为什么在运行exe文件时cx_Freeze会出现此错误？

Question

I am a new user of cx_Freeze and I wanted to package my python file with all needed dependencies.

I use Python3 .

I work under a python virtualenv with dependencies such as tensorflow, sklearn, matplotlib, python-vlc...

Here are all files needed to reproduce the same error on Windows.

My requirements.txt to install

absl-py==0.7.0
altgraph==0.16.1
astor==0.7.1
audioread==2.1.6
certifi==2018.11.29
chardet==3.0.4
cx-Freeze==5.1.1
cycler==0.10.0
decorator==4.3.2
distlib==0.2.8
future==0.17.1
gast==0.2.2
grpcio==1.18.0
h5py==2.9.0
idna==2.8
Jinja2==2.10
joblib==0.13.1
Keras-Applications==1.0.7
Keras-Preprocessing==1.0.9
kiwisolver==1.0.1
librosa==0.6.2
llvmlite==0.27.0
macholib==1.11
Markdown==3.0.1
MarkupSafe==1.1.0
matplotlib==3.0.2
numba==0.42.0
numpy==1.16.1
pefile==2018.8.8
protobuf==3.6.1
py2exe==0.9.2.2
PyInstaller==3.4
pynsist==2.3
pyparsing==2.3.1
pypiwin32==223
pysrt==1.1.1
python-dateutil==2.8.0
python-vlc==3.0.4106
pywin32==224
pywin32-ctypes==0.2.0
requests==2.21.0
requests-download==0.1.2
resampy==0.2.1
scikit-learn==0.20.2
scipy==1.2.0
six==1.12.0
sklearn==0.0
tensorboard==1.12.2
tensorflow==1.12.0
termcolor==1.1.0
tornado==5.1.1
urllib3==1.24.1
watson-developer-cloud==2.8.0
websocket-client==0.48.0
Werkzeug==0.14.1
yarg==0.1.9

I have one main python file IHM.py that imports modules from two others python files sync.py and neurnet.py .

File ihm.py

# -*- coding: utf-8 -*-

# for Python3
from tkinter import *
import matplotlib.backends.backend_tkagg
import matplotlib.pyplot as mp
import tkinter.filedialog
from watson_developer_cloud import LanguageTranslatorV3
import json
import subprocess
import os
import vlc
from sync import *
import warnings

warnings.filterwarnings("ignore")

# Définition des variables globales
vid_file_path = ""
vid_srt_path = ""
vid_srt_path_trans = ""
vid_srt_path_sync = ""


# Création de la fenetre d'IHM
fenetre = Tk()
fenetre.title("Test Tkinter Windows")



fenetre.mainloop()

File sync.py

# -*- coding: utf-8 -*-
from __future__ import division
from neuralNet import 

File neurnet.py

#coding: utf-8
import os
import time
import sys
import librosa
import re
import io
import subprocess
import pysrt
import numpy as np
import matplotlib.pyplot as plt
from time import time
import tensorflow as tf
import pickle
#from sklearn.model_selection import train_test_split
#from tensorflow.contrib.layers import flatten
import sklearn



from subprocess import STDOUT

try:
    from subprocess import DEVNULL # py3k
except ImportError:
    import os
    DEVNULL = open(os.devnull, 'wb')

And here, the cx_Freeze setup.py :

# setup.py 
import sys, os
from cx_Freeze import setup, Executable

os.environ['TCL_LIBRARY'] = 'C:/Program Files/Python36/tcl/tcl8.6' 
os.environ['TK_LIBRARY'] = 'C:/Program Files/Python36/tcl/tk8.6'
__version__ = "1.1.0"

buildOptions = dict(
    packages = [],
    excludes = [],  includes = ["idna.idnadata"],
    include_files = ['C:/Program Files/Python36/DLLs/tcl86t.dll','C:/Program Files/Python36/DLLs/tk86t.dll'] )

import sys

base = 'Win32GUI' if sys.platform=='win32' else None 
executables = [
    Executable('ihm.py', base=base) 
] 
setup(
    name = "mgp320",
    description='Projet Neural Network Speech Detection',
    version=__version__,
    options = dict(build_exe = buildOptions),
    executables = executables
)

Running this setup by using command : python setup.py build So it creates me an exe file but when I run this exe I got this error window :

Error cx_Freeze

And the trackback is the following :

Traceback (most recent call last):
 File
"C:\Users\achraf.bentabib\Desktop\aapsa\aapsa\env\lib\site-p
ackages\cx_Freeze\initscript\__startup__.py", line 14, in run
  module.run()
 File
"C:\Users\achraf.bentabib\Desktop\aapsa\aapsa\env\lib\site-p
ackages\cx_Freeze\initscript\Console.py", line 26, in run
  exec(code, m.__dict__)
 File "ihm.py", line 5, in <module>
 File
"C:\Users\achraf.bentabib\Desktop\aapsa\aapsa\env\lib\site-p
ackages\matplotlib\__init__.py", line 120, in <module>
   import distutils.version
 File
"C:\Users\achraf.bentabib\Desktop\aapsa\aapsa\env\lib\distut
ils\__init__.py", line 35, in <module>
   loader.exec_module(real_distutils)
 File "<frozen importlib._bootstrap_external>", line 674, in
exec_module
 File "<frozen importlib._bootstrap_external>", line 780, in
get_code
 File "<frozen importlib._bootstrap_external>", line 832, in
get_data
FileNotFoundError: [Errno 2] No such file or directory:
"C:\\Users\\achraf.bentabib\\Desktop\\aapsa\\aapsa\\stack\\b
uild\\exe.win-amd64-3.6\\lib\\library.zip\\distutils\\__init__.py'

I really don't know how can I fix that..

Answer 1

I see the following potential problems in your setup script for cx_Freeze 5.1.1:

1. You are using numpy (and matplotlib which depends on numpy ). In order that cx_Freeze correctly freezes numpy , it needs to be added to the packages list of the build_exe options.

2. For cx_Freeze version 5.1.1, the TCL/TK DLLs need to be included in a lib subdirectory of the build directory. You can do that by passing a tuple (source, destination) to the corresponding entry of the include_files list option. Furthermore, it would be safer to dynamically find out the location of the TCL/TK DLLs.

Altogether, try with the following modifications in your setup script:

PYTHON_INSTALL_DIR = os.path.dirname(sys.executable)
os.environ['TCL_LIBRARY'] = os.path.join(PYTHON_INSTALL_DIR, 'tcl', 'tcl8.6')
os.environ['TK_LIBRARY'] = os.path.join(PYTHON_INSTALL_DIR, 'tcl', 'tk8.6')

buildOptions = dict(
    packages = ["numpy"],
    excludes = [],
    includes = ["idna.idnadata"],
    include_files=[(os.path.join(PYTHON_INSTALL_DIR, 'DLLs', 'tk86t.dll'),
                    os.path.join('lib', 'tk86t.dll')),
                   (os.path.join(PYTHON_INSTALL_DIR, 'DLLs', 'tcl86t.dll'),
                    os.path.join('lib', 'tcl86t.dll'))]
)

If it still does not work with this modification, start from a minimal tkinter application, for example by commenting out all imports but tkinter in the example application you have posted and make it work on your system. Re-add then the modules you need one by one ( numpy , matplotlib , ...) and check that the unfrozen and frozen applications work at each step. You can for example add a message box to the example application and print there the version of each module you import.

Answer 2

I finally found a solution for the last traceback:

from distutils import dist, sysconfig # isort:skip 
ImportError: cannot import name "dist" 

The problem is that distutils doesn't have many module that are not installed in the virtualenv. (only __init__.py ) So when we build the exe, it doesn't find the distutils modules...

To fix that we have to manually import disutils

import distutils
import opcode
import os
distutils_path = os.path.join(os.path.dirname(opcode.__file__), 'distutils')

And then, include this path to the include_files options of cx_freeze, and exclude distutils in the excludes section.

buildOptions = dict(
    packages = ['llvmlite', 'pkg_resources._vendor', "tkinter", 'numba', "tkinter.filedialog", "audioread", "librosa", "scipy", "numpy"],
    excludes = ["scipy.spatial.cKDTree", 'distutils'],
    includes = ["idna.idnadata", 'numpy.core._methods', 'numpy.lib.format', 'matplotlib.backends.backend_tkagg'],
    include_files = [(distutils_path, 'distutils'), 'C:/Program Files/Python36/DLLs/tcl86t.dll','C:/Program Files/Python36/DLLs/tk86t.dll']
)

I have also add pkg_resources._vendor in the packages because without this, cx_freeze raise an error like

ImportError: The 'appdirs' package is required; normally this is bundled with th
is package so if you get this warning, consult the packager of your distribution

Finally it work, I added numba to the package to includes because like for disutils, build doesn't create all required modules (for librosa).

The last problem was about multiprocessing lib. I have to manually rename Pool.pyc to pool.py c in the build lib folder.