Dockerfile

# ** Taken from fMRIPREP **
# Use Ubuntu 16.04 LTS
FROM ubuntu:xenial-20161213

# Pre-cache neurodebian key
COPY docker_files/neurodebian.gpg /root/.neurodebian.gpg

# Prepare environment
RUN apt-get update && \
    apt-get install -y --no-install-recommends \
                    curl \
                    bzip2 \
                    ca-certificates \
                    xvfb \
                    cython3 \
                    build-essential \
                    autoconf \
                    libtool \
                    pkg-config \
                    git=1:2.7.4-0ubuntu1 \
                    graphviz=2.38.0-12ubuntu2 && \
    curl -sSL http://neuro.debian.net/lists/xenial.us-ca.full >> /etc/apt/sources.list.d/neurodebian.sources.list && \
    apt-key add /root/.neurodebian.gpg && \
    (apt-key adv --refresh-keys --keyserver hkp://ha.pool.sks-keyservers.net 0xA5D32F012649A5A9 || true) && \
    apt-get clean && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
    
RUN curl -sL https://deb.nodesource.com/setup_7.x | bash -

# Installing Neurodebian packages (FSL, AFNI, git)
RUN apt-get update && \
    apt-get install -y --no-install-recommends \
                    nodejs \
                    fsl-core=5.0.9-5~nd16.04+1 \
                    fsl-mni152-templates=5.0.7-2 \
                    afni=16.2.07~dfsg.1-5~nd16.04+1 \
                    convert3d && \
    apt-get clean && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*

ENV FSLDIR=/usr/share/fsl/5.0 \
    FSLOUTPUTTYPE=NIFTI_GZ \
    FSLMULTIFILEQUIT=TRUE \
    LD_LIBRARY_PATH=/usr/lib/fsl/5.0:$LD_LIBRARY_PATH \
    POSSUMDIR=/usr/share/fsl/5.0 \
    FSLTCLSH=/usr/bin/tclsh \
    FSLWISH=/usr/bin/wish \
    AFNI_MODELPATH=/usr/lib/afni/models \
    AFNI_IMSAVE_WARNINGS=NO \
    AFNI_TTATLAS_DATASET=/usr/share/afni/atlases \
    AFNI_PLUGINPATH=/usr/lib/afni/plugins
ENV PATH=/usr/lib/fsl/5.0:/usr/lib/afni/bin:$PATH

# Installing miniconda --if problems occur, look here for dependency issues
RUN curl -sSLO https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh && \
    bash Miniconda3-latest-Linux-x86_64.sh -b -p /usr/local/miniconda && \
    rm Miniconda3-latest-Linux-x86_64.sh

ENV PATH=/usr/local/miniconda/bin:$PATH \
    LANG=C.UTF-8 \
    LC_ALL=C.UTF-8 \
    PYTHONNOUSERSITE=1


# Installing precomputed python packages --if problems occur, look here first for dependency issues
ADD docker_files/conda_requirements.txt conda_requirements.txt
RUN conda install -y -c conda-forge --file conda_requirements.txt; sync &&  \
    conda build purge-all && sync 

# Precaching fonts, set 'Agg' as default backend for matplotlib
RUN python -c "from matplotlib import font_manager" && \
    sed -i 's/\(backend *: \).*$/\1Agg/g' $( python -c "import matplotlib; print(matplotlib.matplotlib_fname())" )


# Unless otherwise specified each process should only use one thread - nipype
# will handle parallelization
ENV MKL_NUM_THREADS=1 \
    OMP_NUM_THREADS=1

# Installing dev requirements (packages that are not in pypi) -- if problems occur, look here for dependency issues
WORKDIR /home
ADD docker_files/pip_requirements.txt pip_requirements.txt
RUN pip install -r pip_requirements.txt && \                   
    rm -rf ~/.cache/pip
           

# add jupyterlab extensions -- Don't currently work
#RUN jupyter labextension install jupyterlab-flake8
#RUN jupyter labextension install @jupyterlab/toc

# install updated nilearn
RUN pip install git+https://github.com/nilearn/nilearn@e0573ede3ff2f155a2ff21a65d8047a7395d63fd

# Set up data and script directories, ENV variables 
Run mkdir /scripts
WORKDIR /scripts
ENV SHELL=/bin/bash

# Expose Jupyter port & cmd
EXPOSE 8888
RUN mkdir -p /opt/app/data
CMD jupyter lab --ip=0.0.0.0 --port=8888 --no-browser \
    --notebook-dir=/opt/app/data --allow-root \
    --notebook-dir="/" \
    --NotebookApp.token='' \
    --NotebookApp.password=''