It is fairly common practice to configure a fully automated way to start, process, and finish steps (or projects) in XTM Cloud, via particular settings defined in an XTM Cloud project template. The process usually involves a project with a particular project template having a single workflow step assigned to it that is, in the vast majority of cases, matched with machine translation (MT), and MT-matched segments automatically set to “Done”. This setup can be configured to trigger both the automatic starting of a project and automatic completion of its steps.
However, it does not necessarily have to be MT matches that do the work. The existence of other matches can also trigger the mechanism, provided that they are configured properly.
This article describes which options to take into account when configuring an automated workflow for a project.
First, you need to enable automatic completion of a step/project in the global settings for your XTM Cloud instance. To do so:
Select the Automatically finish steps where all segments are green/done option (administrative privileges required!).
Do not forget to save your changes afterwards.
You can also set workflow steps to auto-finish on their due dates (see the Auto-finish step on its due date option).
When this option is selected, the Auto-finish step on its due date command is added at the bottom of the Edit workflow window (displayed when you select Configuration → Settings → Workflow → Workflow definitions → Edit workflow definition or Projects → Project editor → Workflow → Workflow editor).
As set in workflow(default) → Workflow steps will be finished automatically on the same date and at the same time as specified in the workflow.
As set in workflow, but with changed due time → Workflow steps will be finished automatically on the same date as specified in the workflow, but at a different time than specified there.
Setting this parameter also activates the additional Set step to "Finished" by parameter with which you can specify which username is to be used to automatically finish the workflow step. The information about the user is only stored in the Project history.
Open a particular project template (Projects → Templates), and go to the Workflow tab screen. There are three actions here:
Automatically start the project → This option is present in the project template by default. If you would like to have your projects to be started automatically when their analysis has finished, in XTM Cloud, you must select this option.
Automatically finish steps where all segments are green/done → This option is only displayed in the project template if you have previously enabled the Automatically start the project option in the global settings (see the previous bullet point). If you would like to have your workflow step/project finished automatically once all the segments have been set to “Done”, you must select this option.
You must also assign a Linguist to the project’s first workflow step. Otherwise, the project will not start automatically even though the Automatically start the project has been selected.
Machine translation (MT) matches
Once the global/project template settings that enable automation have been configured, you can set up machine translation. If you do not use machine translation, you can safely skip this section.
Once machine translation is enabled, add it to your project template. Open your project template (Projects → Templates) and go to the Project details → Machine translation section. There are a couple of things you need to configure here:
Start off by enabling the machine translation engine itself (select the relevant engine).
Do not search for an MT match when one of the selected matches exists → This option might seem complex at first glance, but it can be explained as follows:
If you are only making use of MT matches in your automated project workflow and do not expect any other TM match to be applied in the target segments, you can keep the TM match types (as shown below) selected, since they will not appear anyway.
If you are making use of MT matches in your automated project workflow, but other TM matches might be applied in the target segments as well, you must deselect the relevant TM match types, as shown below. Otherwise, if a particular segment is filled with one of the TM match types below, the MT match will not be retrieved from the machine translation engine. Not every type of TM match can be configured to automatically set a particular segment to “Done”, in XTM Workbench. (For more details about this, see the next paragraph).
Set MT matches to done → This is the most important option to configure as it specifies whether or not segments populated with MT matches are to be set to “Done”, and if so, in which workflow steps. A particular workflow step will not finish automatically if there are any incomplete segments. Depending on the number of workflow steps in your project, you might choose: All steps, First step only or All except last step. Assuming that you would like to have a fully automated project workflow, from start to finish, you might decide to choose the All steps option.
Populate target with match → This option is as essential as the previous one, in terms of project automation. It specifies whether or not target segments are populated with an MT match: they should be if you would like to have them set to “Done”. For this reason, if you would like to have your workflow step finish automatically, make sure that you select Always, fixing inline differences.
Translation memory (TM) matches
Finally, a project that is based solely on TM matches can also be configured to finish automatically, although this is more difficult to achieve and this type of project automation is much less common. Nonetheless, it is possible so, if you are not using machine translation in your workflow or you are combining it with regular TM matches in your projects, this section explains how to set up your TM matches properly.
Configure your TM matches in the global settings for your XTM Cloud instance. To do so, select Configuration → Settings → Translation → TM.
The TM tab screen contains settings for almost all essential TM matches.
Depending on whether or not a particular TM match exists, the default matching order in a project is:
In-context exact (ICE) matches,
Machine translation (MT) matches,
For example, if you have both an MT match and ICE match/Leveraged match available for a particular target segment, in an XTM Cloud project, and both of them happen to be configured to populate the target segment, the ICE match/Leveraged match takes precedence.
As you can see, MT matches are pretty low in the order of importance as far a project’s TM matching priority is concerned. Therefore, in the case of projects that employ both MT matches and TM matches, it is important to set them in such a way that, no matter which kind of match is inserted in segments, the segments are still be automatically set to “Done”.
ICE matches cannot be permanently disabled. For example, if an ICE match is found for a particular project, you cannot prevent it from being displayed in XTM Workbench. However, what you can do is decide if the match shall be inserted in the target segment or not, by selecting the relevant option in the Populate target with match dropdown.
If the ICE match is “approved”, the target segment will automatically be set to “Done”. This is expected behavior, and there is no way to change that.
If the ICE match is “not approved”, you can specify whether the corresponding target segment is to be set to “Done” by using the following option: Set ICE matches from not approved TM to done, with some additional settings, as shown in the screenshot below.
Leveraged matches can be permanently disabled, using the Show leveraged matches option. Also, in contrast to ICE matches, you can specify whether target segments are to be filled with Leveraged matches or not, using the Populate target with match option. Finally, you can use the Set leveraged matches from approved TM to done and Set leveraged matches from not approved TM to done options for full control over what types of Leveraged match, “approved” and/or “not approved”, make the target segment be set to “Done” automatically.
Fuzzy matches, Repeats, Fuzzy repeats
Finally, there are Fuzzy matches, Repeats and Fuzzy repeats. As you can see below, you can also enable or disable them or fill target segments with these matches as required. However, the difference between these TM matches and ICE or Leveraged matches is that the former cannot be configured to make a target segment be set to “Done” when it has been populated with one of those matches. When you use machine translation and select TM matches from the Do not search for an MT match when one of the selected matches exists option, it is important to remember that target segments filled with Fuzzy matches, Repeats or Fuzzy repeats will not be set to “Done”, especially in the context of Fuzzy Matches, which stand higher than MT matches in the order of TM matching priority.