Merge pull request #12095 from velconia/port_py3

Port py3

CMakeLists.txt

@@ -66,6 +66,12 @@ option(WITH_ANAKIN      "Compile with Anakin library"                   OFF)
 option(WITH_GRPC     "Use grpc as the default rpc framework"            ${WITH_DISTRIBUTE})
 option(WITH_BRPC_RDMA     "Use brpc rdma as the rpc protocal"           OFF)
 option(WITH_SYSTEM_BLAS   "Use system blas library"           OFF)
+option(PY_VERSION       "Compile PaddlePaddle with python3 support"     ${PY_VERSION})
+
+# PY_VERSION
+if(NOT PY_VERSION)
+  set(PY_VERSION 2.7)
+endif()
 
 # CMAKE_BUILD_TYPE
 if(NOT CMAKE_BUILD_TYPE)

Dockerfile

@@ -80,7 +80,7 @@ RUN pip install pre-commit 'ipython==5.3.0' && \
     pip install opencv-python
 
 #For docstring checker
-RUN pip install pylint pytest astroid isort
+RUN pip install pylint pytest astroid isort LinkChecker
 
 COPY ./python/requirements.txt /root/
 RUN pip install -r /root/requirements.txt

cmake/external/python.cmake

@@ -18,8 +18,9 @@ ENDIF()
 
 INCLUDE(python_module)
 
-FIND_PACKAGE(PythonInterp 2.7)
-FIND_PACKAGE(PythonLibs 2.7)
+FIND_PACKAGE(PythonInterp ${PY_VERSION})
+FIND_PACKAGE(PythonLibs ${PY_VERSION})
+
 # Fixme: Maybe find a static library. Get SHARED/STATIC by FIND_PACKAGE.
 ADD_LIBRARY(python SHARED IMPORTED GLOBAL)
 SET_PROPERTY(TARGET python PROPERTY IMPORTED_LOCATION ${PYTHON_LIBRARIES})

paddle/legacy/utils/PythonUtil.cpp

@@ -136,7 +136,13 @@ std::string callPythonFunc(const std::string& moduleName,
                            const std::string& funcName,
                            const std::vector<std::string>& args) {
   PyObjectPtr obj = callPythonFuncRetPyObj(moduleName, funcName, args);
+#if PY_MAJOR_VERSION >= 3
+  Py_ssize_t str_size = 0u;
+  const char* str = PyUnicode_AsUTF8AndSize(obj.get(), &str_size);
+  return std::string(str, (size_t)str_size);
+#else
   return std::string(PyString_AsString(obj.get()), PyString_Size(obj.get()));
+#endif  // PY_MAJOR_VERSION >= 3
 }
 
 PyObjectPtr createPythonClass(

paddle/legacy/utils/PythonUtil.h

@@ -88,6 +88,33 @@ PyObjectPtr createPythonClass(const std::string& moduleName,
 namespace py {
 PyObjectPtr import(const std::string& moduleName);
 
+#if PY_MAJOR_VERSION >= 3
+/**
+ * Cast a PyLong to int type T.
+ * @tparam T return type.
+ * @param [in] obj PyLong object.
+ * @param [out] ok status for casting. False if error occured. nullptr if user
+ *                 don't care is ok or not.
+ * @return The value of python object, or 0 if not ok.
+ */
+template <typename T>
+T castInt(PyObject* obj, bool* ok = nullptr) {
+  // Refer to https://www.python.org/dev/peps/pep-0237/, the int and long object
+  // were unified to long since python3
+  if (PyLong_Check(obj)) {
+    if (ok) *ok = true;
+    return (T)PyLong_AsUnsignedLong(obj);
+  } else {
+    if (ok) *ok = false;
+    return (T)0;
+  }
+}
+
+// Convert PyAPI from 2.x to 3.x
+#define PyString_FromString PyUnicode_FromString
+#define PyString_AsString PyUnicode_AsUTF8
+
+#else
 /**
  * Cast a PyLong or PyInt to int type T.
  * @tparam T return type.
@@ -109,6 +136,7 @@ T castInt(PyObject* obj, bool* ok = nullptr) {
     return (T)0;
   }
 }
+#endif  // PY_MAJOR_VERSION >= 3
 
 /**
  * Invoke repr of python object.

paddle/scripts/paddle_build.sh

@@ -78,6 +78,12 @@ function cmake_gen() {
             PYTHON_FLAGS="-DPYTHON_EXECUTABLE:FILEPATH=/opt/python/cp27-cp27mu/bin/python
         -DPYTHON_INCLUDE_DIR:PATH=/opt/python/cp27-cp27mu/include/python2.7
         -DPYTHON_LIBRARIES:FILEPATH=/opt/_internal/cpython-2.7.11-ucs4/lib/libpython2.7.so"
+        elif [ "$1" == "cp35-cp35m" ]; then
+            export LD_LIBRARY_PATH=/opt/_internal/cpython-3.5.1/lib/:${LD_LIBRARY_PATH}
+            export PATH=/opt/_internal/cpython-3.5.1/bin/:${PATH}
+            export PYTHON_FLAGS="-DPYTHON_EXECUTABLE:FILEPATH=/opt/_internal/cpython-3.5.1/bin/python3
+            -DPYTHON_INCLUDE_DIR:PATH=/opt/_internal/cpython-3.5.1/include/python3.5m
+            -DPYTHON_LIBRARIES:FILEPATH=/opt/_internal/cpython-3.5.1/lib/libpython3.so"
         fi
     fi
 
@@ -108,6 +114,7 @@ function cmake_gen() {
         -DWITH_CONTRIB=${WITH_CONTRIB:-ON}
         -DWITH_ANAKIN=${WITH_ANAKIN:-OFF}
         -DWITH_INFERENCE_DEMO=${WITH_INFERENCE_DEMO:-ON}
+        -DPY_VERSION=${PY_VERSION:-2.7}
     ========================================
 EOF
     # Disable UNITTEST_USE_VIRTUALENV in docker because
@@ -136,7 +143,8 @@ EOF
         -DCMAKE_EXPORT_COMPILE_COMMANDS=ON \
         -DWITH_CONTRIB=${WITH_CONTRIB:-ON} \
         -DWITH_ANAKIN=${WITH_ANAKIN:-OFF} \
-        -DWITH_INFERENCE_DEMO=${WITH_INFERENCE_DEMO:-ON}
+        -DWITH_INFERENCE_DEMO=${WITH_INFERENCE_DEMO:-ON} \
+        -DPY_VERSION=${PY_VERSION:-2.7}
 }
 
 function abort(){

python/paddle/__init__.py

@@ -12,16 +12,16 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 try:
-    from version import full_version as __version__
-    from version import commit as __git_commit__
+    from paddle.version import full_version as __version__
+    from paddle.version import commit as __git_commit__
 
 except ImportError:
     import sys
-    sys.stderr.write('''Warning with import paddle: you should not 
+    sys.stderr.write('''Warning with import paddle: you should not
      import paddle from the source directory; please install paddlepaddle*.whl firstly.'''
                      )
 
-import reader
-import dataset
-import batch
+import paddle.reader
+import paddle.dataset
+import paddle.batch
 batch = batch.batch

python/paddle/dataset/__init__.py

@@ -15,20 +15,20 @@
 Dataset package.
 """
 
-import mnist
-import imikolov
-import imdb
-import cifar
-import movielens
-import conll05
-import uci_housing
-import sentiment
-import wmt14
-import wmt16
-import mq2007
-import flowers
-import voc2012
-import image
+import paddle.dataset.mnist
+import paddle.dataset.imikolov
+import paddle.dataset.imdb
+import paddle.dataset.cifar
+import paddle.dataset.movielens
+import paddle.dataset.conll05
+import paddle.dataset.uci_housing
+import paddle.dataset.sentiment
+import paddle.dataset.wmt14
+import paddle.dataset.wmt16
+import paddle.dataset.mq2007
+import paddle.dataset.flowers
+import paddle.dataset.voc2012
+import paddle.dataset.image
 
 __all__ = [
     'mnist',

python/paddle/reader/__init__.py

@@ -66,9 +66,9 @@ An example implementation for multiple item data reader creator:
 TODO(yuyang18): Should we add whole design doc here?
 """
 
-import decorator
-from decorator import *
+import paddle.reader.decorator
+from paddle.reader.decorator import *
 
-import creator
+import paddle.reader.creator
 
 __all__ = decorator.__all__ + ['creator']

python/paddle/reader/decorator.py

@@ -20,7 +20,7 @@ __all__ = [
 from threading import Thread
 import subprocess
 
-from Queue import Queue
+from six.moves.queue import Queue
 import itertools
 import random
 import zlib

python/requirements.txt

@@ -8,4 +8,4 @@ scipy>=0.19.0
 Pillow
 nltk>=3.2.2
 graphviz
-LinkChecker
+six

python/setup.py.in

@@ -17,7 +17,8 @@ def git_commit():
         git_commit = subprocess.Popen(cmd, stdout = subprocess.PIPE).communicate()[0].strip()
     except:
         git_commit = 'Unknown'
-    return git_commit
+    git_commit = git_commit.decode()
+    return str(git_commit)
 
 def _get_version_detail(idx):
     assert idx < 3, "vesion info consists of %(major)d.%(minor)d.%(patch)d, \
@@ -44,6 +45,7 @@ def is_taged():
     try:
         cmd = ['git', 'describe', '--exact-match', '--tags', 'HEAD', '2>/dev/null']
         git_tag = subprocess.Popen(cmd, stdout = subprocess.PIPE).communicate()[0].strip()
+        git_tag = git_tag.decode()
     except:
         return False
 
@@ -67,13 +69,13 @@ with_mkl        = '%(with_mkl)s'
 
 def show():
     if istaged:
-        print 'full_version:', full_version
-        print 'major:', major
-        print 'minor:', minor
-        print 'patch:', patch
-        print 'rc:', rc
+        print('full_version:', full_version)
+        print('major:', major)
+        print('minor:', minor)
+        print('patch:', patch)
+        print('rc:', rc)
     else:
-        print 'commit:', commit
+        print('commit:', commit)
 
 def mkl():
     return with_mkl