Implement context TM functionality for NEC-TM in MemoQ style

translate5 should support context-TM functionality for NEC-TM based on the "smeta"-attribute, when saving a segment to NEC-TM and when querying NEC-TM for matches.

See how to use smeta in the NEC-TM REST-API documentation at

https://www.nec-tm.eu/wp-content/uploads/2020/03/NEC-TM-Admin-UI-description.pdf

We will use smeta in an analogous way to how MemoQ handles context TM. This means, we save the previous segment with the attribute key "x-context-pre" in smeta and save the entire previous segment into the value. For the next segment we will use the attribute key "x-context-post". Internal tags inside the context attributes are saved in the same way as they are saved to NEC-TM when segments are saved.

As in MemoQ a segment will be surrounded by "<seg>" and "</seg>" (surround the content only for x-context-pre and x-context-post), what enables the possibility to also save more than one segment as context. If more than one segment is saved as context, the first segment in the x-context-pre key corresponds to the first segment in the x-context-post key as a segment pair for a complete context ect.

When we calculate the matchrate we will compare the previous and following segment of the segment in the task with the segments pairs in the smeta attribute. If one of them match, we have a NEC-TM context match (or 103% match).

There also is the special case of a segment, that is the start or the end of a file (not a task! we have to look for the context file specific!). For this case we have to save the information in x-context-pre or post, that there is no pre or post-segment. How this information is saved has to be tried out with MemoQ and done in the same way. Pangeanic needs to send us a TMX from MemoQ with context matches of this kind, so we can follow the way MemoQ does it.