-
Team, Professional, Business and Enterprise plans
Get in touch with Sales for licensing questions.
-
Team, Ultimate and Enterprise plans (Legacy)
Get in touch with Sales for licensing questions.
MT glossaries are a feature available in the Phrase Language AI add-on. They can be attached to a Phrase Language AI profile (or activated in Phrase Language AI if profiles are not used) to ensure terminology is handled correctly by machine translation (MT) engines.
MT glossaries function similarly to term bases but are used by MT engines instead of linguists. Terms added to a glossary are applied by the machine translation engine before being seen by the linguist ensuring provided terms are consistently substituted.
Available MT glossaries are accessible by clicking on the
tab in the Phrase Language AI page. The tab presents a searchable list of glossaries, with different columns providing specific information about each glossary.Note
Depending on file size, it can take up to 15 minutes until a glossary is attached.
MT Glossary Limits and Requirements
-
Up to 10 glossaries can be attached to the MT profile, with a maximum of 50,000 entries in total.
-
Total number of glossaries is limited by purchased plan.
-
Glossaries are not supported by all MT engines and are only available through Phrase Language AI. If supported, this will be indicated in the machine translation profile along with supported glossary languages.
-
If glossary languages are not supported by the MT engine, the glossary upload might fail with an error. Rozetta T-4OO, in particular, supports a limited list of languages.
To resolve the issue, disable the MT engine in the machine translation profile and reattach the glossary.
-
If glossaries contain duplicates, they might be rejected by some MT engines.
DeepL, in particular, treats any term appearing more than once for a language as a duplicate. If duplicates exist in any language within the glossary, the entire glossary is considered invalid for all languages, and will be rejected.
Tip
The conditional formatting functionality in spreadsheet applications can be used to identify and remove duplicates.
-
The glossary and job languages (and their codes) should correlate in order for the glossary to be applied.
Note
If using Google AutoML MT engine, multiple locales within a single MT glossary are not supported.
Use cases
-
Product names
Phrase Language AI must translate to Phrase Language AI.
-
Ambiguous words
The word bat can mean a piece of sports equipment or an animal. If translating words about sports, use a glossary to feed the Cloud Translation API the sports translation of bat and not the translation for the animal.
-
Borrowed words
Bouillabaisse in French translates to bouillabaisse in English. The English language borrowed the word from the French in the 19th century. An English speaker lacking French cultural context may not know that bouillabaisse is a fish stew dish. A glossary can override a translation so that bouillabaisse from French translates to fish stew in English.
To add a glossary, follow these steps:
To use a glossary, it must be attached to an existing MT Profile.
To attach a glossary, follow these steps:
To add a term, follow these steps:
External terminology (or glossary) files can be imported in Excel (.XLSX) or .TBX file formats. The size limit for a file that can be uploaded is 1GB. An exported term base contains all languages of the given term base.
.XLSX file preparation
.XLSX files must be formatted in specific manner before being imported.
To prepare the file, follow these steps:
-
In the .XLSX file, organize all terms into columns with each column representing one language.
-
In the first row, apply the language code for each language.
Example:
-
Save the file.
File import
To import terms to a glossary, follow these steps:
Note
Files exported from term bases may be imported into glossaries, but they would not be accepted if they contain synonyms (more than 1 term per language in a concept). Files with terms containing a pipe character are also rejected.
-
From a glossary page, click Import.
The
window opens. -
Click Choose file and select a file to import.
-
Click Import.
Terms from the selected file are added to the glossary.