[DRAFT] Docker setup for Beagle Imputation WDL

3rd-party-tools/beagle/Dockerfile

@@ -0,0 +1,62 @@
+# Adding a platform tag to ensure that images built on ARM-based machines doesn't break pipelines
+FROM --platform="linux/amd64" adoptopenjdk/openjdk8:alpine-slim
+
+ARG BEAGLE_VERSION=01Mar24.d36 \
+    BREF3_VERSION=22Jul22.46e \
+    BCFTOOLS_VERSION=1.10.2
+
+ENV TERM=xterm-256color \
+    TINI_VERSION=v0.19.0
+
+LABEL MAINTAINER="Broad Institute DSDE <dsde-engineering@broadinstitute.org>" \
+        BEAGLE_VERSION=${BEAGLE_VERSION} \
+        BCFTOOLS_VERSION=${BCFTOOLS_VERSION}
+
+WORKDIR /usr/gitc
+
+# Install dependencies
+RUN set -eux; \
+    apk add --no-cache \
+            autoconf \
+            automake \
+            bash \
+            bzip2-dev \
+            curl \
+            g++ \
+            gcc \
+            gsl-dev \
+            make \
+            musl-dev \
+            perl \
+            perl-dev \
+            tini \
+            wget \
+            xz-dev \
+            zlib-dev \
+    ; \
+# Install BCFTools
+    wget https://github.com/samtools/bcftools/releases/download/${BCFTOOLS_VERSION}/bcftools-${BCFTOOLS_VERSION}.tar.bz2; \
+    tar xf bcftools-${BCFTOOLS_VERSION}.tar.bz2; \
+    cd bcftools-${BCFTOOLS_VERSION}; \
+    \
+    ./configure; \
+    make; \
+    make install; \
+    \
+    cd ../; \
+    rm -r bcftools-${BCFTOOLS_VERSION}; \
+    rm bcftools-${BCFTOOLS_VERSION}.tar.bz2 \
+    ; \
+# Download Beagle jars
+    # beagle runs phasing and imputation
+    curl -L https://faculty.washington.edu/browning/beagle/beagle.${BEAGLE_VERSION}.jar > beagle.${BEAGLE_VERSION}.jar \
+    ; \
+    # bref3 converts a reference panel from vcf to the bref3 format that Beagle needs
+       curl -L https://faculty.washington.edu/browning/beagle/bref3.${BREF3_VERSION}.jar > bref3.${BREF3_VERSION}.jar \
+    ; \
+# Install tini
+    wget https://github.com/krallin/tini/releases/download/$TINI_VERSION/tini -O /sbin/tini; \
+    chmod +x /sbin/tini;
+
+# Set tini as default entrypoint
+ENTRYPOINT ["/sbin/tini", "--" ]

3rd-party-tools/beagle/README.md

@@ -0,0 +1,35 @@
+# Imputation Beagle
+
+## Quick reference
+
+Copy and paste to pull this image
+
+#### `us-central1-docker.pkg.dev/morgan-fieldeng-gcp/imputation-beagle-development/imputation-beagle:0.0.1-01Mar24.d36-wip-temp-20240301`
+
+- __What is this image:__ This image is a lightweight alpine-based image for running Beagle in the [ImputationBeagle pipeline](../../../../pipelines/broad/arrays/imputation_beagle/ImputationBeagle.wdl).
+- __What is Beagle:__ Beagle is a software package for phasing genotypes and imputing ungenotyped markers. Beagle version 5.4 has improved memory and computational efficiency when analyzing large sequence data sets. See [here](https://faculty.washington.edu/browning/beagle/beagle.html) for more information.
+- __How to see Beagle version used in image:__ Please see below.
+
+## Versioning
+
+The Imputation Beagle image uses the following convention for versioning:
+
+#### `us-central1-docker.pkg.dev/morgan-fieldeng-gcp/imputation-beagle-development/imputation-beagle:<image-version>-<beagle-version>-<manual-timestamp>`
+
+We keep track of all past versions in [docker_versions](docker_versions.tsv) with the last image listed being the currently used version in WARP.
+
+You can see more information about the image, including the tool versions, by running the following command:
+
+```bash
+$ docker pull us-central1-docker.pkg.dev/morgan-fieldeng-gcp/imputation-beagle-development/imputation-beagle:0.0.1-01Mar24.d36-wip-temp-20240301
+$ docker inspect us-central1-docker.pkg.dev/morgan-fieldeng-gcp/imputation-beagle-development/imputation-beagle:0.0.1-01Mar24.d36-wip-temp-20240301
+```
+
+## Usage
+
+### Display default menu
+
+```bash
+$ docker run --rm -it \
+    us-central1-docker.pkg.dev/morgan-fieldeng-gcp/imputation-beagle-development/imputation-beagle:0.0.1-01Mar24.d36-wip-temp-20240301 /usr/gitc/beagle
+```

3rd-party-tools/beagle/docker_build.sh

@@ -0,0 +1,78 @@
+#!/bin/bash
+set -e
+
+# Update version when changes to Dockerfile are made
+DOCKER_IMAGE_VERSION=0.0.1
+TIMESTAMP=$(date +"%s")
+DIR=$(cd $(dirname $0) && pwd)
+
+# Registries and tags
+# GCR_URL="us.gcr.io/broad-gotc-prod/imputation-beagle"
+
+# GAR setup
+GAR_REGION="us-central1"
+GAR_PROJECT="morgan-fieldeng-gcp"
+GAR_REPOSITORY="imputation-beagle-development"
+GAR_IMAGE="imputation-beagle"
+GAR_URL="${GAR_REGION}-docker.pkg.dev/${GAR_PROJECT}/${GAR_REPOSITORY}/${GAR_IMAGE}"
+
+# Beagle version
+BEAGLE_VERSION="01Mar24.d36"
+
+# Necessary tools and help text
+TOOLS=(docker gcloud)
+HELP="$(basename "$0") [-h|--help] [-b|--beagle] [-t|--tools] -- script to build the Imputation Beagle image and push to GAR
+
+where:
+    -h|--help Show help text
+    -b|--beagle Version of Beagle to use (default: BEAGLE_VERSION=${BEAGLE_VERSION})
+    -t|--tools Show tools needed to run script
+    "
+
+function main(){
+    for t in "${TOOLS[@]}"; do which $t >/dev/null || ok=no; done
+        if [[ $ok == no ]]; then
+            echo "Missing one of the following tools: "
+            for t in "${TOOLS[@]}"; do echo "$t"; done
+            exit 1
+        fi
+
+    while [[ $# -gt 0 ]]
+    do 
+    key="$1"
+    case $key in
+        -b|--beagle)
+        BEAGLE_VERSION="$2"
+        shift
+        shift
+        ;;
+        -h|--help)
+        echo "$HELP"
+        exit 0
+        ;;
+        -t|--tools)
+        for t in "${TOOLS[@]}"; do echo $t; done
+        exit 0
+        ;;
+        *)
+        shift
+        ;;
+    esac
+    done
+
+    IMAGE_TAG="$DOCKER_IMAGE_VERSION-$BEAGLE_VERSION-$TIMESTAMP"
+
+    echo "building and pushing GCR Image - $GAR_URL:$IMAGE_TAG"
+
+    # TODO: add `--squash` when ready to productionize. https://docs.docker.com/reference/cli/docker/image/build/#squash
+    docker build -t "$GAR_URL:$IMAGE_TAG" \
+        --build-arg BEAGLE_VERSION="$BEAGLE_VERSION" \
+        $DIR   
+        # --no-cache $DIR\
+    docker push "$GAR_URL:$IMAGE_TAG"
+
+    echo -e "$GAR_URL:$IMAGE_TAG" >> "$DIR/docker_versions.tsv"
+    echo "done"
+}
+
+main "$@"

3rd-party-tools/beagle/docker_versions.tsv

@@ -0,0 +1,2 @@
+us-central1-docker.pkg.dev/morgan-fieldeng-gcp/imputation-beagle-development/imputation-beagle:0.0.1-22Jul22.46e-wip-temp-20240227
+us-central1-docker.pkg.dev/morgan-fieldeng-gcp/imputation-beagle-development/imputation-beagle:0.0.1-01Mar24.d36-wip-temp-20240301