Creating a corpus in a translation project
November 1, 2019
The Corpus contains sentence Pairs used to adapt a machine translation Model.
Sentence Pairs are matching sentences in the Source and Target language used to fill the Corpus that will be used later to adapt the machine translation.
Create a Corpus
To start adding or importing Pairs into the Corpus, begin by clicking on the "Create corpus" tab located on the top right.
This action will open a "Create corpus" pop-up window. Here, you'll be able to import your data via a CSV file (ensure the file is correctly formatted). Remember, it's not mandatory to import any data at this stage; Pairs and Special terms can be uploaded later, if you prefer.
Once a Corpus has been created, you can select it from your dashboard by clicking on it, and then begin adding your first Pair. At the right of the dashboard, a "Search" feature can be used to assist you in locating a specific Corpus from your list.
Selecting a Corpus from the list allows you to manually add a Pair or import your Pairs by using a CSV file (ensure the file is correctly formatted).
Clicking on the "Add pairs" tab allows you to directly create a matching sentence Pair in the Source and Target languages.
Once the Pairs are correctly imported or manually added, they will appear in the Corpus as a list of items. Here, you'll be able to delete the selected Pairs, if needed.
You can modify the matching sentences in a Pair by clicking on the Source text or on Target text.