Speed up internal fuzzy analysis by copying binary files

The t5memory integration for internal fuzzy analysis should be changed. So far for internal fuzzy analysis translate5 downloads the memory as binary from t5memory and uploads it again under a new name as a copy.

This process is time and resource consuming.

Therefore with T5TMS-137 t5memory will provide a new end point, where translate5 identifies a TM with its name and tells t5memory a new name for a copy that t5memory should create itself.

Please note: The existing integration must be kept for old OpenTM2.

Please coordinate the development with Orest.