You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
if this is meant to be used as an app I would have pretty strict pinning policy where I would explicitly specify the version of all direct dependencies at least with a == and possibly also that of all transitive dependencies with a pip freeze
also it seems that dependencies are installed "twice" in the docker recipe:
RUN pip3 install nilearn==0.9.2 templateflow pybids h5py tqdm&& \
mkdir -p /code && mkdir -p /templateflow <-- first install
WORKDIR /code
RUN python3 -c "from templateflow.api import get; get(['MNI152NLin2009cAsym', 'MNI152NLin6Asym'])"
COPY [".", "/code"]
RUN pip install --upgrade pip && pip3 install -e . <-- second 'indirect' install
I would also probably pin version of the python base imaged used in the docker build to ensure reproducibility.
The text was updated successfully, but these errors were encountered:
if this is meant to be used as an app I would have pretty strict pinning policy where I would explicitly specify the version of all direct dependencies at least with a
==
and possibly also that of all transitive dependencies with apip freeze
also it seems that dependencies are installed "twice" in the docker recipe:
I would also probably pin version of the python base imaged used in the docker build to ensure reproducibility.
The text was updated successfully, but these errors were encountered: