pytest fails with OSError: dlopen: cannot load any more object with static TLS

[jimthompson5802]

OS: MacOS 10.14.6 Docker for Mac: 2.1.01 docker base image: continuumio/minconda3:4.3.27 mlbox: 0.8.0 Python: 3.6

pytest unit tests fail with error OSError: dlopen: cannot load any more object with static TLS. This contains full error log mlbox_pytest_failure_static_tls_py36.txt

In researching this issue, found this posting. The dlopen error in this posting was solved by changing the order of imports. From this, I changed the oder of imports in mlbox/model/classification/classifier.py for 'lightgbm'. See this screenshot for the change of import order, i.e., import lightgbm before the sklearn imports.

With this change, the all the unit test cases pass. Here is the pytest run log after changing import order. mlbox_pytest_run_afer_import_change.txt

I'll submit a PR to address this issue.

For reference, this is the Dockerfile used for this test.

FROM continuumio/miniconda3:4.3.27

# 
# install additional packages
#
RUN pip install mlflow mlbox

WORKDIR /opt/project
ENV MLFLOW_TRACKING_URI /opt/project/tracking