[REVIEW] Adding Cython to Code Coverage (#3111)

* Adding ability to build with --linetrace=1 to support cython codecov * Adding PR to CHANGELOG * Style cleanup * Converting BUILD_PYTHON_ARGS to be a argument in build.sh
rapidsai · Nov 6, 2020 · 25b29f5 · 25b29f5
1 parent 93eef64
commit 25b29f5
Show file tree

Hide file tree

Showing 6 changed files with 100 additions and 32 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -22,6 +22,7 @@
 - PR #3052: Speeding up MNMG KNN Cl&Re testing
 - PR #3115: Speeding up MNMG UMAP testing
 - PR #3112: Speed test_array
+- PR #3111: Adding Cython to Code Coverage
 
 ## Bug Fixes
 - PR #3065: Refactoring prims metrics function names from camelcase to underscore format

diff --git a/build.sh b/build.sh
@@ -19,7 +19,7 @@ ARGS=$*
 REPODIR=$(cd $(dirname $0); pwd)
 
 VALIDTARGETS="clean libcuml cuml cpp-mgtests prims bench prims-bench cppdocs pydocs"
-VALIDFLAGS="-v -g -n --allgpuarch --buildfaiss --buildgtest --singlegpu --nvtx --show_depr_warn -h --help "
+VALIDFLAGS="-v -g -n --allgpuarch --buildfaiss --buildgtest --singlegpu --nvtx --show_depr_warn --codecov -h --help "
 VALIDARGS="${VALIDTARGETS} ${VALIDFLAGS}"
 HELP="$0 [<target> ...] [<flag> ...]
  where <target> is:
@@ -43,6 +43,8 @@ HELP="$0 [<target> ...] [<flag> ...]
    --singlegpu      - Build libcuml and cuml without multigpu components
    --nvtx           - Enable nvtx for profiling support
    --show_depr_warn - show cmake deprecation warnings
+   --codecov        - Enable code coverage support by compiling with Cython linetracing
+                      and profiling enabled (WARNING: Impacts performance)
    -h               - print this text
 
  default action (no args) is to build and install 'libcuml', 'cuml', and 'prims' targets only for the detected GPU arch
@@ -58,7 +60,7 @@ BUILD_TYPE=Release
 INSTALL_TARGET=install
 BUILD_ALL_GPU_ARCH=0
 SINGLEGPU_CPP_FLAG=""
-SINGLEGPU_PYTHON_FLAG=""
+BUILD_PYTHON_ARGS=${BUILD_PYTHON_ARGS:=""}
 NVTX=OFF
 CLEAN=0
 BUILD_DISABLE_DEPRECATION_WARNING=ON
@@ -115,7 +117,7 @@ if hasArg --allgpuarch; then
     BUILD_ALL_GPU_ARCH=1
 fi
 if hasArg --singlegpu; then
-    SINGLEGPU_PYTHON_FLAG="--singlegpu"
+    BUILD_PYTHON_ARGS="${BUILD_PYTHON_ARGS} --singlegpu"
     SINGLEGPU_CPP_FLAG=ON
 fi
 if hasArg cpp-mgtests; then
@@ -133,6 +135,9 @@ fi
 if hasArg --show_depr_warn; then
     BUILD_DISABLE_DEPRECATION_WARNING=OFF
 fi
+if hasArg --codecov; then
+    BUILD_PYTHON_ARGS="${BUILD_PYTHON_ARGS} --linetrace=1 --profile"
+fi
 if hasArg clean; then
     CLEAN=1
 fi
@@ -224,9 +229,9 @@ fi
 if completeBuild || hasArg cuml || hasArg pydocs; then
     cd ${REPODIR}/python
     if [[ ${INSTALL_TARGET} != "" ]]; then
-        python setup.py build_ext -j${PARALLEL_LEVEL:-1} ${SINGLEGPU_PYTHON_FLAG} --library-dir=${LIBCUML_BUILD_DIR} install --single-version-externally-managed --record=record.txt
+        python setup.py build_ext -j${PARALLEL_LEVEL:-1} ${BUILD_PYTHON_ARGS} --library-dir=${LIBCUML_BUILD_DIR} install --single-version-externally-managed --record=record.txt
     else
-        python setup.py build_ext -j${PARALLEL_LEVEL:-1} --library-dir=${LIBCUML_BUILD_DIR} ${SINGLEGPU_PYTHON_FLAG}
+        python setup.py build_ext -j${PARALLEL_LEVEL:-1} ${BUILD_PYTHON_ARGS} --library-dir=${LIBCUML_BUILD_DIR}
     fi
 
     if hasArg pydocs; then

diff --git a/ci/gpu/build.sh b/ci/gpu/build.sh
@@ -98,7 +98,7 @@ if [[ -z "$PROJECT_FLASH" || "$PROJECT_FLASH" == "0" ]]; then
     ################################################################################
 
     gpuci_logger "Build from source"
-    $WORKSPACE/build.sh clean libcuml cuml prims bench -v
+    $WORKSPACE/build.sh clean libcuml cuml prims bench -v --codecov
 
     gpuci_logger "Resetting LD_LIBRARY_PATH"
 
@@ -190,7 +190,7 @@ else
     conda install -c $WORKSPACE/ci/artifacts/cuml/cpu/conda-bld/ libcuml
 
     gpuci_logger "Building cuml"
-    "$WORKSPACE/build.sh" -v cuml
+    "$WORKSPACE/build.sh" -v cuml --codecov
 
     gpuci_logger "Python pytest for cuml"
     cd $WORKSPACE/python

diff --git a/python/.coveragerc b/python/.coveragerc
@@ -2,3 +2,4 @@
 [run]
 include = cuml/*
 omit = cuml/test/*
+plugins = Cython.Coverage
diff --git a/python/cython_build_ext.py b/python/cython_build_ext.py
@@ -46,10 +46,15 @@ class cython_build_ext(_build_ext, object):
 
     Parameters
     ----------
-    language_level : {"2", "3", "3str"}, default="2"
-        Globally set the Python language level to be used for module
-        compilation. Default is compatibility with Python 2. To enable Python 3
-        source code semantics, set this to 3 (or 3str)
+    annotate : bool, default=False
+        If True, will produce a HTML file for each of the .pyx or .py files
+        compiled. The HTML file gives an indication of how much Python
+        interaction there is in each of the source code lines, compared to
+        plain C code. It also allows you to see the C/C++ code generated for
+        each line of Cython code. This report is invaluable when optimizing a
+        function for speed, and for determining when to release the GIL: in
+        general, a nogil block may contain only “white” code. See examples in
+        Determining where to add types or Primes.
     binding : bool, default=True
         Controls whether free functions behave more like Python’s CFunctions
         (e.g. len()) or, when set to True, more like Python’s functions. When
@@ -59,46 +64,74 @@ class cython_build_ext(_build_ext, object):
         annotations.
 
         Changed in version 3.0.0: Default changed from False to True
-    profile : bool, default=False
-        Write hooks for Python profilers into the compiled C code.
+    cython_exclude : list of str
+        When passing glob patterns as module_list, you can exclude certain
+        module names explicitly by passing them into the exclude option.
     embedsignature : bool, default=False
         If set to True, Cython will embed a textual copy of the call signature
         in the docstring of all Python visible functions and classes. Tools
         like IPython and epydoc can thus display the signature, which cannot
         otherwise be retrieved after compilation.
-    cython_exclude : list of str
-        When passing glob patterns as module_list, you can exclude certain
-        module names explicitly by passing them into the exclude option.
     gdb_debug : bool, default=False
         Passes the `gdb_debug` argument to `cythonize()`. Setting up debugging
         for Cython can be difficult. See the debugging docs here
         https://cython.readthedocs.io/en/latest/src/userguide/debugging.html
+    language_level : {"2", "3", "3str"}, default="2"
+        Globally set the Python language level to be used for module
+        compilation. Default is compatibility with Python 2. To enable Python 3
+        source code semantics, set this to 3 (or 3str)
+    linetrace : bool, default=False
+        Write line tracing hooks for Python profilers or coverage reporting
+        into the compiled C code. This also enables profiling. Default is
+        False. Note that the generated module will not actually use line
+        tracing, unless you additionally pass the C macro definition
+        ``CYTHON_TRACE=1`` to the C compiler (e.g. using the setuptools option
+        define_macros). Define ``CYTHON_TRACE_NOGIL=1`` to also include nogil
+        functions and sections.
+    profile : bool, default=False
+        Write hooks for Python profilers into the compiled C code.
     """
     user_options = [
-        ('language-level=', None,
-         'Sets the python language syntax to use "2", "3", "3str".'),
-        ("binding", None,
+        ("annotate=",
+         None,
+         "Passes the `annotate` argument to `cythonize()`. See the Cython "
+         "docs for more info."),
+        ("binding",
+         None,
          "Sets the binding Cython compiler directive. See the Cython docs for "
          "more info."),
-        ("profile", None,
-         "Sets the profile Cython compiler directive. See the Cython docs for "
-         "more info."),
-        ("embedsignature", None,
-         "Sets the `embedsignature` Cython compiler directive. See the Cython "
-         "docs for more info."),
-        ("cython-exclude=", None,
+        ("cython-exclude=",
+         None,
          "Sets the exclude argument for `cythonize()`. See the Cython docs for"
          " more info."),
-        ("gdb-debug=", None,
+        ("embedsignature",
+         None,
+         "Sets the `embedsignature` Cython compiler directive. See the Cython "
+         "docs for more info."),
+        ("gdb-debug=",
+         None,
          "Passes the `gdb_debug` argument to `cythonize()`. See the Cython "
-         "docs for more info.")
+         "docs for more info."),
+        ('language-level=',
+         None,
+         'Sets the python language syntax to use "2", "3", "3str".'),
+        ("linetrace=",
+         None,
+         "Passes the `linetrace` argument to `cythonize()`. See the Cython "
+         "docs for more info."),
+        ("profile",
+         None,
+         "Sets the profile Cython compiler directive. See the Cython docs for "
+         "more info."),
     ] + _build_ext.user_options
 
     boolean_options = [
+        "annotate",
         "binding",
-        "profile",
         "embedsignature",
         "gdb-debug",
+        "linetrace",
+        "profile",
     ] + _build_ext.boolean_options
 
     def initialize_options(self):
@@ -107,12 +140,15 @@ def initialize_options(self):
         detect if they were set by the user
         """
 
-        self.language_level = None
+        self.annotate = None
         self.binding = None
-        self.profile = None
-        self.embedsignature = None
         self.cython_exclude = None
+        self.embedsignature = None
         self.gdb_debug = None
+        self.language_level = None
+        self.linetrace = None
+        self.profile = None
+
         super().initialize_options()
 
     def finalize_options(self):
@@ -153,13 +189,37 @@ def finalize_options(self):
                 self.profile = bool(self.profile)
                 compiler_directives.update({"profile": self.profile})
 
+            if (self.linetrace is not None):
+                self.linetrace = bool(self.linetrace)
+                compiler_directives.update({"linetrace": self.linetrace})
+
+                # Also need the compiler directive. Only add if it hasnt been
+                # specified yet
+                for ext in self.distribution.ext_modules:
+                    if (not hasattr(ext, "define_macros")):
+                        ext.define_macros = []
+
+                    found_macro = False
+
+                    for mac in ext.define_macros:
+                        if (mac[0] == "CYTHON_TRACE_NOGIL"):
+                            found_macro = True
+                            break
+
+                    if not found_macro:
+                        ext.define_macros.append(("CYTHON_TRACE_NOGIL", 1))
+
             if (self.embedsignature is not None):
                 self.embedsignature = bool(self.embedsignature)
                 compiler_directives.update(
                     {"embedsignature": self.embedsignature})
 
             cythonize_kwargs = {}
 
+            if (self.annotate is not None):
+
+                cythonize_kwargs.update({"annotate": self.annotate})
+
             if (self.cython_exclude is not None):
 
                 if (isinstance(self.cython_exclude, str)):

diff --git a/python/setup.cfg b/python/setup.cfg
@@ -23,4 +23,5 @@ inplace = True
 binding = True
 language_level = 3
 profile = False
+linetrace = False
 embedsignature = True