v1.0.6

• added complexity_cost as a parameter
• fixed an off-by-one bug in the interval pruner
• refactored the specialized and general solver
• compute gini values per feature only based on actual possible splits
• add question length to the python api

v1.0.5

• increased minimum supported python to 3.10 and sklearn to 1.6
• added the predict_proba method (issue #1)
• moved pycontree folder into a subfolder python

v1.0.4

• First release on pypi
• Fix memory usage of bitset representation of data
• Removed global_skip (possibly suboptimal)