MAINT Adding meson compiler directives

doc/whats_new/v0.8.rst

@@ -14,6 +14,11 @@ Changelog
 ---------
 - |Fix| Trunk simulators now correctly generate random values with a fixed seed,
     by `Sambit Panda`_ (:pr:`#236`)
+- |Efficiency| All scikit-tree estimators are now at least 2X faster than they were
+    in previous versions. This was due to adding in compiler-directives to turn on 
+    optimizations '-03' when compiling the C++ generated code from Cython. In addition,
+    we explicitly turned off bounds-checking and related runtime checks in the Cython code,
+    which would lead to performance degradation during runtime. by `Adam Li`_ (:pr:`#242`)
 
 Code and Documentation Contributors
 -----------------------------------

meson.build

@@ -1,12 +1,12 @@
 project(
   'scikit-tree',
-  'c', 'cpp',
+  'c', 'cpp', 'cython',
   # Note that the git commit hash cannot be added dynamically here
   # That only happens when importing from a git repository.
   # See `sktree/__init__.py`
   version: '0.8.0dev0',
   license: 'BSD-3',
-  meson_version: '>= 0.64.0',
+  meson_version: '>= 1.1.0',
   default_options: [
     'buildtype=debugoptimized',
     'c_std=c99',
@@ -29,6 +29,20 @@ elif cc.get_id() == 'msvc'
   endif
 endif
 
+# Suppress warning for deprecated Numpy API.
+# (Suppress warning messages emitted by #warning directives).
+# Replace with numpy_nodepr_api after Cython 3.0 is out
+# '-Wno-maybe-uninitialized'
+# numpy_nodepr_api = '-DNPY_NO_DEPRECATED_API=NPY_1_7_API_VERSION'
+_global_c_args = cc.get_supported_arguments(
+  '-Wno-unused-but-set-variable',
+  '-Wno-unused-function',
+  '-Wno-conversion',
+  '-Wno-misleading-indentation',
+  '-DNPY_NO_DEPRECATED_API=NPY_1_7_API_VERSION',
+)
+add_project_arguments(_global_c_args, language : 'c')
+
 # We need -lm for all C code (assuming it uses math functions, which is safe to
 # assume for scikit-tree). For C++ it isn't needed, because libstdc++/libc++ is
 # guaranteed to depend on it. For Fortran code, Meson already adds `-lm`.
@@ -50,22 +64,22 @@ r = run_command('mv', 'sktree/_lib/sklearn_fork/sklearn', 'sktree/_lib/sklearn',
 
 # Setup Python:
 # https://mesonbuild.com/Python-module.html
-py3_mod = import('python')
+py = import('python').find_installation(pure: false)
+# py3_mod = import('python')
 
-# NOTE: with Meson >=0.64.0 we can add `pure: false` here and remove that line
-# everywhere else, see https://github.com/mesonbuild/meson/pull/10783.
-py3 = py3_mod.find_installation(
-  'python3',
-  pure: false  # Will be installed next to binaries
-)
-# py3.install_env('venv')
+# # NOTE: with Meson >=0.64.0 we can add `pure: false` here and remove that line
+# # everywhere else, see https://github.com/mesonbuild/meson/pull/10783.
+# py3 = py3_mod.find_installation(
+#   'python3',
+#   pure: false  # Will be installed next to binaries
+# )
+# py.install_env('venv')
 
 # print some debugging output
-message(py3.full_path())
-message(py3.get_install_dir())
-if py3.language_version().version_compare('<3.9')
+message(py.full_path())
+message(py.get_install_dir())
+if py.language_version().version_compare('<3.9')
     error('At least Python 3.9 is required.')
 endif
-py3_dep = py3.dependency()
 
 subdir('sktree')

sktree/__init__.py

@@ -4,6 +4,7 @@
 import os
 import sys
 
+print("using current cython branch.")
 __version__ = "0.8.0dev0"
 logger = logging.getLogger(__name__)
 

sktree/_lib/meson.build

@@ -4,20 +4,31 @@ if not fs.exists('sklearn')
 endif
 
 # install tree/ submodule
-extensions = [
-  '_tree',
-  '_utils',
-  '_criterion',
-  '_splitter',
-]
+tree_extension_metadata = {
+  '_tree':
+    {'sources': ['./sklearn/tree/' + '_tree.pyx'],
+     'override_options': ['cython_language=cpp', 'optimization=3']},
+  '_splitter':
+    {'sources': ['./sklearn/tree/' + '_splitter.pyx'],
+     'override_options': ['cython_language=cpp', 'optimization=3']},
+  '_criterion':
+    {'sources': ['./sklearn/tree/' + '_criterion.pyx'],
+     'override_options': ['cython_language=cpp', 'optimization=3']},
+  '_utils':
+    {'sources': ['./sklearn/tree/' + '_utils.pyx'],
+     'override_options': ['cython_language=cpp', 'optimization=3']},
+}
 
-foreach ext: extensions
-  py3.extension_module(ext,
-    cython_gen_cpp.process('./sklearn/tree/' + ext + '.pyx'),
-    c_args: cython_c_args,
-    include_directories: [incdir_numpy,],
-    install: true,
+
+foreach ext_name, ext_dict : tree_extension_metadata
+  py.extension_module(
+    ext_name,
+    ext_dict.get('sources'),
+    dependencies: [np_dep],
+    override_options : ext_dict.get('override_options', []),
+    cython_args: cython_c_args,
     subdir: 'sktree/_lib/sklearn/tree/',
+    install: true
   )
 endforeach
 
@@ -28,7 +39,7 @@ python_sources = [
   './sklearn/tree/_reingold_tilford.py',
 ]
 
-py3.install_sources(
+py.install_sources(
   python_sources,
   subdir: 'sktree/_lib/sklearn/tree'   # Folder relative to site-packages to install to
 )
@@ -38,7 +49,7 @@ python_sources = [
   '_forest.py',
 ]
 foreach py_source: python_sources
-  py3.install_sources(
+  py.install_sources(
     './sklearn/ensemble/' + py_source,
     subdir: 'sktree/_lib/sklearn/ensemble'
   )
@@ -51,10 +62,13 @@ extensions = [
 ]
 
 foreach ext: extensions
-  py3.extension_module(ext,
-    cython_gen_cpp.process('./sklearn/neighbors/' + ext + '.pyx'),
-    c_args: cython_c_args,
-    include_directories: [incdir_numpy,],
+  py.extension_module(
+    ext,
+    ['./sklearn/neighbors/' + ext + '.pyx'],
+    c_args: c_args,
+    dependencies: [np_dep],
+    cython_args: cython_c_args,
+    override_options : ['optimization=3', 'cython_language=cpp'],
     install: true,
     subdir: 'sktree/_lib/sklearn/neighbors/',
   )
@@ -67,10 +81,12 @@ extensions = [
 ]
 
 foreach ext: extensions
-  py3.extension_module(ext,
-    cython_gen_cpp.process('./sklearn/utils/' + ext + '.pyx'),
-    c_args: cython_c_args,
-    include_directories: [incdir_numpy,],
+  py.extension_module(ext,
+    ['./sklearn/utils/' + ext + '.pyx'],
+    c_args: c_args,
+    dependencies: [np_dep],
+    cython_args: cython_c_args,
+    override_options : ['optimization=3', 'cython_language=cpp'],
     install: true,
     subdir: 'sktree/_lib/sklearn/utils/',
   )

sktree/datasets/meson.build

@@ -4,7 +4,7 @@ python_sources = [
   'hyppo.py',
 ]
 
-py3.install_sources(
+py.install_sources(
   python_sources,
   pure: false,
   subdir: 'sktree/datasets'

sktree/datasets/tests/meson.build

@@ -4,7 +4,7 @@ python_sources = [
   'test_multiview.py',
 ]
 
-py3.install_sources(
+py.install_sources(
   python_sources,
   pure: false,
   subdir: 'sktree/datasets/tests'

sktree/ensemble/meson.build

@@ -8,7 +8,7 @@ python_sources = [
   '_extensions.py',
 ]
 
-py3.install_sources(
+py.install_sources(
   python_sources,
   pure: false,
   subdir: 'sktree/ensemble'

sktree/experimental/meson.build

@@ -6,7 +6,7 @@ python_sources = [
   'monte_carlo.py',
 ]
 
-py3.install_sources(
+py.install_sources(
   python_sources,
   pure: false,
   subdir: 'sktree/experimental'

sktree/experimental/tests/meson.build

@@ -6,7 +6,7 @@ python_sources = [
   'test_monte_carlo.py',
 ]
 
-py3.install_sources(
+py.install_sources(
   python_sources,
   pure: false,
   subdir: 'sktree/experimental/tests'