Skip to content

Latest commit

 

History

History
146 lines (97 loc) · 4.01 KB

CHANGES.md

File metadata and controls

146 lines (97 loc) · 4.01 KB

Release history:

v1.1.5

Changes

BLD: Set rpath for libomp on MacOS to fix compatibility with latest XGBoost version

v1.1.4

Changes

PKG: Suppress error for double initialisation of OpenMP

v1.1.3

Changes

BLD: Unrestrict nanobind by @RUrlus in #112

Internal

FIX: [C++] Remove usage of nb::raw_doc

v1.1.2

Changes

BLD: Restrict nanobind version to <2.0 by @RUrlus in #111

Internal

CICD: Add trusted publishing and updated MacOS runners by @RUrlus in #108 MAINT: Correct name typo in readme by @mmtevelde in #110 CICD: Bump pypa/cibuildwheel from 2.18.0 to 2.18.1 by @dependabot in #109

v1.1.1

Internal

  • FIX: Prevent Scipy from dropping columns that are all zero for sub-matrices

v1.1.0

Add new function to select top-n from blocks of a sparse matrix matmul. Function will return a zipped matrix Z, where Z = [sorted top n results > lower_bound for each row of C_j], where C_j = A.dot(B_j) and where B has been split row-wise into sub-matrices B_j.

API

  • ENH: new function zip_sp_matmul_topn can zip matrices zip_j A.dot(B_j)

v1.0.0

This introduces major and potentially breaking changes to the API. Please see the migration guide in the README for details.

API

  • API: awesome_cossim_topn is superseded with sp_matmul_topn.

  • API: awesome_cossim_topn has been deprecated and will be removed in a future version.

  • API: ntop parameter has been renamed to topn

  • API: lower_bound parameter has been renamed to threshold

  • API: use_threads and n_jobs parameters have been combined into n_threads

  • API: return_best_ntop parameter has been removed

  • API: test_nnz_max parameter has been removed

  • API: default parameter value for threshold changed from 0.0 to None (disabled)

  • API: default parameter value for sort changed to False

  • ENH: Add support for 32 and 64bit integers

  • BLD: Add support for CPython 3.12

Internal

  • BLD: Switch to pyproject.toml based setup (scikit-build-core)
  • FIX: [C++] Resolve unneeded memory allocation that solved hidden buffer-overrun in multithreaded implementation
  • BLD: [C++] Switch to Nanobind bindings
  • CHG: [C++] Switch to OpenMP for multithreading
  • ENH: [C++] Use MaxHeap to collect top-n results over vector of candidates

v0.3.6

  • Adds support for Cython >= 3.0

v0.3.5

  • Restrict Cython version to <3.0

v0.3.4

  • Add Python 3.11 wheels
  • Fix a compilation error when std:: is missing

v0.3.3

  • Upgrade to 0.3.3, since PyPI had versioning problem when we release 0.3.2

v0.3.2

  • Fix the Numpy ABI compatibility issues issue-48 and all related issues
  • Add Github Actions to build Python release in Linux, MacOS and Windows automatically

v0.3.1

  • Adding the possibility to use smaller data type float32 instead of the default float64
  • Adding unit tests to cover both data types
  • Adding unit tests to cover return_best_ntop==False

v0.3.0

  • defragmented memory used during computation PR-53
  • Enable Github Action for unit test
  • add license metadata PR-47

v0.2.9

  • added unit tests.
  • awesome_cossim_topn checks for zero input matrices.
  • fix flake8 warnings.

v0.2.8

  • quick fix for nt installation problem [ISSUE-26]

v0.2.7

  • add n_jobs value validation
  • matrix dimension validation
  • change windows mingw32 compiler
  • fix numpy and Cython install dependency

v0.2.6

  • import parallel implementation

v0.2.5

  • fix python 2 install problem

v0.2.5

  • fix python 2 install problem, but failed

v0.2.3

  • Make long description use markdown format

v0.2.2

  • update readme file name

v0.2.1

  • fix the problem when pip install by tar.gz;
  • update meta-information of package

v0.2, Nov 14, 2018:

  • improve the function import layer. Now awesome_cossim_topn can be used directly after from sparse_dot_topn import awesome_cossim_topn.
  • improve the module installation

v0.1, July 13, 2017:

  • Release the first version