Troubleshooting

If you are unable to install SDV Enterprise please check the technical requirements first for your platform. If you have met those requirements and there are still errors, you may be encountering an uncommon situation. Check below for some known cases and suggestions.

Unable to provide a valid username/password combo?

After purchasing SDV Enterprise, you should have received an email containing your username and password. Your username is your email address, and your password is license key we provided for you.

If this is not working, or you have lost your credentials, please contact DataCebo so that we can reset your credentials and provide you new ones.

Do you have a Firewall?

If you or your IT department is enforcing a Firewall, it may be blocking your machine from accessing our server. This may result many retry messages, and an error such as the one shown below.

Could not fetch URL https://pypi.datacebo.com/sdv-enterprise/:
There was a problem confirming the ssl certificate

In this case, please retry your download or installation command with the following flags added to the end of the command .

--trusted-host pypi.org --trusted-host pypi.python.org --trusted-host files.pythonhosted.org --trusted-host pypi.datacebo.com

This should grant explicit approval to download the SDV Enterprise package from our servers. For example:

pip install -U sdv-enterprise --no-index --index-url https://pypi.datacebo.com --trusted-host pypi.org --trusted-host pypi.python.org --trusted-host files.pythonhosted.org --trusted-host pypi.datacebo.com

Are you running a macOS in Rosetta mode?

Rosetta mode was introduced to allow macOS devices with M-chips (M1, M2, …) to run software compatible with Intel chips. In most cases, this mode is not necessary for installing SDV and you may have better luck turning it off. Check if you're running Rosetta mode using the Python code below.

import platform
import subprocess

py_arch = platform.machine()
arch = subprocess.check_output(['arch']).strip().decode('utf-8')

if py_arch == 'x86_64' and arch == 'arm64':
  print('Rosetta mode?:', False)
else:
  print('Rosetta mode?:', True)

Solution: We recommend turning Rosetta mode off, as it is not necessary for using SDV or any of its dependencies. If you are required to use Rosetta mode for some reason, you can try creating a special environment for use with Intel.

Below is a conda snippet you can use in your terminal.

conda create -n my_env
conda activate my_env
conda config --env --set subdir osx-arm64
conda install python=3.9

Do you have an older Mac with an Intel chip?

Older Macs with Intel chips may run into issues when running neural network-based synthesizers such as CTGAN, TVAE, or PAR. You may see the following error:

A module that was compiled using NumPy 1.x cannot be run in
NumPy 2.2.0 as it may crash. To support both 1.x and 2.x
versions of NumPy, modules must be compiled with NumPy 2.0.
Some module may need to rebuild instead e.g. with 'pybind11>=2.12'.

If you are a user of the module, the easiest solution will be to
downgrade to 'numpy<2' or try to upgrade the affected module.
We expect that some modules will need time to support NumPy 2.

This error is occurring because an underlying data science library we use has stopped providing support for Intel-based Macs. But you should still be able to use all SDV features by downgrading your version of the numpy data science library.

pip install "numpy<2.0.0"

Note that this should not affect you if you are using HSA or GaussianCopula.

Are you using a virtual machine or container to install SDV?

In this case, please ensure you meet the technical requirements in the final destination where you plan to run SDV (as opposed to your local machine).

Still having problems?

Reach out to us! Please run the code below in Python to print out some additional troubleshooting info. Please reach out to the DataCebo team with a detailed explanation of your error and the output of the code below.

import pip
import platform
import sys
import subprocess
import sysconfig
import re

os_version = platform.version()
os_system = platform.system()
linux_distro = None
is_linux = sys.platform == 'linux'

if sys.platform == 'darwin':
    os_version = subprocess.check_output(['sw_vers', '-productVersion']).strip().decode('utf-8')
    os_system = 'macOS'
if is_linux:
    os_version = platform.release()

    linux_distro = subprocess.run("cat /etc/*-release", shell=True, capture_output=True, text=True)
    linux_distro = linux_distro.stdout.strip()
    distro_name_match = re.search(r'PRETTY_NAME="(.*)"', linux_distro)
    if distro_name_match:
        linux_distro = distro_name_match.group(1)

is_64bits = sys.maxsize > 2**32
bit = "64-bit" if is_64bits else "32-bit"

if platform.system().lower() == "windows":
    arch = sysconfig.get_platform()
else:
    arch = subprocess.check_output(['uname', '-m'], text=True).strip().lower()
print(f'Operating System: {os_system}')

if is_linux:
    os_system += ' Kernel'
print(f'{os_system} Version: {os_version}')

if is_linux:
    print(f'Linux Distribution: {linux_distro}')

print(f'Architecture: {arch}')
print(f'Python Version: {platform.python_version()}')
print(f'System Bit: {bit}')
print(f'pip Version: {pip.__version__}')

print(f'\nAdditional Information For Troubleshooting')
platform_tag = sysconfig.get_platform()
platform_tag = platform_tag.replace('-', '_').replace('.', '_')
print(f'Processor: {platform.processor()}')
print(f'Platform Tag: {platform_tag}')