From bbb8be6e5afacb3e690fa3bf7dc557905a02078e Mon Sep 17 00:00:00 2001
From: Louis Bergelson <louisb@broadinstitute.org>
Date: Mon, 3 Apr 2023 15:34:43 -0400
Subject: [PATCH] Updates to reduce size of docker (#8259)

* reduce unnecessary files in our docker
* update base docker to broadinstitute/gatk:gatkbase-3.1.0
---
 Dockerfile                         | 35 ++++++++++++++++++++----------
 scripts/docker/gatkbase/Dockerfile | 20 +++++++++--------
 2 files changed, 34 insertions(+), 21 deletions(-)

diff --git a/Dockerfile b/Dockerfile
index 98f8a3ffdfc..a513863128d 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -1,5 +1,7 @@
+ARG BASE_DOCKER=broadinstitute/gatk:gatkbase-3.1.0
+
 # stage 1 for constructing the GATK zip
-FROM broadinstitute/gatk:gatkbase-3.0.0 AS gradleBuild
+FROM ${BASE_DOCKER} AS gradleBuild
 LABEL stage=gatkIntermediateBuildImage
 ARG RELEASE=false
 
@@ -8,12 +10,16 @@ ADD . /gatk
 WORKDIR /gatk
 
 # Get an updated gcloud signing key, in case the one in the base image has expired
-RUN rm /etc/apt/sources.list.d/google-cloud-sdk.list
-RUN apt update
-RUN apt-key list
-RUN curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | apt-key --keyring /usr/share/keyrings/cloud.google.gpg add -
-RUN add-apt-repository universe && apt update
-RUN apt-get --assume-yes install git-lfs
+RUN rm /etc/apt/sources.list.d/google-cloud-sdk.list && \
+    apt update &&\
+    apt-key list && \
+    curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | apt-key --keyring /usr/share/keyrings/cloud.google.gpg add - && \
+    add-apt-repository universe && apt update && \
+    apt-get --assume-yes install git-lfs && \
+    apt-get -y clean  && \
+    apt-get -y autoclean  && \
+    apt-get -y autoremove && \
+    rm -rf /var/lib/apt/lists/*
 RUN git lfs install --force
 
 #Download only resources required for the build, not for testing
@@ -22,15 +28,21 @@ RUN git lfs pull --include src/main/resources/large
 RUN export GRADLE_OPTS="-Xmx4048m -Dorg.gradle.daemon=false" && /gatk/gradlew clean collectBundleIntoDir shadowTestClassJar shadowTestJar -Drelease=$RELEASE
 RUN cp -r $( find /gatk/build -name "*bundle-files-collected" )/ /gatk/unzippedJar/
 RUN unzip -o -j $( find /gatk/unzippedJar -name "gatkPython*.zip" ) -d /gatk/unzippedJar/scripts
+RUN chmod -R a+rw /gatk/unzippedJar
 
-FROM broadinstitute/gatk:gatkbase-3.0.0
+FROM ${BASE_DOCKER}
 
-RUN rm /etc/apt/sources.list.d/google-cloud-sdk.list
-RUN apt update
-RUN apt-key list
+RUN rm /etc/apt/sources.list.d/google-cloud-sdk.list && \
+    apt update && \
+    apt-key list && \
+    apt-get -y clean  && \
+    apt-get -y autoclean  && \
+    apt-get -y autoremove && \
+    rm -rf /var/lib/apt/lists/*
 
 WORKDIR /gatk
 
+RUN chmod -R a+rw /gatk
 # Location of the unzipped gatk bundle files
 COPY --from=gradleBuild /gatk/unzippedJar .
 
@@ -76,7 +88,6 @@ ENV CLASSPATH /gatk/gatk.jar:$CLASSPATH
 # Start GATK Python environment
 
 WORKDIR /gatk
-RUN chmod -R a+rw /gatk
 ENV PATH $CONDA_PATH/envs/gatk/bin:$CONDA_PATH/bin:$PATH
 RUN conda env create -n gatk -f /gatk/gatkcondaenv.yml && \
     echo "source activate gatk" >> /gatk/gatkenv.rc && \
diff --git a/scripts/docker/gatkbase/Dockerfile b/scripts/docker/gatkbase/Dockerfile
index 95bafaa9350..7f4d8cb6a7b 100644
--- a/scripts/docker/gatkbase/Dockerfile
+++ b/scripts/docker/gatkbase/Dockerfile
@@ -22,7 +22,8 @@ RUN apt-get update && \
     software-properties-common && \
     apt-get -y clean  && \
     apt-get -y autoclean  && \
-    apt-get -y autoremove
+    apt-get -y autoremove && \
+    rm -rf /var/lib/apt/lists/*
 
 RUN java -version
 
@@ -35,6 +36,10 @@ RUN echo "deb [signed-by=/usr/share/keyrings/cloud.google.gpg] http://packages.c
     apt-get -y clean  && \
     apt-get -y autoclean  && \
     apt-get -y autoremove && \
+    rm -rf /var/lib/apt/lists/* && \
+#  Remove the anthos cli tool and related files since they are very large and we / anyone using the docker are unlikely to use them
+#  Remove the bundled python because we have python installed separately
+    rm -rf /usr/lib/google-cloud-sdk/bin/anthoscli /usr/lib/google-cloud-sdk/platform/anthoscli_licenses /usr/lib/google-cloud-sdk/platform/bundledpythonunix && \
     find / -wholename "*__pycache__/*.pyc" -exec rm {} + 
 
 # Set environment variables.
@@ -50,19 +55,16 @@ ENV JAVA_LIBRARY_PATH /usr/lib/jni
 
 # Install miniconda
 ENV DOWNLOAD_DIR /downloads
-ENV CONDA_URL https://repo.continuum.io/miniconda/Miniconda3-4.7.12.1-Linux-x86_64.sh
-ENV CONDA_MD5 = "81c773ff87af5cfac79ab862942ab6b3"
+ENV CONDA_URL https://repo.anaconda.com/miniconda/Miniconda3-py310_23.1.0-1-Linux-x86_64.sh 
+ENV CONDA_MD5 = "32d73e1bc33fda089d7cd9ef4c1be542616bd8e437d1f77afeeaf7afdb019787"
 ENV CONDA_PATH /opt/miniconda
+ENV PATH $CONDA_PATH/bin:$PATH
 RUN mkdir $DOWNLOAD_DIR && \
     wget -nv -O $DOWNLOAD_DIR/miniconda.sh $CONDA_URL && \
     test "`md5sum $DOWNLOAD_DIR/miniconda.sh | awk -v FS='  ' '{print $1}'` = $CONDA_MD5" && \
     bash $DOWNLOAD_DIR/miniconda.sh -p $CONDA_PATH -b && \
     rm $DOWNLOAD_DIR/miniconda.sh && \
-    ${CONDA_PATH}/bin/conda clean -afy && \
+    conda clean -afy && \
     find /opt/miniconda/ -follow -type f -name '*.a' -delete && \
     find /opt/miniconda/ -follow -type f -name '*.pyc' -delete && \
-    rm -rf /root/.cache/pip
-    
-
-# Deleting unneeded caches
-RUN rm -rf /var/lib/apt/lists/*
+    rm -rf /root/.cache/pip