Offline Translation Integration

Owned by Aaron Sakievich (Unlicensed)
Last updated: Mar 04, 2024 by Kenny Teves
8 min read

The Enterprise API makes it easy for integrators to export/re-import vendor-assigned tasks from/to the Enterprise TMS, so that they can be translated within a third-party CAT tool.

The client's project and community settings in the TMS must allow for offline translation before you can use these APIs.

API Diagram

There is one recommended flow to support offline translation using Enterprise's API:

Receiving Callbacks when Tasks Are Available to Work On

Polling

It is also possible to poll the GET /task endpoint to see if there are additional tasks to translate but this is not recommended. If your security requirements oblige you to pursue this method, please reach out to support@lingotek.com for more information.

Example API Responses

GET /task

Status

200

cache-control

private, no-cache, no-store, max-age=0, must-revalidate

content-type

application/json

pragma

no-cache

{ "class": [ "tasks", "Collection" ], "properties": { "limit": 10, "total": 1, "offset": 0, "size": 1 }, "entities": [ { "class": [ "task" ], "rel": [ "task" ], "properties": { "project_name": "example.com", "translation_locale": "fr-FR", "title": "Example Document", "document_name": "example_document.txt", "author_email": "", "author_name": "", "business_division": "", "business_unit": "", "campaign_id": "", "campaign_rating": 0, "contact_email": "", "contact_name": "", "content_description": "", "domain": "http://example.com", "external_application_id": "a79ee237-4c69-4126-946d-cf9cfcf13ef1", "external_document_id": "40", "external_style_id": "", "purchase_order": "", "reference_url": "http://example.com/document", "weighted_words": null, "words": 1209, "document_id": "20f93bc9-ed5d-45af-c225-1e4384d73408", "translation_id": "682ee649-9449-2925-8f2a-c2ff76b0fe43", "project_id": "ccc57f52-2f92-418f-b884-8348e95c1be9", "content_type": "node", "progress": 0, "status": "WAITING", "id": "4186f3d5-c335-4e04-a123-94d40cdd20af", "channel": "", "region": "" }, "links": [ { "rel": [ "self" ], "href": "/task/4186f3d5-c335-4e04-a123-94d40cdd20af" } ] } ], "links": [ { "rel": [ "self" ], "href": "/task?offset=0&limit=10" } ] }

GET /task/{id}/content

Status

200

cache-control

private, no-cache, no-store, max-age=0, must-revalidate

content-disposition

attachment); filename="article (node): Elit Oppeto Usitas Validus Vicis Zelus.fr_FR.xlf"

content-type

application/x-xliff+offline

pragma

no-cache

"<?xml version='1.0' encoding='UTF-8'?> <xliff xmlns="urn:oasis:names:tc:xliff:document:1.2" version="1.2" xmlns:lgtk="http://www.lingotek.com/xliff/1.2"> ... </xliff>"

PATCH /task/{id}/content

Status

202

cache-control

private, no-cache, no-store, max-age=0, must-revalidate

content-type

application/json

pragma

no-cache

{ "processes": [ "78f0e4d3-ee66-4b81-2aed-a621e46afe2a" ] }

GET /process/{id}

Status

200

cache-control

private, no-cache, no-store, max-age=0, must-revalidate

content-type

application/json

pragma

no-cache

GET /document/{id}/translation/{locale}

Status

200

cache-control

private, no-cache, no-store, max-age=0, must-revalidate

content-type

application/json

pragma

no-cache

Example HTTP Notification

Request Origin: Enterprise

Request Destination: Integrator

Query Params

community_id

1c5f137b-d9d1-44c3-8ff0-24d0460943bb

complete

true

doc_author_name

j.doe

doc_business_division

text

doc_business_unit

text

doc_campaign_id

text

doc_campaign_rating

0.0

doc_channel

text

doc_contact_email

text

doc_contact_name

text

doc_content_description

text

doc_content_type

txt

doc_cts

1537210928500

doc_domain_type

http://example.com/document

doc_external_application_id

a79ee237-4c69-4126-946d-cf9cfcf13ef1

doc_external_document_id

1

doc_external_style_id

text

doc_mts

1537211244114

doc_purchase_order

text

doc_reference_url

http://example.com/document

doc_region

text

doc_status

COMPLETE

documentId

4222311

document_id

20f93bc9-ed5d-45af-c225-1e4384d73408

locale

fr_FR

locale_code

fr-FR

original_project_id

0

phase_name

translate

progress

100

projectId

70821

project_id

ccc57f52-2f92-418f-b884-8348e95c1be9

status

COMPLETE

targetId

14638015

target_id

db65816a-e854-4b3e-b4f4-9eea10bae34c

type

target

Downloading Tasks

Use the following API calls to view and export your tasks. 

  • GET /task allows you to view the list of tasks assigned to you in the TMS.

  • GET /task/{id}/content downloads the XLIFF for a given task.

Keep in mind the following as you are exporting tasks:

  • Via the UI, a user is not able to download tasks (XLIFF files) that aren't in their Task List. Likewise, the API calls will need to authenticate using the credentials of the user to whom the task is assigned in Enterprise.

  • You can only download XLIFF for 'Active' tasks (i.e. tasks that are ready to be worked on, but have not been skipped, canceled or 'marked complete' within Enterprise).

Editing XLIFF files

While you are editing the XLIFF files you should be aware of the following, so that you can re-import the files back into the TMS when you are finished editing:

  • We follow the XLIFF 1.2 specification

  • Be sure not to modify the number of trans-units, their order or identifiers

  • Adding style to the corrections is not advisable. For example, if a translation were to change from "dog" to "puppy", "puppy" should simply replace "dog" and not be something like "dogpuppy."

  • Do not merge or split segments within your tool. If our system detects that one or more segments were merged or split, the re-import will fail.

Attributes and metadata

  • On export, we add a lgtk:phaseUUID attribute in the header of the XLIFF. This is what we match on when you re-import the XLIFF. Be sure that you do not remove this attribute. 

If the lgtk:phaseUUID attribute is missing, you will not be able to successfully re-import the XLIFF file.

  • The XLIFF files you download from Enterprise's TMS will only ever correspond to one target language. As such it is important that if you have an xml:lang for a segment it should match the target language in the header.

  • Enterprise doesn't import any meta from the uploaded XLIFF. All we do is extract and import the translations form the trans-units. We do not import anything else from the XLIFF.

  • All target attributes are optional, however, below is more information about target attributes if you include them. 

    • state: Different behavior can happen on XLIFF re-import depending on the state attribute.

    • xml:lang: If present, this should contain the locale code for the target locale, which is defined in the <file> tag's "target-language" attribute.

    • state-qualifier: This helps to clarify the origin of the translation. Click here for instructions on how to make these values appear in SDL Trados Studio. 

    • phase-name: This provides clarification as to which phase this target is in. The export provides a number, which has a description in the <phase group> section in the <header>. 

  • mrk: <mrk> is optional, but if present, there must only be one per target, and it must surround the entire segment. If there is more than one <mrk> segment in a <target>, the import fails. If there is one that does not completely encapsulate the segment, only the text within the <mrk> is imported. 

  • mid: Optional. Used as an identifier for a specific <mrk> within the target.

  • mtype: REQUIRED. This is required by the XLIFF 1.2 specification. It describes the type of <mrk>. The export uses "seg" as the mtype. If mtype is not present, the import will fail.

Uploading A Completed Task

Upon completing a Task outside of Enterprise’s TMS, it is a simple process to push them back into the TMS. The following API calls will re-import content into the TMS. View the documentation here to troubleshoot any errors you may receive when uploading tasks to the TMS. 

  • PATCH /task/content

  • PATCH /task/{id}/content

Keep the following in mind as you re-import XLIFF file(s) to complete your assigned task(s):

Process Completions

  • PATCH /task/content and PATCH /task/{id}/content both schedule new jobs in the TMS for executing the the re-import process.

  • The IDs for the re-import process(es) are included in the API response.

  • GET /process/{id} can be used to get the status of a given process (e.g. to check if it failed or not; or if it successfully finished).

File Structure

  • Be sure that the structure of the XLIFF has not changed. 

  • We accept both zip folders containing XLIFF files or single XLIFF files.

  • We only re-import the translations and not any metadata from the XLIFF that may have been added to the file.

Authentication

  • Via the UI, a user is not able to re-import XLIFF for a task that isn't in their Task List. Likewise, the API calls will need to authenticate using the credentials of the user to whom the task is assigned in Enterprise.

  • Authorship is attributed to the user re-importing the tasks through the API. 

  • You can only re-import XLIFF for 'Active' tasks (i.e. tasks that are ready to be worked on, but have not been skipped, canceled or 'marked complete' within Enterprise).

Effect on the TMS

  • Existing translations in the phase are overwritten by the ones in the imported XLIFF file.

  • When reviewing translations in a third-party CAT tool, you should always re-import the XLIFF file to the TMS even if you didn't need to make any edits to the translations. Re-importing an unedited XLIFF lets the TMS know that the assigned task was completed and did not need any edits; and will be reflected as such in your Task List upon successfully re-importing the XLIFF file. 

  • If there is no translation for a segment when the XLIFF file was downloaded, and there still is not a translation when it is re-imported back into the TMS, the import process will skip over that segment and it will appear as a non-translated segment.

  • If a segment has a translation in the XLIFF upon download, but the <target> is blank upon re-import, the TMS will keep the previous translation.