math/py-umap-learn: New port: Uniform Manifold Approximation and Projection

math/Makefile

@@ -984,6 +984,7 @@
     SUBDIR += py-timple
     SUBDIR += py-topologic
     SUBDIR += py-triangle
+    SUBDIR += py-umap-learn
     SUBDIR += py-uncertainties
     SUBDIR += py-unyt
     SUBDIR += py-vincenty

math/py-umap-learn/Makefile

@@ -0,0 +1,24 @@
+PORTNAME=	umap-learn
+DISTVERSION=	0.5.3
+CATEGORIES=	math python
+MASTER_SITES=	CHEESESHOP
+PKGNAMEPREFIX=	${PYTHON_PKGNAMEPREFIX}
+
+MAINTAINER=	yuri@FreeBSD.org
+COMMENT=	Uniform Manifold Approximation and Projection
+
+LICENSE=	BSD3CLAUSE
+
+RUN_DEPENDS=	${PYTHON_PKGNAMEPREFIX}numba>=0.49:devel/py-numba@${PY_FLAVOR} \
+		${PYNUMPY} \
+		${PYTHON_PKGNAMEPREFIX}pynndescent>=0.5:math/py-pynndescent@${PY_FLAVOR} \
+		${PYTHON_PKGNAMEPREFIX}scikit-learn>=0.22:science/py-scikit-learn@${PY_FLAVOR} \
+		${PYTHON_PKGNAMEPREFIX}scipy>=1.0:science/py-scipy@${PY_FLAVOR} \
+		${PYTHON_PKGNAMEPREFIX}tqdm>=3.4.0:misc/py-tqdm@${PY_FLAVOR}
+
+USES=		python:3.6+
+USE_PYTHON=	distutils autoplist pytest
+
+NO_ARCH=	yes
+
+.include <bsd.port.mk>

math/py-umap-learn/distinfo

@@ -0,0 +1,3 @@
+TIMESTAMP = 1659803954
+SHA256 (umap-learn-0.5.3.tar.gz) = dbd57cb181c2b66d238acb5635697526bf24c798082daed0cf9b87f6a3a6c0c7
+SIZE (umap-learn-0.5.3.tar.gz) = 88193

math/py-umap-learn/pkg-descr

@@ -0,0 +1,9 @@
+Uniform Manifold Approximation and Projection (UMAP) is a dimension reduction
+technique that can be used for visualisation similarly to t-SNE, but also for
+general non-linear dimension reduction. The algorithm is founded on three
+assumptions about the data:
+* The data is uniformly distributed on a Riemannian manifold;
+* The Riemannian metric is locally constant (or can be approximated as such);
+* The manifold is locally connected.
+
+WWW: https://github.com/lmcinnes/umap