Problem

In some cases uplifting results in worse tag placing then we can achieve.

That is caused by our tag handling logic. As we guess we delete some tags in original query but when we apply them back we do so one by one.

Solution

After guessing is done - pass data of what protected content was skipped in process and try to filter those tags when we apply original tags back

How to check that all is good

1. Create task with segment that has two integers or some other content to protect in it. should use 1 rule for both
2. Create a TM that has only one 2nd entity protected in it
3. Pre-translate the task or open this segment in editor and check match table
4. Confirm that second protected entity is same thing as in original not protected segment