From f928a582de0f723de9f304254503d1c686b60d5d Mon Sep 17 00:00:00 2001
From: Pieter Verschaffelt <pieter.verschaffelt@ugent.be>
Date: Tue, 19 Mar 2024 09:20:41 +0100
Subject: [PATCH] Implement Docker image for new Unipept suffix array

---
 .github/workflows/build_index_image.yml  | 53 ++++++++++++++++++++
 examples/suffix-array/docker-compose.yml | 26 ++++++++++
 unipept-index/Dockerfile                 | 64 ++++++++++++++++++++++++
 unipept-index/initialize_container.sh    | 41 +++++++++++++++
 4 files changed, 184 insertions(+)
 create mode 100644 .github/workflows/build_index_image.yml
 create mode 100644 examples/suffix-array/docker-compose.yml
 create mode 100644 unipept-index/Dockerfile
 create mode 100644 unipept-index/initialize_container.sh

diff --git a/.github/workflows/build_index_image.yml b/.github/workflows/build_index_image.yml
new file mode 100644
index 0000000..e98d958
--- /dev/null
+++ b/.github/workflows/build_index_image.yml
@@ -0,0 +1,53 @@
+# This is a basic workflow to help you get started with Actions
+
+name: Build Index Image
+
+# Controls when the workflow will run
+on:
+  workflow_dispatch:
+  push:
+    branches:
+      - 'main'
+    tags:
+      - 'v*.*.*'
+
+permissions:
+  packages: write
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v2
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v2
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v2
+      - name: Login to GHCR
+        uses: docker/login-action@v2
+        with:
+          registry: ghcr.io
+          username: ${{ github.repository_owner }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+      - name: Docker meta
+        id: metadata
+        uses: docker/metadata-action@v3
+        with:
+          images: |
+            ghcr.io/unipept/unipept-index
+          tags: |
+            type=semver,pattern={{version}}
+            type=semver,pattern={{major}}.{{minor}}
+            type=sha
+      - name: Build and push
+        uses: docker/build-push-action@v3
+        with:
+          context: 'unipept-index'
+          platforms: linux/amd64,linux/arm64
+          push: true
+          file: 'unipept-index/Dockerfile'
+          tags: ${{ steps.metadata.outputs.tags }}
+          labels: ${{ steps.metadata.outputs.labels }}
+
+
diff --git a/examples/suffix-array/docker-compose.yml b/examples/suffix-array/docker-compose.yml
new file mode 100644
index 0000000..51ef9ea
--- /dev/null
+++ b/examples/suffix-array/docker-compose.yml
@@ -0,0 +1,26 @@
+version: '3.7'
+
+services:
+  db:
+    image: 26aa16358195
+    environment:
+      # Valid values: swissprot, trembl
+      DB_TYPES: swissprot
+      DB_SOURCES: https://ftp.expasy.org/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot.dat.gz
+    restart: always
+    ports:
+      - "3000:3000"
+    volumes:
+      # Volume that's being used to store the Unipept index. This index is a special kind of structure that can easily
+      # be queried and filtered by taxa (if required). The index will be created on first run of an image (if it is
+      # not yet present in the folder that's mounted) and will greatly increase the build speed of the database.
+      - type: bind
+        # Change this to a folder on your hard drive where the Unipept index could be stored.
+        source: /Users/pverscha/Documents/Unipept/data/index
+        target: /index
+      - type: bind
+        source: /Users/pverscha/Documents/Unipept/data/temp
+        target: /tmp
+      - type: bind
+        source: /Users/pverscha/Documents/Unipept/data/suffix-array
+        target: /suffix-array
diff --git a/unipept-index/Dockerfile b/unipept-index/Dockerfile
new file mode 100644
index 0000000..836d173
--- /dev/null
+++ b/unipept-index/Dockerfile
@@ -0,0 +1,64 @@
+FROM ubuntu:latest
+
+LABEL maintainer="Pieter Verschaffelt <pieter.verschaffelt@ugent.be>"
+
+RUN apt update && \
+    apt install -y \
+    git \
+    wget \
+    unzip \
+    gawk \
+    binutils \
+    gcc \
+    libssl-dev \
+    uuid-runtime \
+    pv \
+    pigz \
+    parallel \
+    curl \
+    sudo \
+    lz4 \
+    dos2unix \
+    pkg-config \
+    clang \
+    cmake
+
+RUN git clone --depth 1 https://github.com/unipept/unipept-database
+# Use the webserver branch for now
+RUN git clone --depth 1 -b webserver https://github.com/BramDevlaminck/Thesis_rust_implementations
+# Initialize all submodules for the suffix array
+RUN cd Thesis_rust_implementations && git submodule update --init --recursive
+
+# Make a directory that contains the initialization data for this script
+RUN mkdir "/scripts"
+COPY "initialize_container.sh" "/scripts/initialize_container.sh"
+# Make sure that this script is executable in the container
+RUN chmod u+x "/scripts/initialize_container.sh"
+
+# Install Rust toolchain (https://rustup.rs/)
+RUN curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | bash -s -- -y
+RUN echo 'source $HOME/.cargo/env' >> $HOME/.bashrc
+ENV PATH="/root/.cargo/bin:${PATH}"
+
+# Compile Rust binaries for the database
+RUN /unipept-database/scripts/build_binaries.sh
+
+# Clean up database build artifacts so they don't end up in the image
+RUN rm -rf /unipept-database/scripts/helper_scripts/unipept-database-rs/target/
+
+# Compile Rust binaries for the unipept index
+RUN cd /Thesis_rust_implementations && cargo build --release
+
+# Uninstall Rust again to keep the image size down
+RUN rustup self uninstall -y
+
+# Database types that should be processed by this image. Delimited by comma's.
+ENV DB_TYPES swissprot
+# Database URLs that should be downloaded and processed by this container. Delimited by comma's, n'th item in this list
+# should correspond to n'th db type given in DB_TYPES arg.
+ENV DB_SOURCES https://ftp.expasy.org/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot.dat.gz
+# How much memory is the database construction script allowed to use for sorting?
+ENV SORT_MEMORY 2G
+
+# Start initialization of the index construction process once the container is constructed.
+CMD "/scripts/initialize_container.sh"
diff --git a/unipept-index/initialize_container.sh b/unipept-index/initialize_container.sh
new file mode 100644
index 0000000..91376e4
--- /dev/null
+++ b/unipept-index/initialize_container.sh
@@ -0,0 +1,41 @@
+#! /bin/bash
+
+# Exit immediately when an error occurs
+set -e
+# Exit when a command in one of the pipelines fails
+set -o pipefail
+
+# Check if the index is already present and only needs to be started (or needs to be constructed if no data is available)
+if [ ! -f "/suffix-array/.completed" ]
+then
+  echo "Started construction of the suffix array"
+
+  # Clear the suffix_array directory and restart the build of the index
+  rm -rf /suffix-array && mkdir -p /suffix-array
+
+  # Clear the temp directory
+  rm -rf /tmp/*
+
+  # First, build the input files for the database and store them in the temporary directory
+  /unipept-database/scripts/build_database.sh -i '/index' -d '/tmp' -m $SORT_MEMORY database $DB_TYPES $DB_SOURCES '/tmp/tables'
+
+  # Move the required files for the suffix array to its directory (and extract only the required columns)
+  lz4cat /tmp/tables/uniprot_entries.tsv.lz4 | cut -f2,4,7 > /suffix-array/proteins.tsv
+  lz4cat /tmp/tables/taxons.tsv.lz4 > /suffix-array/taxons.tsv
+
+  # Remove all other database files
+  rm -rf '/tmp/tables'
+
+  # Now, construct the actual suffix array
+  /Thesis_rust_implementations/target/release/suffixarray_builder -d /suffix-array/proteins.tsv -t /suffix-array/taxons.tsv --sample-rate 3 -o /suffix-array/sa.bin
+
+  # Write the `.completed` file to indicate that this suffix array build is complete
+  touch /suffix-array/.completed
+
+  echo "Finished construction of the suffix array"
+fi
+
+echo "Start loading suffix array"
+
+# At this point, we know that the suffix array exists and we can start it.
+/Thesis_rust_implementations/target/release/suffixarray_server -d /suffix-array/proteins.tsv -t /suffix-array/taxons.tsv -i /suffix-array/sa.bin