Merge pull request #25 from instituteofcancerresearch/docker

Docker

.github/workflows/docker.yml

@@ -0,0 +1,29 @@
+name: Upload to Docker Hub
+
+on:
+  push:
+    branches:
+      - main
+
+jobs:
+  docker:
+    runs-on: ubuntu-latest
+    steps:
+      -
+        name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+      -
+        name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+      -
+        name: Login to Docker Hub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+      -
+        name: Build and push
+        uses: docker/build-push-action@v5
+        with:
+          push: true
+          tags: icrsc/soprano:latest

.gitignore

@@ -83,4 +83,8 @@ app_sources/
 !app_sources/immunopeptidomes/.gitkeep
 
 # Ignore release cache file
-src/SOPRANO/data/ensembl.releases
+src/SOPRANO/data/ensembl.releases
+
+# Ensembl downloads
+ensembl_downloads/**/*.fa
+ensembl_downloads/**/*.fa.gz

Dockerfile

@@ -0,0 +1,60 @@
+FROM condaforge/mambaforge:23.3.1-1 as conda
+
+WORKDIR /app
+
+# Construct conda environment and clean up cache files
+COPY src/SOPRANO/local.yml ./
+RUN mamba env create -f local.yml --name soprano &&  \
+    mamba clean -afy &&  \
+    rm local.yml
+
+# Install R dependencies pulled from GH
+COPY install_R_pkgs.R ./
+RUN mamba run --no-capture-output -n soprano Rscript install_R_pkgs.R &&  \
+    rm install_R_pkgs.R
+
+# Prepare Python dependencies (this could be better...)
+COPY pyproject.toml ./
+RUN sed -i '71,84d' pyproject.toml &&  \
+    sed -i '38d' pyproject.toml &&  \
+    sed -i '/name = "SOPRANO"/a version="0.0.1"' pyproject.toml && \
+    touch README.md
+
+# Run pip install
+COPY src ./src
+RUN mamba run --no-capture-output -n soprano pip install -e . &&  \
+    mamba run --no-capture-output -n soprano pip cache purge
+
+# Decompress transcript ID
+RUN if [[ ! -f "./src/SOPRANO/data/ensemble_transcriptID.fasta" ]] ; then gunzip -v ./src/SOPRANO/data/ensemble_transcriptID.fasta.gz ; fi
+
+# Clean up additional stuff ...
+RUN rm -rf ./src/SOPRANO/immunopeptidomes/mouse && \
+    find ./src/SOPRANO/scripts -name "*.R" -delete && \
+    find -name "*.yml" -delete &&  \
+    find -name "*.gz" -delete &&  \
+    find -name '__pycache__' -type d -exec rm -rf {} + && \
+    cd /opt/conda &&  \
+    rm -rf conda-meta && \
+    find -name '__pycache__' -type d -exec rm -rf {} + && \
+    cd /opt/conda/lib && \
+    find -name 'tests' -type d -exec rm -rf {} +
+
+# Expose port for streamlit interface
+EXPOSE 8501
+
+# Define streamlit run as entry point
+ENTRYPOINT ["mamba", "run", "-n", "soprano", "streamlit", "run", "./src/SOPRANO/app.py", "--server.port=8501", "--server.address=0.0.0.0"]
+
+# Bind mount to downloads with:
+# docker run -d -p 8501:8501 --name devtest -v "$(pwd)"/ensembl_downloads/homo_sapiens:/app/ensembl_downloads/homo_sapiens soprano
+# docker run -d -p 8501:8501 --name devtest -e SOPRANO_DISABLE_DOWNLOADS=True -v "$(pwd)"/ensembl_downloads/homo_sapiens:/app/ensembl_downloads/homo_sapiens soprano
+# Should be visible on e.g. http://localhost:8501/
+
+# Download and from registry:
+# docker run -d -p 8501:8501 --name hub_soprano -e SOPRANO_DISABLE_DOWNLOADS=True icrsc/soprano:latest
+
+# Mount data volumes with
+# docker run -d -p 8501:8501 --name hub_soprano -e SOPRANO_DISABLE_DOWNLOADS=True -v /path/to/ensembl/downloads/:/app/ensembl_downloads/homo_sapiens icrsc/soprano:latest
+# e.g.
+# docker run -d -p 8501:8501 --name hub_soprano -e SOPRANO_DISABLE_DOWNLOADS=True -v "$(pwd)"/ensembl_downloads/homo_sapiens:/app/ensembl_downloads/homo_sapiens icrsc/soprano:latest

src/SOPRANO/app.py

@@ -1,3 +1,4 @@
+import os
 import pathlib
 
 import streamlit as st
@@ -180,6 +181,17 @@ def with_tab_genomes(tab: DeltaGenerator):
             "may be used to accelerate the annotation procedure; though this "
             "is _not_ essential."
         )
+        feature_disabled = (
+            os.environ.get("SOPRANO_DISABLE_DOWNLOADS", "False") == "True"
+        )
+
+        if feature_disabled:
+            st.warning(
+                "Downloading genome references is currently disabled at this "
+                "app hosting.\n\n"
+                "Contact the administrators [email protected] "
+                "to request additional genomic files to be added."
+            )
 
         species_selection = st.text_input(
             "Define the species",
@@ -210,7 +222,7 @@ def with_tab_genomes(tab: DeltaGenerator):
         )
         type_processed = DownloaderUIProcessing.type(type_selection)
 
-        if st.button("Download", disabled=False):
+        if st.button("Download", disabled=feature_disabled):
             RunTab.download(
                 species=species_processed,
                 assembly=assembly_processed,

src/SOPRANO/core/objects.py

@@ -293,7 +293,10 @@ class _GatherReferences:
     sizes_done: Set[int] = set()
 
     def _dest_directory(self, release: int):
-        return Directories.data(self.species) / f"{release}_{self.assembly}"
+        return (
+            Directories.ensembl_downloads(self.species)
+            / f"{release}_{self.assembly}"
+        )
 
     def _dest_fa_gz(self, release: int, _toplevel: bool):
         return self._dest_directory(release) / filename_from_url(