Apache Arrow: 7.0.0

[Matthew Danielson <matthewd%fastmail.us@localhost>]

Module Name:	pkgsrc-wip
Committed By:	Matthew Danielson <matthewd%fastmail.us@localhost>
Pushed By:	matthewd
Date:		Sun Apr 17 05:57:51 2022 -0700
Changeset:	bd46b6b175be84c887928ea77c419e23b786935e

Added Files:
	apache-arrow/DESCR
	apache-arrow/Makefile
	apache-arrow/TODO
	apache-arrow/distinfo
	apache-arrow/options.mk

Log Message:
Apache Arrow: 7.0.0

Language Independent Columnar memory format for data

To see a diff of this commit:
https://wip.pkgsrc.org/cgi-bin/gitweb.cgi?p=pkgsrc-wip.git;a=commitdiff;h=bd46b6b175be84c887928ea77c419e23b786935e

Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.

diffstat:
 apache-arrow/DESCR      |  4 ++++
 apache-arrow/Makefile   | 62 +++++++++++++++++++++++++++++++++++++++++++++++++
 apache-arrow/TODO       |  2 ++
 apache-arrow/distinfo   | 11 +++++++++
 apache-arrow/options.mk | 56 ++++++++++++++++++++++++++++++++++++++++++++
 5 files changed, 135 insertions(+)

diffs:
diff --git a/apache-arrow/DESCR b/apache-arrow/DESCR
new file mode 100644
index 0000000000..b0b54767b1
--- /dev/null
+++ b/apache-arrow/DESCR
@@ -0,0 +1,4 @@
+iApache Arrow defines a language-independent columnar memory format for flat
+and hierarchical data, organized for efficient analytic operations on modern
+hardware like CPUs and GPUs. The Arrow memory format also supports zero-copy
+reads for lightning-fast data access without serialization overhead.
diff --git a/apache-arrow/Makefile b/apache-arrow/Makefile
new file mode 100644
index 0000000000..7e926abd44
--- /dev/null
+++ b/apache-arrow/Makefile
@@ -0,0 +1,62 @@
+# $NetBSD$
+
+DISTNAME=	apache-arrow-7.0.0
+CATEGORIES=	devel
+MASTER_SITES=	https://dlcdn.apache.org/arrow/arrow-7.0.0/
+
+
+MAINTAINER=	matthewda%zillowgroup.com@localhost
+HOMEPAGE=	https://arrow.apache.org/
+COMMENT=	Cross-language development platform for in-memory analytics
+LICENSE=	apache-2.0
+
+# These packages are built within arrow, and it
+# looks difficult to decouple them
+XSIMD=		aeec9c872c8b475dedd7781336710f2dd2666cb2.tar.gz
+JEMALLOC=	jemalloc-5.2.1.tar.bz2
+
+DISTFILES+=		${DISTNAME}${EXTRACT_SUFX}
+DISTFILES+=		${XSIMD}
+DISTFILES+=		${JEMALLOC}
+SITES.${XSIMD}=		https://github.com/xtensor-stack/xsimd/archive/
+SITES.${JEMALLOC}=	${MASTER_SITE_GITHUB:=jemalloc/}
+
+.include "../../mk/bsd.prefs.mk"
+.include "options.mk"
+
+CONFIGURE_DIRS=	cpp
+BUILD_DIRS=	cpp
+
+USE_LANGUAGES=	c c++
+USE_TOOLS=	cmake gmake
+USE_CMAKE=	yes
+CMAKE_ARGS+=	-DCMAKE_INSTALL_LIBDIR=lib
+CMAKE_ARGS+=	-DARROW_DEPENDENCY_SOURCE=SYSTEM
+# Not yet
+CMAKE_ARGS+=	-DARROW_GANDIVA=OFF
+CMAKE_ARGS+=	-DARROW_FILESYSTEM=ON
+CMAKE_ARGS+=	-DARROW_BUILD_UTILITIES=ON
+CMAKE_ARGS+=	-DARROW_CSV=ON
+
+# Set environment variable to find the extra source packages
+CONFIGURE_ENV+=	ARROW_JEMALLOC_URL=/${DISTDIR}/${JEMALLOC}
+CONFIGURE_ENV+=	ARROW_XSIMD_URL=/${DISTDIR}/${XSIMD}
+
+# To enable tests, devel/googletest needs -fPIE removed
+CMAKE_ARGS+=	-DARROW_BUILD_TESTS=OFF
+
+BUILD_DEPENDS+=	cmake-[0-9]*:../../devel/cmake
+
+.include "../../converters/utf8proc/buildlink3.mk"
+.include "../../devel/boost-libs/buildlink3.mk"
+.include "../../devel/cmake/buildlink3.mk"
+.include "../../devel/libthrift/buildlink3.mk"
+.include "../../devel/flatbuffers/buildlink3.mk"
+
+# .include "../../devel/googletest/buildlink3.mk"
+
+.include "../../devel/protobuf/buildlink3.mk"
+.include "../../devel/re2/buildlink3.mk"
+.include "../../net/grpc/buildlink3.mk"
+
+.include "../../mk/bsd.pkg.mk"
diff --git a/apache-arrow/TODO b/apache-arrow/TODO
new file mode 100644
index 0000000000..66212fcc85
--- /dev/null
+++ b/apache-arrow/TODO
@@ -0,0 +1,2 @@
+* Build python support as a separate package
+* Try and run tests?
diff --git a/apache-arrow/distinfo b/apache-arrow/distinfo
new file mode 100644
index 0000000000..df258a7ddd
--- /dev/null
+++ b/apache-arrow/distinfo
@@ -0,0 +1,11 @@
+$NetBSD$
+
+BLAKE2s (aeec9c872c8b475dedd7781336710f2dd2666cb2.tar.gz) = 55568fa5a7dc66e1e3d1d7e5196dc9ff0976ce1887b39a6de42362214b729acb
+SHA512 (aeec9c872c8b475dedd7781336710f2dd2666cb2.tar.gz) = ecc58e7909648aaa22aefbf76ee2c5a2bece4b1b88da9c7eda1b69a46247e619ff99c4366afd6015f95debd68c1a6b89292c677b9049462d0fc0255fbc0c2a33
+Size (aeec9c872c8b475dedd7781336710f2dd2666cb2.tar.gz) = 173025 bytes
+BLAKE2s (apache-arrow-7.0.0.tar.gz) = 3c14a2ca983fd0e41146d23d869c0e33fea4b1c6a39c0d7cf44efb2e9af71b14
+SHA512 (apache-arrow-7.0.0.tar.gz) = 6727ea625b1a4dc3fd452e94ef5e8fd7a5bdebcf44b105f0ba684687cf126163317caef9e91cfd3722a8defca1f0295741ed0dde2393e85d15559bddd7ecb9a2
+Size (apache-arrow-7.0.0.tar.gz) = 11949886 bytes
+BLAKE2s (jemalloc-5.2.1.tar.bz2) = a580a7fdd97dd82f1daf1b0059f229a6d34c65d76452cfc5673da1ac1a29a140
+SHA512 (jemalloc-5.2.1.tar.bz2) = 0bbb77564d767cef0c6fe1b97b705d368ddb360d55596945aea8c3ba5889fbce10479d85ad492c91d987caacdbbdccc706aa3688e321460069f00c05814fae02
+Size (jemalloc-5.2.1.tar.bz2) = 554279 bytes
diff --git a/apache-arrow/options.mk b/apache-arrow/options.mk
new file mode 100644
index 0000000000..dc16c8d70e
--- /dev/null
+++ b/apache-arrow/options.mk
@@ -0,0 +1,56 @@
+# $NetBSD: options.mk,v 1.5 2019/04/26 13:13:46 maya Exp $
+
+PKG_OPTIONS_VAR=	PKG_OPTIONS.apache-arrow
+PKG_SUPPORTED_OPTIONS=	brotli bzip2 google-log json lz4 parquet python snappy zlib zstd
+PKG_SUGGESTED_OPTIONS=	brotli bzip2 google-log json lz4 parquet python snappy zlib zstd
+
+.include "../../mk/bsd.options.mk"
+
+.if !empty(PKG_OPTIONS:Mbrotli)
+.include "../../archivers/brotli/buildlink3.mk"
+CMAKE_ARGS+=	-DARROW_WITH_BROTLI=ON
+.endif
+
+.if !empty(PKG_OPTIONS:Mbzip2)
+.include "../../archivers/bzip2/buildlink3.mk"
+CMAKE_ARGS+=	-DARROW_WITH_BZ2=ON
+.endif
+
+.if !empty(PKG_OPTIONS:Mgoogle-log)
+.include "../../devel/google-glog/buildlink3.mk"
+CMAKE_ARGS+=	-DARROW_USE_GLOG=OFF=ON
+.endif
+
+.if !empty(PKG_OPTIONS:Mjson)
+.include "../../textproc/rapidjson/buildlink3.mk"
+CMAKE_ARGS+=	-DARROW_JSON=ON
+.endif
+
+.if !empty(PKG_OPTIONS:Mlz4)
+.include "../../archivers/lz4/buildlink3.mk"
+CMAKE_ARGS+=	-DARROW_WITH_LZ4=ON
+.endif
+
+.if !empty(PKG_OPTIONS:Mparquet)
+CMAKE_ARGS+=	-DARROW_PARQUET=ON
+.endif
+
+.if !empty(PKG_OPTIONS:Mpython)
+.include "../../math/py-numpy/buildlink3.mk"
+.include "../../lang/python/pyversion.mk"
+CMAKE_ARGS+=	-DARROW_PYTHON=ON
+.endif
+.if !empty(PKG_OPTIONS:Msnappy)
+.include "../../devel/snappy/buildlink3.mk"
+CMAKE_ARGS+=	-DARROW_WITH_SNAPPY=ON
+.endif
+
+.if !empty(PKG_OPTIONS:Mzlib)
+.include "../../devel/zlib/buildlink3.mk"
+CMAKE_ARGS+=	-DARROW_WITH_ZLIB=ON
+.endif
+
+.if !empty(PKG_OPTIONS:Mzstd)
+.include "../../archivers/zstd/buildlink3.mk"
+CMAKE_ARGS+=	-DARROW_WITH_ZSTD=ON
+.endif