New protocol distribution

See merge request !4

Each Ethnicity has 10 t-references and 10 z-probes, 5 for each gender.

Had to add multiple samples of the same subjects for Asiatic women.

Moved metadata to sampleset

See merge request !3

[py] Added Z and T normalization objects

See merge request !2

Forgot to save the last change -_-

[py][fix] Improved the speed of the db generation.

See merge request !1

The problem was the search of the annotations in the eyes dataframe.
Made it faster by using the filename as index and creating series
instead of dataframes for each row.

'Facial_hair' and 'glasses' fields are always NULL for all the subjects.

Returned a list of Samples instead of a list of SampleSets for world set