Implement proper segmentation for InstantTranslate

InstantTranslate currently implements only a rudimentary segmentation. Within this feature a reusable segmentation engine shall be implemented (as part of the Okapi Plugin), that uses the Tranlste5-default BCONF's segmentation rules to perform the segmentation. This segmentation then should be utilized for InstantTranslate

Useful Links:

https://okapiframework.org/wiki/index.php/SRX

https://unicode-org.github.io/icu/userguide/strings/regexp.html#regular-expressions
https://en.wikipedia.org/wiki/Unicode_character_property

Essential for the implementation is, that we need to make use of the existing regular expressions in the okapi srx.

Okapi is a Java app and srx uses ICU Regex, which is not completly compatible with Java.

https://okapiframework.org/wiki/index.php/SRX_and_Java shows, how this is solved in Okapi.

We must check, what this means for the the regex used in the srx files. Are they icu regex and Okapi simply maps a regex something that works with Java and is equal? Or are actually Java regex in the file? What does that mean for PHP, if we use the same regexes? We should keep this as simple as possible. One way would be, to implement a check, if a regex works in php without an error. If yes, all is fine. If no, we skip the rule. Reason for this strategy: Everything else will become very complex. And: On first sight all the rules currently present in the okapi srx file should  work just as you would expect it in PHP. So in practical use this strategy will reach the goal (pragmatic approach).