rpm/python-openTSNE
SHA256
1
0
Fork 0
forked from pool/python-openTSNE
Code Pull Requests Activity

Accepting request 963465 from devel:languages:python:numeric

Browse Source 
OBS-URL: https://build.opensuse.org/request/show/963465
OBS-URL: https://build.opensuse.org/package/show/openSUSE:Factory/python-openTSNE?expand=0&rev=4
This commit is contained in:
Dominique Leuenberger 2022-03-21 19:11:37 +00:00 committed by Git OBS Bridge
commit 7455d9d7c5
4 changed files with 33 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
openTSNE-0.6.1-gh.tar.gz
View File

@ -1,3 +0,0 @@
version https://git-lfs.github.com/spec/v1
oid sha256:020caa3b210e01476c029d9faa9d1a905b8adb123a2010f5c2b91f95a25f8b37
size 23306737

3
openTSNE-0.6.2-gh.tar.gz Normal file
View File

@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:54ba0dcf782bc88b544d22d5dc9d342b27dbaf47abb87ef8d1d80918029a2421
size 22641032

22
python-openTSNE.changes
View File

@ -1,3 +1,25 @@
-------------------------------------------------------------------
Sun Mar 20 19:38:50 UTC 2022 - Ben Greiner <code@bnavigator.de>
- Update to 0.6.2
* By default, we now use the MultiscaleMixture affinity model,
enabling us to pass in a list of perplexities instead of a
single perplexity value. This is fully backwards compatible.
* Previously, perplexity values would be changed according to the
dataset. E.g. we pass in perplexity=100 with N=150. Then
TSNE.perplexity would be equal to 50. Instead, keep this value
as is and add an effective_perplexity_ attribute (following the
convention from scikit-learn, which puts in the corrected
perplexity values.
* Fix bug where interpolation grid was being prepared even when
using BH optimization during transform.
* Enable calling .transform with precomputed distances. In this
case, the data matrix will be assumed to be a distance matrix.
* Fix potential problem with clang-13, which actually does
optimization with infinities using the -ffast-math flag
- Enable python310 build
- Skip a test in 32bit failing due to rounding errors
-------------------------------------------------------------------
Fri Feb 4 16:10:13 UTC 2022 - Ben Greiner <code@bnavigator.de>

13
python-openTSNE.spec
View File

@ -18,10 +18,8 @@
%{?!python_module:%define python_module() python3-%{**}}
%define skip_python2 1
# not compatile with Python 3.10 API: https://github.com/pavlin-policar/openTSNE/issues/205
%define skip_python310 1
Name: python-openTSNE
Version: 0.6.1
Version: 0.6.2
Release: 0
Summary: Extensible, parallel implementations of t-SNE
License: BSD-3-Clause
@ -29,18 +27,19 @@ URL: https://github.com/pavlin-policar/openTSNE
# tests are not packaged in the PyPI sdist, use GitHub instead
Source: %{url}/archive/v%{version}.tar.gz#/openTSNE-%{version}-gh.tar.gz
Patch0: python-openTSNE-disable-CPU-autodetection.patch
BuildRequires: %{python_module Cython}
BuildRequires: %{python_module devel >= 3.7}
BuildRequires: %{python_module numpy-devel >= 1.16.6}
BuildRequires: %{python_module scikit-learn >= 0.20}
BuildRequires: %{python_module scipy}
BuildRequires: %{python_module setuptools}
BuildRequires: c++_compiler
BuildRequires: fdupes
BuildRequires: fftw3-devel
BuildRequires: python-rpm-macros
# SECTION test requirements
BuildRequires: %{python_module pytest}
# /SECTION
BuildRequires: fdupes
Requires: python-numpy >= 1.16.6
Requires: python-scikit-learn >= 0.20
Requires: python-scipy
@ -65,7 +64,11 @@ export CFLAGS="%{optflags}"
%python_expand %fdupes %{buildroot}%{$python_sitearch}
%check
%pytest_arch --import-mode append
%ifarch %ix86 %arm32
# precision errors on 32bit
%define donttest -k "not TestTSNECorrectnessUsingPrecomputedDistanceMatrix"
%endif
%pytest_arch --import-mode append %{?donttest}
%files %{python_files}
%doc README.rst