Refactoring of Jump and Drop models

[VargBurz]

This PR adds parent Stair class for Jump and Drop models.

Motivation:

Jump and Drop models are almost identical. We need to remove code duplication which presents because Jump algorithm is a reverted Drop algorithm.

Changes:

• new StairModel class
• models/jump_model.py, models/drop_model.py:
  • inherit Jump/Drop classes from StairModel class
  • Jump/Drop methods do_fit, do_detect, filter_detection are StairModel class methods now
  • new get_extremum_type method
• models/model.py:
  • new ModelType enum: jump, drop, peak, trough, general
  • new ExtremumType enum: min or max, used to connect models with utils algorithms
  • fix imports order
  • get_model_type method returns ModelType
  • use new remove_duplicates_and_sort utils method instead of list(set(array[]))
  • _update_fiting_result -> _update_fitting_result
• utils/common.py:
  • move find_drop and find_jump methods to Stair model, rename to get_stair_indexes
  • new remove_duplicates_and_sort method
  • get_parameters_from_segments: remove unused model_type arg
• tests/test_models.py: tests for get_stair_indexes method
• tests/test_utils.py: test for remove_duplicates_and_sort method

[jonyrock]

what about "ExtremumType"