Translation Memory import and export

Concepts

Concept Explanation
Translation Unit A pair made of source and target segments
Fields The fields contain metadata about the translation unit.

Importing a translation memory

The translation memory that you want to import must have one of the following formats: tmx, sdltm, zip, tmx.gz or sdlxliff.

Note

Before the import action, make sure you have already created a translation memory using the Create Translation Memory endpoint.

To import a translation memory, perform a POST request to the Import Translation Memory endpoint and provide these details: translationMemoryId, the translation memory file and a couple of properties.

Warning

Pay special attention to the order of properties and file. It must be exactly as described in the contract (properties first).

If you are not sure what properties to include, here's a short description for each of them:

  1. sourceLanguageCode and targetLanguageCode - These fields represent the language direction of your import. Make sure to provide both of them as they are required fields.
  2. importAsPlainText - When true, the translation units will be imported as plain text, excluding any text markup.
  3. exportInvalidTranslationUnits - When true, all the translation units that failed to import are saved in a tmx file.
  4. triggerRecomputeStatistics - When true, will recompute the fuzzy index statistics once the import is done.
  5. targetSegmentsDifferOption - Option to specify how to handle translation units in the Translation Memory, if the target segments differ.
    • addNew - Add a new translation unit and leave the original translation unit with the same source untouched.
    • overwrite - Overwrite all the existing translation units where the source segment matches with the new content.
    • leaveUnchanged - Do not change the existing translation units with the same source segment and ignore the new translation unit.
    • keepMostRecent - Delete all the existing translation units with matching source segments and retain only the most recent translation unit
  6. unknownFieldsOption - Option to specify how to handle translation units and unknown fields in the Translation Memory.
    • addToTranslationMemory - The translation unit is processed and the unknown user-defined fields are automatically added to the setup.
    • skipTranslationUnit - The translation units containing unknown user-defined fields are skipped.
    • ignore - The translation unit is processed and the unknown user-defined fields are ignored (not added to the setup).
    • failTranslationUnitImport - If any translation unit contains an unknown user-defined field an error is thrown.
  7. onlyImportSegmentsWithConfirmationLevels - Only segments with the specified confirmation levels will be imported.
    • translated - The segment has been fully translated, but not yet reviewed.
    • approvedTranslation - The translation has been reviewed and approved, but not signed-off.
    • approvedSignOff - The translation has been approved and signed-off.
    • draft - The target segment may have been changed, but it's not yet considered fully translated.
    • rejectedTranslation - The translation has been reviewed and rejected.
    • rejectedSignOff - The translation was rejected in the sign-off process.

Note that only the sourceLanguageCode and targetLanguageCode fields are required, the others already have default values, so you don't have to set them.

If the import is successful, you will get back the import id and the status(normally queued).

For more details about translation memory imports you can also check out the Importing TM content page.

Import TM

Polling a translation memory import

Some imports may take longer, depending on the size of the Translation Memory.

To check if the import finished, you can perform a GET request to the Poll Translation Memory Import endpoint and supply the importId and the translationMemoryId. The import is complete when the status is done.

Exporting a translation memory

To export a translation memory, perform a POST request to the Export Translation Memory endpoint and provide the translationMemoryId and a valid languageDirection of your translation memory.

If the export is successful, you will get back an exportId and the export status. If the export did not succeed, the response will contain an error message.

Export TM

Polling a translation memory export

Some exports may take longer, depending on the size of the Translation Memory.

To check if the export finished, you can perform a GET request to the Poll Translation Memory Export endpoint and supply the exportId. The export is complete when the status is done.

Downloading a translation memory export

If an export is completed successfully (the status received at the polling is done), it can be downloaded by performing a GET request to the Download Exported Translation Memory endpoint. Just supply as path parameter the exportId and the response will contain the exported translation memory file in tmx.gz format.