tnkm v0.1.0

• Initial public release of tnkm, an open-source Python library for
  tensor network kernel machines.
• Unified framework for constructing and training TNKM models with
  support for nonlinear feature representations and tensor-network
  parameterizations.
• Support for multiple tensor-network parameterizations, including
  canonical polyadic (CP) and tensor-train (TT) formats.
• Structure-aware optimization methods based on alternating least squares
  (ALS), together with gradient-based optimization using JAX.
• Utilities for nonlinear system identification workflows, including
  data preprocessing, simulation, and evaluation.
• Reproducible experiment scripts and benchmark implementations
  accompanying the research paper.
• Documentation and usage examples for building and training TNKM models.
• Automated code quality checks and continuous integration.
• Package configuration and installation support.