When your glossary is applied to part of a sentence, the Document Translation API checks whether the glossary term matches the case in the source text. If the casing doesn't match, the glossary isn't applied.
- Complete sentence application. When your glossary is applied to a complete sentence, the service becomes case-insensitive. It matches the glossary term regardless of its case in the source text. This provision applies the correct results for use cases involving idioms and quotes.
Do you think this is causing the behavior you are seeing?
If this answers your query, do click Accept Answer
and Yes
for was this answer helpful. And, if you have any further query do let us know.