Compresso

Latest version: v3.3.0

Safety actively analyzes 629900 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 2 of 3

2.3.2

-----

* chore: update MANIFEST.in
* test: protect against rle zero overflow
* fix: logic error in RLE window encoder
* fix: calculation for valid stream used bits instead of bytes
* refactor: use templates to shorten the logic for decompress
* chore: tighten up whitespace

2.3.1

-----

* chore: exclude test data files from sdist

2.3.0

-----

* docs: fix whitespace
* perf(remap): substantially faster
* docs: show how to use remap and valid
* docs: add docstring to remap
* feat(remap): adds remap without compression
* refactor: use cython type notation

2.2.0

-----

* feat: if RLE encode error, automatically recompress as 8,8,1

2.1.0

-----

* fix: throw error if unable to RLE encode windows
* fix: catch cpp exceptions in Python
* feat(cli): adds steps and six flags

2.0.0

-----

* chore: update changelog
* docs: note that this implementation is format version 0
* docs: note that a second stage compressor is needed
* docs: make CLI example clearer
* chore: update changelog
* fix: really add CLI
* feat: add compresso cli
* chore: delete obsolete dockerfile
* chore: update dockerfile with oldest-supported-numpy
* chore: rename connectomics.cpso.lzma to connectomics.npy.cpso.lzma
* chore: add sample data to repo
* chore: compresso.hxx to compresso.hpp and add MANIFEST.in
* docs: note that any possible input is possible in header
* docs: name author of modifications
* chore: skip pypy builds
* docs: describe the stream format better
* chore: add github builds
* fix(encode\_boundaries): ensure 1 is correct type
* chore: update all runtime\_errors to start with "compresso"
* fix: ensure that zero locations can never segfault
* feat: allow nonstandard step sizes
* fix: incorrect casting of pointers
* chore(setup.cfg): replace deprecated hyphens
* docs: describe support for 6 connected
* feat: support for both 4 and 6 connectivity (4 is default)
* feat: add connectivity to header
* refactor(cc3d): change name of main fn to connected\_components
* perf(decompress): improve decode\_boundary performance
* docs: change wording slightly
* docs: description of the algorithm
* fix: uint8 windows
* docs: discuss implementation changes
* docs: save edits needed to experiment with 6-way ccl
* docs: update readme
* fix: use new interface to cc3d
* fix(cc3d): add back in optimizations
* chore: update year
* chore: update license for cc3d.hpp to LGPLv3
* docs: compresson performance in docs
* fix: add figure back in to docs
* docs: describe modifications to the protocol
* fix: support window8 in cpp
* test: add tests for steps
* fix: start to support uint32 windows
* test: check black arrays
* test: ensure compresso.labels matches actual unique labels
* chore: update appveyor
* fix: magic should be cpso not cpso\0
* test: make sure maximum values are handled
* feat: add more tests
* fix: handle zero length arrays
* fix(cc3d): can't use optimizations with start\_label
* docs: add comments
* feat: add python methods for inspecting the raw stream
* perf: vectorize decoding of ccids
* chore: whitespace
* fix: prevent amplification of start label
* fix(cc3d): start label was not assigned correctly
* fix: better error messages for decoding locations
* fix: incorrect nx and ny
* fix: add check for zero step size
* fix: more rounding errors for nx,ny,nz
* fix: bug in boundary decoding
* fix: bug with rounding nblocks up
* fix: memory leak
* fix: ensure labels handles 0 properly
* fix(labels): use new decoding protocol for locations
* fix: able to encode last 6 units for indeterminate locations
* docs: fix error in docstring
* feat: add functions for reading labels and dtypes from compressed stream
* docs: add a header
* fix: lots of cleanup
* fix(cc3d): start\_label required some finessing
* fix: make sure output pointer is mapped to right data type
* fix: cc3d needs to offset start label as well in relabel
* fix: bug in rle compression
* fix: skipping first values for uniform arrays
* fix(cc3d): need to have a new offset label for each z level
* perf: specialization for uint16\_t window renumber
* wip: trying to figure out decoding error
* fix: compiles!
* fix: forgot to decode RLE
* feat: add decompression
* feat: support uint16 for 4x4x1
* refactor: cc3d switched to use black as on
* perf: maybe faster unique
* fix: some cleanup
* refactor: move stream writing into a function
* fix: some encoding errors
* feat: read header function
* fix: bounds check was in the wrong direction
* wip: overhauling the code
* fix: wrong order of arguments
* fix: try to match the 2d slice processing of original connected components
* fix: memory errors
* wip: attempting to fix ordering
* test: add order test
* fix: unsigned long -> uint64\_t
* chore: no support for 32-bit windows
* chore: update build system
* docs: place note on unofficialness in a more attractive location
* docs: remove inapplicable section on running experiments
* docs: add example, badge, and update requirements
* chore: a few changes for the 1.0 release

Page 2 of 3

© 2024 Safety CLI Cybersecurity Inc. All Rights Reserved.