Merge pull request #38440 from veprbl/arrow_cpp_reopen

Add arrow-cpp, parquet-cpp, pythonPackages.pyarrow

pkgs/development/libraries/arrow-cpp/default.nix

@@ -0,0 +1,44 @@
+{ stdenv, symlinkJoin, fetchurl, boost, brotli, cmake, flatbuffers, gtest, gflags, lz4, python, rapidjson, snappy, zlib, zstd }:
+
+stdenv.mkDerivation rec {
+  name = "arrow-cpp-${version}";
+  version = "0.9.0";
+
+  src = fetchurl {
+    url = "mirror://apache/arrow/arrow-${version}/apache-arrow-${version}.tar.gz";
+    sha256 = "16l91fixb5dgx3v6xc73ipn1w1hjgbmijyvs81j7ywzpna2cdcdy";
+  };
+
+  sourceRoot = "apache-arrow-${version}/cpp";
+
+  nativeBuildInputs = [ cmake ];
+  buildInputs = [ boost python.pkgs.python python.pkgs.numpy ];
+
+  preConfigure = ''
+    substituteInPlace cmake_modules/FindBrotli.cmake --replace CMAKE_STATIC_LIBRARY CMAKE_SHARED_LIBRARY
+    substituteInPlace cmake_modules/FindLz4.cmake --replace CMAKE_STATIC_LIBRARY CMAKE_SHARED_LIBRARY
+    substituteInPlace cmake_modules/FindSnappy.cmake --replace CMAKE_STATIC_LIBRARY CMAKE_SHARED_LIBRARY
+  '';
+
+  BROTLI_HOME = symlinkJoin { name="brotli-wrap"; paths = [ brotli.lib brotli.dev ]; };
+  FLATBUFFERS_HOME = flatbuffers;
+  GTEST_HOME = gtest;
+  GFLAGS_HOME = gflags;
+  LZ4_HOME = symlinkJoin { name="lz4-wrap"; paths = [ lz4 lz4.dev ]; };
+  RAPIDJSON_HOME = rapidjson;
+  SNAPPY_HOME = symlinkJoin { name="snappy-wrap"; paths = [ snappy snappy.dev ]; };
+  ZLIB_HOME = symlinkJoin { name="zlib-wrap"; paths = [ zlib.dev zlib.static ]; };
+  ZSTD_HOME = zstd;
+
+  cmakeFlags = [
+    "-DARROW_PYTHON=ON"
+  ];
+
+  meta = {
+    description = "A  cross-language development platform for in-memory data";
+    homepage = https://arrow.apache.org/;
+    license = stdenv.lib.licenses.asl20;
+    platforms = stdenv.lib.platforms.unix;
+    maintainers = with stdenv.lib.maintainers; [ veprbl ];
+  };
+}

pkgs/development/libraries/parquet-cpp/default.nix

@@ -0,0 +1,37 @@
+{ stdenv, symlinkJoin, fetchurl, arrow-cpp, boost, cmake, gtest, snappy, thrift, zlib }:
+
+stdenv.mkDerivation rec {
+  name = "parquet-cpp-${version}";
+  version = "1.4.0";
+
+  src = fetchurl {
+    url = "https://github.com/apache/parquet-cpp/archive/apache-${name}.tar.gz";
+    sha256 = "1kn7pjzi5san5f05qbl8l8znqsa3f9cq9bflfr4s2jfwr7k9p2aj";
+  };
+
+  nativeBuildInputs = [ cmake ];
+  buildInputs = [ boost ];
+
+  preConfigure = ''
+    substituteInPlace cmake_modules/FindThrift.cmake --replace CMAKE_STATIC_LIBRARY CMAKE_SHARED_LIBRARY
+    substituteInPlace cmake_modules/FindSnappy.cmake --replace CMAKE_STATIC_LIBRARY CMAKE_SHARED_LIBRARY
+  '';
+
+  ARROW_HOME = arrow-cpp;
+  THRIFT_HOME = thrift;
+  GTEST_HOME = gtest;
+  SNAPPY_HOME = symlinkJoin { name="snappy-wrap"; paths = [ snappy snappy.dev ]; };
+  ZLIB_HOME = symlinkJoin { name="zlib-wrap"; paths = [ zlib.dev zlib.static ]; };
+
+  cmakeFlags = [
+    "-DPARQUET_BUILD_BENCHMARKS=OFF"
+  ];
+
+  meta = {
+    description = "A C++ library to read and write the Apache Parquet columnar data format";
+    homepage = http://parquet.apache.org;
+    license = stdenv.lib.licenses.asl20;
+    platforms = stdenv.lib.platforms.unix;
+    maintainers = with stdenv.lib.maintainers; [ veprbl ];
+  };
+}

pkgs/development/libraries/rapidjson/default.nix

@@ -13,14 +13,16 @@ stdenv.mkDerivation rec {
 
   nativeBuildInputs = [ pkgconfig cmake ];
 
-  # detected by gcc7
-  NIX_CFLAGS_COMPILE = [ "-Wno-error=implicit-fallthrough" ];
+  preConfigure = ''
+    substituteInPlace CMakeLists.txt --replace "-Werror" ""
+    substituteInPlace example/CMakeLists.txt --replace "-Werror" ""
+  '';
 
   meta = with lib; {
     description = "Fast JSON parser/generator for C++ with both SAX/DOM style API";
     homepage = "http://rapidjson.org/";
     license = licenses.mit;
-    platforms = platforms.linux;
+    platforms = platforms.unix;
     maintainers = with maintainers; [ cstrahan ];
   };
 }

pkgs/development/python-modules/pyarrow/default.nix

@@ -0,0 +1,65 @@
+{ lib, buildPythonPackage, python, isPy3k, fetchurl, arrow-cpp, cmake, cython, futures, numpy, pandas, pytest, pytestrunner, parquet-cpp, pkgconfig, setuptools_scm, six }:
+
+let
+  _arrow-cpp = arrow-cpp.override { inherit python;};
+  _parquet-cpp = parquet-cpp.override { arrow-cpp = _arrow-cpp; };
+in
+
+buildPythonPackage rec {
+  pname = "pyarrow";
+  version = "0.9.0";
+
+  src = fetchurl {
+    url = "mirror://apache/arrow/arrow-${version}/apache-arrow-${version}.tar.gz";
+    sha256 = "16l91fixb5dgx3v6xc73ipn1w1hjgbmijyvs81j7ywzpna2cdcdy";
+  };
+
+  sourceRoot = "apache-arrow-${version}/python";
+
+  nativeBuildInputs = [ cmake cython pkgconfig setuptools_scm ];
+  propagatedBuildInputs = [ numpy six ] ++ lib.optionals (!isPy3k) [ futures ];
+  checkInputs = [ pandas pytest pytestrunner ];
+
+  PYARROW_BUILD_TYPE = "release";
+  PYARROW_CMAKE_OPTIONS = "-DCMAKE_INSTALL_RPATH=${ARROW_HOME}/lib;${PARQUET_HOME}/lib";
+
+  preBuild = ''
+    substituteInPlace CMakeLists.txt --replace "\''${ARROW_ABI_VERSION}" '"0.0.0"'
+    substituteInPlace CMakeLists.txt --replace "\''${ARROW_SO_VERSION}" '"0"'
+
+    # fix the hardcoded value
+    substituteInPlace cmake_modules/FindParquet.cmake --replace 'set(PARQUET_ABI_VERSION "1.0.0")' 'set(PARQUET_ABI_VERSION "${_parquet-cpp.version}")'
+  '';
+
+  preCheck = ''
+    rm pyarrow/tests/test_hdfs.py
+
+    # fails: "ArrowNotImplementedError: Unsupported numpy type 22"
+    substituteInPlace pyarrow/tests/test_feather.py --replace "test_timedelta_with_nulls" "_disabled"
+
+    # runs out of memory on @grahamcofborg linux box
+    substituteInPlace pyarrow/tests/test_feather.py --replace "test_large_dataframe" "_disabled"
+
+    # probably broken on python2
+    substituteInPlace pyarrow/tests/test_feather.py --replace "test_unicode_filename" "_disabled"
+
+    # fails "error: [Errno 2] No such file or directory: 'test'" because
+    # nix_run_setup invocation somehow manages to import deserialize_buffer.py
+    # when it is not intended to be imported at all
+    rm pyarrow/tests/deserialize_buffer.py
+    substituteInPlace pyarrow/tests/test_feather.py --replace "test_deserialize_buffer_in_different_process" "_disabled"
+  '';
+
+  ARROW_HOME = _arrow-cpp;
+  PARQUET_HOME = _parquet-cpp;
+
+  setupPyBuildFlags = ["--with-parquet" ];
+
+  meta = with lib; {
+    description = "A cross-language development platform for in-memory data";
+    homepage = https://arrow.apache.org/;
+    license = lib.licenses.asl20;
+    platforms = platforms.unix;
+    maintainers = with lib.maintainers; [ veprbl ];
+  };
+}

pkgs/top-level/all-packages.nix

@@ -8405,6 +8405,8 @@ with pkgs;
 
   armadillo = callPackage ../development/libraries/armadillo {};
 
+  arrow-cpp = callPackage ../development/libraries/arrow-cpp {};
+
   assimp = callPackage ../development/libraries/assimp { };
 
   asio = callPackage ../development/libraries/asio { };
@@ -10876,6 +10878,8 @@ with pkgs;
 
   paperkey = callPackage ../tools/security/paperkey { };
 
+  parquet-cpp = callPackage ../development/libraries/parquet-cpp {};
+
   pangoxsl = callPackage ../development/libraries/pangoxsl { };
 
   pcaudiolib = callPackage ../development/libraries/pcaudiolib {

pkgs/top-level/python-packages.nix

@@ -311,6 +311,10 @@ in {
 
   pyamf = callPackage ../development/python-modules/pyamf { };
 
+  pyarrow = callPackage ../development/python-modules/pyarrow {
+    inherit (pkgs) arrow-cpp cmake pkgconfig;
+  };
+
   pyatspi = disabledIf (!isPy3k) (callPackage ../development/python-modules/pyatspi { });
 
   pyaxmlparser = callPackage ../development/python-modules/pyaxmlparser { };