GitLab

Added logic to shrink the training, when there's not enough data to be used as test to be in the training percentage

- Added more tests for extract_relation_between_type_and_name;
- Added tests for get_simple_name_for_type;
- Added more tests for randoop;

- Added tests for split_words;
- Added tests for extract_name_properties;
- Added tests for extract_relation_between_type_and_name;

- Added tests for process_scope;
- Added tests for process_annotations;

- Added tests for test_process_names;
- Added tests for process_comments;

- Added test for var_type;
- Improved tests for get_type_arguments.

- Renamed tests with duplicated names;
- Removed unnecessary spaces.