ARROW-3075: [C++] Incorporate parquet-cpp codebase into Arrow C++ build

system. Add unit test label granularity options, ability to add component group targets like 'make parquet' that build libraries and tests
apache · Sep 6, 2018 · 23093d7 · 23093d7
1 parent e5662a3
commit 23093d7
Show file tree

Hide file tree

Showing 92 changed files with 852 additions and 868 deletions.
diff --git a/.gitmodules b/.gitmodules
@@ -0,0 +1,3 @@
+[submodule "cpp/submodules/parquet-testing"]
+	path = cpp/submodules/parquet-testing
+	url = https://github.com/apache/parquet-testing.git
diff --git a/.travis.yml b/.travis.yml
@@ -60,6 +60,7 @@ matrix:
     - ARROW_TRAVIS_ORC=1
     - ARROW_TRAVIS_CLANG_FORMAT=1
     - ARROW_TRAVIS_COVERAGE=1
+    - ARROW_TRAVIS_PARQUET=1
     - ARROW_TRAVIS_PYTHON_DOCS=1
     - ARROW_BUILD_WARNING_LEVEL=CHECKIN
     - ARROW_TRAVIS_PYTHON_JVM=1
@@ -77,11 +78,11 @@ matrix:
     - $TRAVIS_BUILD_DIR/ci/travis_install_clang_tools.sh
     - $TRAVIS_BUILD_DIR/ci/travis_lint.sh
     # If either C++ or Python changed, we must install the C++ libraries
+    - git submodule update --init
     - $TRAVIS_BUILD_DIR/ci/travis_before_script_cpp.sh
     script:
     # All test steps are required for accurate C++ coverage info
     - $TRAVIS_BUILD_DIR/ci/travis_script_cpp.sh
-    - $TRAVIS_BUILD_DIR/ci/travis_build_parquet_cpp.sh
     # Build Arrow Java to test the pyarrow<->JVM in-process bridge
     - $TRAVIS_BUILD_DIR/ci/travis_script_java.sh
     # Only run Plasma tests with valgrind in one of the Python builds because
@@ -102,14 +103,15 @@ matrix:
     - ARROW_TRAVIS_USE_TOOLCHAIN=1
     - ARROW_TRAVIS_PLASMA=1
     - ARROW_TRAVIS_ORC=1
+    - ARROW_TRAVIS_PARQUET=1
     - ARROW_BUILD_WARNING_LEVEL=CHECKIN
     before_script:
     - if [ $ARROW_CI_PYTHON_AFFECTED != "1" ]; then exit; fi
     # If either C++ or Python changed, we must install the C++ libraries
+    - git submodule update --init
     - $TRAVIS_BUILD_DIR/ci/travis_before_script_cpp.sh
     script:
     - if [ $ARROW_CI_CPP_AFFECTED == "1" ]; then $TRAVIS_BUILD_DIR/ci/travis_script_cpp.sh; fi
-    - $TRAVIS_BUILD_DIR/ci/travis_build_parquet_cpp.sh
     - $TRAVIS_BUILD_DIR/ci/travis_script_python.sh 2.7
     - $TRAVIS_BUILD_DIR/ci/travis_script_python.sh 3.6
   # [manylinux1] Python

diff --git a/appveyor.yml b/appveyor.yml
@@ -48,7 +48,6 @@ environment:
       GENERATOR: Ninja
       CONFIGURATION: "Release"
       APPVEYOR_BUILD_WORKER_IMAGE: Visual Studio 2017
-      BOOST_ROOT: C:\Libraries\boost_1_64_0
     - JOB: "Toolchain"
       GENERATOR: Visual Studio 14 2015 Win64
       CONFIGURATION: "Release"
@@ -67,8 +66,8 @@ environment:
       USE_CLCACHE: false
 
   MSVC_DEFAULT_OPTIONS: ON
-  BOOST_ROOT: C:\Libraries\boost_1_63_0
-  BOOST_LIBRARYDIR: C:\Libraries\boost_1_63_0\lib64-msvc-14.0
+  BOOST_ROOT: C:\Libraries\boost_1_67_0
+  BOOST_LIBRARYDIR: C:\Libraries\boost_1_67_0\lib64-msvc-14.0
   APPVEYOR_SAVE_CACHE_ON_ERROR: true
 
 install:

diff --git a/ci/cpp-python-msvc-build.bat b/ci/cpp-python-msvc-build.bat
@@ -68,16 +68,20 @@ if "%JOB%" == "Build_Debug" (
   exit /B 0
 )
 
-conda create -n arrow -q -y python=%PYTHON% ^
+conda create -n arrow -q -y -c conda-forge ^
+      python=%PYTHON% ^
       six pytest setuptools numpy pandas cython ^
-      thrift-cpp=0.11.0
+      thrift-cpp=0.11.0 boost-cpp
 
 call activate arrow
 
+@rem Use Boost from conda-forge
+set BOOST_ROOT=%CONDA_PREFIX%\Library
+set BOOST_LIBRARYDIR=%CONDA_PREFIX%\Library\lib
+
 if "%JOB%" == "Toolchain" (
   @rem Install pre-built "toolchain" packages for faster builds
   conda install -q -y -c conda-forge ^
-      boost-cpp ^
       brotli ^
       cmake ^
       flatbuffers ^
@@ -94,6 +98,10 @@ if "%JOB%" == "Toolchain" (
 
 set ARROW_HOME=%CONDA_PREFIX%\Library
 
+@rem Retrieve git submodules, configure env var for Parquet unit tests
+git submodule update --init
+set PARQUET_TEST_DATA=%CD%\cpp\submodules\parquet-testing\data
+
 @rem Build and test Arrow C++ libraries
 
 mkdir cpp\build
@@ -104,6 +112,7 @@ cmake -G "%GENERATOR%" ^
       -DARROW_BOOST_USE_SHARED=OFF ^
       -DCMAKE_BUILD_TYPE=%CONFIGURATION% ^
       -DARROW_CXXFLAGS="/WX /MP" ^
+      -DARROW_PARQUET=ON ^
       -DARROW_PYTHON=ON ^
       ..  || exit /B
 cmake --build . --target install --config %CONFIGURATION%  || exit /B
@@ -117,26 +126,7 @@ ctest -VV  || exit /B
 set PYTHONHOME=%OLD_PYTHONHOME%
 popd
 
-@rem Build parquet-cpp
-
-git clone https://github.com/apache/parquet-cpp.git || exit /B
-mkdir parquet-cpp\build
-pushd parquet-cpp\build
-
-set PARQUET_BUILD_TOOLCHAIN=%CONDA_PREFIX%\Library
-set PARQUET_HOME=%CONDA_PREFIX%\Library
-cmake -G "%GENERATOR%" ^
-     -DCMAKE_INSTALL_PREFIX=%PARQUET_HOME% ^
-     -DCMAKE_BUILD_TYPE=%CONFIGURATION% ^
-     -DPARQUET_BOOST_USE_SHARED=OFF ^
-     -DPARQUET_BUILD_TESTS=OFF ^
-     .. || exit /B
-cmake --build . --target install --config %CONFIGURATION% || exit /B
-popd
-
 @rem Build and install pyarrow
-@rem parquet-cpp has some additional runtime dependencies that we need to figure out
-@rem see PARQUET-1018
 
 pushd python
 

diff --git a/ci/travis_before_script_cpp.sh b/ci/travis_before_script_cpp.sh
@@ -82,6 +82,10 @@ if [ $ARROW_TRAVIS_ORC == "1" ]; then
   CMAKE_COMMON_FLAGS="$CMAKE_COMMON_FLAGS -DARROW_ORC=ON"
 fi
 
+if [ $ARROW_TRAVIS_PARQUET == "1" ]; then
+  CMAKE_COMMON_FLAGS="$CMAKE_COMMON_FLAGS -DARROW_PARQUET=ON"
+fi
+
 if [ $ARROW_TRAVIS_VALGRIND == "1" ]; then
   CMAKE_COMMON_FLAGS="$CMAKE_COMMON_FLAGS -DARROW_TEST_MEMCHECK=ON"
 fi

diff --git a/ci/travis_build_parquet_cpp.sh b/ci/travis_build_parquet_cpp.sh
diff --git a/ci/travis_env_common.sh b/ci/travis_env_common.sh
@@ -44,8 +44,6 @@ export ARROW_CPP_BUILD_DIR=$TRAVIS_BUILD_DIR/cpp-build
 export ARROW_C_GLIB_INSTALL_AUTOTOOLS=$TRAVIS_BUILD_DIR/c-glib-install-autotools
 export ARROW_C_GLIB_INSTALL_MESON=$TRAVIS_BUILD_DIR/c-glib-install-meson
 
-export ARROW_PYTHON_PARQUET_HOME=$TRAVIS_BUILD_DIR/parquet-env
-
 export CMAKE_EXPORT_COMPILE_COMMANDS=1
 
 export ARROW_BUILD_TYPE=${ARROW_BUILD_TYPE:=debug}
@@ -70,3 +68,5 @@ fi
 if [ $TRAVIS_OS_NAME == "osx" ]; then
   export GOPATH=$TRAVIS_BUILD_DIR/gopath
 fi
+
+export PARQUET_TEST_DATA=$TRAVIS_BUILD_DIR/cpp/submodules/parquet-testing/data
diff --git a/ci/travis_install_toolchain.sh b/ci/travis_install_toolchain.sh
@@ -40,8 +40,4 @@ if [ ! -e $CPP_TOOLCHAIN ]; then
         thrift-cpp=0.11.0 \
         zlib \
         zstd
-
-    # HACK(wesm): We started experiencing OpenSSL failures when Miniconda was
-    # updated sometime on October 2 or October 3
-#     conda update -y -q -p $CPP_TOOLCHAIN ca-certificates -c defaults
 fi
diff --git a/ci/travis_script_python.sh b/ci/travis_script_python.sh
@@ -23,8 +23,8 @@ source $TRAVIS_BUILD_DIR/ci/travis_env_common.sh
 source $TRAVIS_BUILD_DIR/ci/travis_install_conda.sh
 
 export ARROW_HOME=$ARROW_CPP_INSTALL
-export PARQUET_HOME=$ARROW_PYTHON_PARQUET_HOME
-export LD_LIBRARY_PATH=$ARROW_HOME/lib:$PARQUET_HOME/lib:$LD_LIBRARY_PATH
+export PARQUET_HOME=$ARROW_CPP_INSTALL
+export LD_LIBRARY_PATH=$ARROW_HOME/lib:$LD_LIBRARY_PATH
 export PYARROW_CXXFLAGS="-Werror"
 
 PYARROW_PYTEST_FLAGS=" -r sxX --durations=15 --parquet"

diff --git a/cpp/CMakeLists.txt b/cpp/CMakeLists.txt
@@ -275,6 +275,20 @@ Pass multiple labels by dividing with semicolons")
       "Build Arrow with statically linked CRT"
       OFF)
   endif()
+
+  # Parquet-related build options
+  option(ARROW_PARQUET
+    "Build the Parquet libraries"
+    OFF)
+
+  option(PARQUET_MINIMAL_DEPENDENCY
+    "Depend only on Thirdparty headers to build libparquet. \
+Always OFF if building binaries"
+    OFF)
+
+  set(PARQUET_ARROW_LINKAGE "shared" CACHE STRING
+    "How to link Arrow libraries with libparquet.so. static|shared (default shared)")
+
 endif()
 
 if(ARROW_BUILD_TESTS OR ARROW_BUILD_BENCHMARKS)
@@ -584,6 +598,8 @@ endif(UNIX)
 ############################################################
 
 set(ARROW_LINK_LIBS)
+
+# Libraries to link statically with libarrow.so
 set(ARROW_STATIC_LINK_LIBS)
 
 if (ARROW_WITH_BROTLI)
@@ -643,11 +659,8 @@ set(ARROW_BENCHMARK_LINK_LIBS
   gtest
   ${ARROW_STATIC_LINK_LIBS})
 
-set(ARROW_LINK_LIBS
-  ${ARROW_STATIC_LINK_LIBS}
-  ${ARROW_LINK_LIBS})
-
 set(ARROW_SHARED_PRIVATE_LINK_LIBS
+  ${ARROW_STATIC_LINK_LIBS}
   ${BOOST_SYSTEM_LIBRARY}
   ${BOOST_FILESYSTEM_LIBRARY}
   ${BOOST_REGEX_LIBRARY})
@@ -729,3 +742,7 @@ endif()
 if(ARROW_HIVESERVER2)
   add_subdirectory(src/arrow/dbi/hiveserver2)
 endif()
+
+if(ARROW_PARQUET)
+  add_subdirectory(src/parquet)
+endif()
diff --git a/cpp/build-support/lint_cpp_cli.py b/cpp/build-support/lint_cpp_cli.py
@@ -19,6 +19,8 @@
 import argparse
 import re
 import os
+import sys
+import traceback
 
 parser = argparse.ArgumentParser(
     description="Check for illegal headers for C++/CLI applications")
@@ -59,23 +61,29 @@ def lint_file(path):
     'arrow/util/macros.h',
     'arrow/python/iterators.h',
     'arrow/util/parallel.h',
-    'arrow/io/hdfs-internal.h'
+    'arrow/io/hdfs-internal.h',
+    'parquet/arrow/test-util.h',
+    'parquet/encoding-internal.h',
+    'parquet/test-util.h'
 ]
 
-
-for dirpath, _, filenames in os.walk(arguments.source_path):
-    for filename in filenames:
-        full_path = os.path.join(dirpath, filename)
-
-        exclude = False
-        for exclusion in EXCLUSIONS:
-            if exclusion in full_path:
-                exclude = True
-                break
-
-        if exclude:
-            continue
-
-        # Only run on header files
-        if filename.endswith('.h'):
-            lint_file(full_path)
+try:
+    for dirpath, _, filenames in os.walk(arguments.source_path):
+        for filename in filenames:
+            full_path = os.path.join(dirpath, filename)
+
+            exclude = False
+            for exclusion in EXCLUSIONS:
+                if exclusion in full_path:
+                    exclude = True
+                    break
+
+            if exclude:
+                continue
+
+            # Only run on header files
+            if filename.endswith('.h'):
+                lint_file(full_path)
+except Exception:
+    traceback.print_exc()
+    sys.exit(1)