c++ 调用python错误总结

参考：

1. C++ 调用Python脚本
2. C/C++调用python（新手必看）
3. c调用python遇到的那些坑
4. 解决PyRun_SimpleFile/PyRun_SimpleString报错
5. “探坑”——在C++中执行python脚本的艰辛尝试

背景介绍：

做 deepsort 跟踪部署 c++ 调用 python 时出现好多错误，记录下来，方便后续排查。用的IDE 是clion，众所周知它生成的可执行文件在 ./cmake-build-debug 目录下，所以查找 python 文件的位置要相对于./cmake-build-debug下的可执行文件，而不是相对于源文件。

错误1：PyImport_ImportModule老返回NULL

tree-查看目录和文件
本例中 test.cpp 调用 ./Python 中的 main_test.py

$ tree -L 2
.
├── cmake-build-debug
│   ├── CMakeCache.txt
│   ├── CMakeFiles
│   ├── cmake_install.cmake
│   ├── Makefile
│   ├── test
│   └── test1.cbp
├── CMakeLists.txt
├── main.cpp
├── __pycache__
├── Python
│   ├── deep_sort
│   ├── detector.py
│   ├── __init__.py
│   ├── main_test.py
│   ├── models
│   ├── __pycache__
│   ├── run.py
│   ├── test_test.py
│   ├── tracker.py
│   ├── utils
│   ├── weights
│   └── yolov5
├── test.cpp
└── video
    ├── getvideo_189_2019-07-20_09-17-48.avi
    └── test.mp4

CMakeLists.txt 为：

cmake_minimum_required(VERSION 3.10)
project(test1)

set(CMAKE_CXX_STANDARD 14)

set(OpenCV_DIR /home/wxf/softwares/3rdparty/opencv_env/opencv349/share/OpenCV)
find_package(OpenCV REQUIRED)

include_directories(${OpenCV_INCLUDE_DIRS})
include_directories(/home/wxf/softwares/3rdparty/anaconda/envs/deepsort_py37/include)

add_executable(test test.cpp)
target_link_libraries(test
        ${OpenCV_LIBS}
        /home/wxf/softwares/3rdparty/anaconda/envs/deepsort_py37/lib/libpython3.7m.so)

add_executable(main main.cpp)
target_link_libraries(main
        ${OpenCV_LIBS}
        /home/wxf/softwares/3rdparty/anaconda/envs/deepsort_py37/lib/libpython3.7m.so)

出现这个错误时：
受到启发：c++加载python模块，但是PyImport_ImportModule老返回NULL
PyRun_SimpleString改为：

PyRun_SimpleString("sys.path.append(r'../Python')");

并且删除 cmake-build-debug 文件夹，重新编译。

错误2： ModuleNotFoundError: No module named ‘tracker’
test.cpp 调用 ./Python 中的 main_test.py
main_test.py 导入 的包是：

import numpy as np
import cv2
import  tracker
from  detector import Detector

numpy 、opencv 是装的包
tracker 是 tracker.py ，detector 是 detector.py 自己写得包
参考： ModuleNotFoundError: No module named 'xxx’可能的解决方案大全
从 Python中import导入模块时出现ImportError: No module named的解决办法 、ImportError: No module named *** 问题？——理解绝对导入和相对导入 受到启发，然后修改代码：

PyRun_SimpleString("import sys");
PyRun_SimpleString("import cv2");
PyRun_SimpleString("import numpy as np");
PyRun_SimpleString("sys.path.append(r'../Python')");
//PyRun_SimpleString("sys.path.append(r'../Python')"); 一定要放在tracker 和 detector 的前面，指明它们的工作目录
PyRun_SimpleString("import tracker");
PyRun_SimpleString("from detector import Detector");
PyRun_SimpleString("import os");

错误3：

Could not find platform independent libraries <prefix>
Consider setting $PYTHONHOME to <prefix>[:<exec_prefix>]
Fatal Python error: initfsencoding: Unable to get the locale encoding
ModuleNotFoundError: No module named 'encodings'

Current thread 0x00007fc83087e300 (most recent call first):

debug:
添加 conda 虚拟环境的目录

Py_SetPythonHome(L"/xxx/3rdparty/anaconda/envs/deepsort_py37");

错误4： The _imagingft C module is not installed
参照 Python: The _imagingft C module is not installed错误的解决
错误5：

error: ‘import_array’ was not declared in this scope
     import_array();
     ^~~~~~~~~~~~

解决方法：
参考： Python + C/C++ 嵌入式编程（2）：基于Code::Blocks进行环境配置和开发（入门版)
cmakelist.txt 中修改如下：

include_directories(/xxx/anaconda/envs/deepsort_py37/include/python3.7m)
include_directories(/xxx/anaconda/envs/deepsort_py37/lib/python3.7/site-packages/numpy/core/include)

所在代码中添加以下头文件：

#include 
#include 

后续错误：

anaconda/envs/deepsort_py37/lib/python3.7/site-packages/numpy/core/include/numpy/__multiarray_api.h:1532:151: error: return-statement with a value, in function returning 'void' [-fpermissive]
 #define import_array() {if (_import_array() < 0) {PyErr_Print(); PyErr_SetString(PyExc_ImportError, "numpy.core.multiarray failed to import"); return NULL; } }
                                                                                                                                                       ^
/xxx/yolo_deepsort_test/main.cpp:133:5: note: in expansion of macro ‘import_array’
     import_array();
     ^

参考：import_array()报错，返回值类型与函数类型不匹配

错误6： 待更新…