Last radar has always zero weights

[jm5791]

Hi devs !

First of all, many thanks for making pydda available, it's a very interesting tool !

For information, during my different explorations and trials, I noticed that the current implementation of retrieval.wind_retrieve._get_dd_wind_field_tensorflow() always affect zero weights to the observation data associated to the last radar provided, when more than one radar are provided.

Note: the same seems to occur with the equivalent scipy implementation.

[rcjackson]

Hello. Thank you for bringing attention to this! I've released PyDDA 1.3.0, which fixes this issue in #90.