pct_tools.utils

Maintenance of stored system-matrix blocks: recompression and row subsampling.

Both operations read a compressed block written by pct_tools.io, transform it on the C++ side and write a new file, without loading the matrix into Python. Subsampling keeps a random fraction of the rows of a matrix block and the matching entries of its b-vector, to emulate a lower dose from the same simulation.

pct_tools.utils.subsample_matrix_vector_pair(input_path, projection, output_path, subsampling_ratio, pixel_width, *, seed=0, compression_level=9, shuffle_method=ShuffleMethod.BLOSC_SHUFFLE, codec='lz4hc', verbosity_level=0)[source]

Subsample matrix and b-vector for a given projection.

Reads <input_path>/matrix-<pixel_width>mm/<projection>.mat and <input_path>/b/b_<projection>.vec, keeps a random fraction of the rows and writes the result to the same relative locations under output_path.

Parameters:

  • input_path (str) – The path to the input matrices.

  • projection (str) – The projection to subsample.

  • output_path (str) – The path to the output matrices.

  • subsampling_ratio (float) – The fraction of rows to keep, in [0, 1].

  • pixel_width (float) – The pixel width in mm; selects the matrix-<pixel_width>mm directory.

  • seed (int) – Seed of the random number generator used to pick rows.

  • compression_level (int) – The compression level to use when storing the result.

  • shuffle_method (ShuffleMethod) – The blosc shuffle method to use.

  • codec (Codec) – The Blosc codec to use. Default is LZ4HC.

  • verbosity_level (int) – The verbosity level.

pct_tools.utils.recompress_matrix(filename, output, compression_level=9, shuffle_method=ShuffleMethod.BLOSC_SHUFFLE, codec='lz4hc', verbosity_level=0)[source]

Recompress a matrix with a given codec and compression level.

Parameters:

  • filename (str) – The name of the file containing the matrix.

  • output (str) – The name of the output file.

  • compression_level (int) – The compression level to use when storing it.

  • shuffle_method (ShuffleMethod) – The blosc shuffle method to use.

  • codec (Codec) – The Blosc codec to use. Default is LZ4HC.

  • verbosity_level (int) – The verbosity level.