Dockerfile

# Using this image will ensure a specific Python version (3.10) and specific CUDA Version (12.1)
# Note that `nvidiaDriverVersion` must be at least 525.60.13 on linux to support Cuda 12.1:
# https://docs.nvidia.com/cuda/archive/12.1.0/cuda-toolkit-release-notes/
# Google cloud makes driver version recommendations for GCE VMs here:
# https://cloud.google.com/compute/docs/gpus/install-drivers-gpu
FROM pytorch/pytorch:2.3.1-cuda12.1-cudnn8-runtime

# Overrides the default pytorch image workdir
WORKDIR /

COPY requirements.txt /
COPY setup.py /
RUN pip install --no-cache-dir -r /requirements.txt
RUN pip install torch-scatter -f https://data.pyg.org/whl/torch-2.1.0+cu121.html

ADD permutect/ /permutect

RUN pip install build
RUN python3 -m build --sdist
RUN pip install dist/*.tar.gz

# extra utilities for WDL tasks -- command line tools, not python packages
# for easy upgrade later. ARG variables only persist during build time
ARG bcftoolsVer="1.21"

# install dependencies, cleanup apt garbage
RUN apt-get update && apt-get install --no-install-recommends -y \
 wget \
 bzip2 \
 autoconf \
 automake \
 make \
 gcc \
 zlib1g-dev \
 libbz2-dev \
 liblzma-dev \
 libcurl4-gnutls-dev \
 libssl-dev \
 libgsl0-dev && \
 rm -rf /var/lib/apt/lists/* && apt-get autoclean

# get bcftools and make /data
RUN wget https://github.com/samtools/bcftools/releases/download/${bcftoolsVer}/bcftools-${bcftoolsVer}.tar.bz2 && \
 tar -vxjf bcftools-${bcftoolsVer}.tar.bz2 && \
 rm bcftools-${bcftoolsVer}.tar.bz2 && \
 cd bcftools-${bcftoolsVer} && \
 make && \
 make install

CMD ["/bin/sh"]