CR-Nimble v0.4.0

This is a maintenance release to align the library with JAX 0.4.x.

CR-Nimble v0.3.2

Documentation

Added

Vectors:

• Circular buffers
• Binary heap

DSP

• sliding_windows

Synthetic Signals

• picket_fence
• heavi_sine
• bumps
• blocks
• doppler
• ramp
• cusp
• sing
• hi_sine
• lo_sine
• lin_chirp
• two_chirp
• quad_chirp
• mish_mash
• werner_sorrows
• leopold

Miscellaneous

• cr.sparse.io moved to cr.nimble.io

Fixed

• Handling of complex signals in build_signal_from_indices_and_values
• Handling of 100 percentile in num_largest_coeffs_for_energy_percent

CR-Nimble v0.3.1

This release includes functions which were introduced during the development of block sparse bayesian learning in CR-Sparse library.

Documentation

Added

Metrics

• normalized_mse
• percent_rms_diff
• compression_ratio
• cr_to_pss
• pss_to_cr

Noise

• awgn_at_snr_std

Matrices

• mat_column_blocks
• block_diag

Vectors

• has_equal_values_vec

Special matrices

• toeplitz_mat
• toeplitz_mult
• circulant_mat
• circulant_mult

Misc

• to_tex_matrix

CR-Nimble v0.3.0

Added

Data Compression

• Binary data encoding/decoding
• Run length encoding/decoding
• Fixed length encoding/decoding

Digital Signal Processing

• Scaling functions
• Quantized
• Energy fraction based thresholding

Metrics

• Percentage root mean square difference

Removed

Changed

• Statistical normalization renamed with changes in return type
• Digital signal processing related functions moved under
  cr.nimble.dsp

Improved

• Documentation improved
• API organization improved

CR-Nimble v0.2.4

Added

• Digital signal processing utilities moved from cr-sparse to cr-nimble
• Moved discrete number related functions from cr-sparse.
• Some sparse vector and matrix processing functionality moved from cr-sparse.

Removed

• Unnecessary __init__.py files removed.

Notes

• Jax 0.3.14 compatibility
• Aligning version numbering across sister projects.

CR-Nimble v0.1.1

Added

• distance, matrix, ndarray, metrics, modules were moved from cr-sparse to cr-nimble
• some more vector functions were moved from cr-sparse to cr-nimble

Improved

• All unit tests were moved to 64-bit floating-point data.

CR-Nimble v0.1.0

This is the first-ever release of CR-Nimble. It has been created by extracting code from the cr-sparse project.