Add UMAP support for dimension reduction - #588

Overview

PR under Issue #574 - Add UMAP support to Dimension Reduction Operator
I've added UMAP (Uniform Manifold Approximation and Projection) support to the dimension reduction operator. UMAP is a dimensionality reduction technique that can be used as an alternative to t-SNE, offering similar visualization quality with better preservation of global structure and faster performance on larger datasets.

Changes Made

• Added UMAPReduction implementation to the dimension reduction operator
• Updated the DimensionReductionFactory to support creating UMAP models
• Added umap-learn dependency to pyproject.toml and requirements
• Added tests for UMAP functionality
• Fixed import sorting issues

Testing

• Added unit tests specifically for UMAP reduction
• Ensured existing t-SNE functionality continues to work
• Verified that the dimension reduction operator properly handles both methods

These changes allow users to choose between t-SNE and UMAP for their dimension reduction needs, providing more flexibility and potentially better performance for large datasets.