Merge pull request #42 from workflowhub-eu/refactor-workflow

Refactored Snakemake workflow and generate example output

Dockerfile

@@ -1,29 +1,44 @@
-FROM python:3.11-slim
+# Stage 1: Build environment
+FROM python:3.11-slim AS build-stage
 
-RUN pip install poetry
+# Install build tools and Poetry
+RUN apt-get update && apt-get install -y build-essential \
+    && pip install poetry
 
-# Set the working directory
 WORKDIR /app
 
-# Install build tools for Snakemake (gcc, make, etc.)
-RUN apt-get update && apt-get install -y build-essential
+# Copy dependency files and install dependencies
+COPY pyproject.toml poetry.lock /app/
+RUN poetry config virtualenvs.create false \
+    && poetry install --no-interaction --no-ansi
 
-# Copy the pyproject.toml file
-COPY pyproject.toml /app/
+# Copy and install the application
+COPY . /app
+RUN poetry install
 
-# Install the dependencies
-RUN poetry install --no-root
+# Stage 2: Snakemake runtime environment
+FROM snakemake/snakemake:latest
 
-# Copy the rest of the application files
-COPY . /app
+# Install Poetry
+RUN pip install poetry
 
-# Install the package
-RUN poetry install
+WORKDIR /app
+
+# Copy the application from the build stage
+COPY --from=build-stage /app /app
+
+# Install dependencies
+RUN pip install -r <(poetry export --format requirements.txt --without-hashes) \
+    && pip install -e .
+
+# Set up non-root user
+RUN groupadd -r snakemake && useradd -r -g snakemake snakemake \
+    && chown -R snakemake:snakemake /app
 
-# Install Snakemake using Poetry
-RUN poetry add snakemake
+USER snakemake
 
-# Set the entry point for the container
-ENTRYPOINT ["poetry", "run"]
+# Configure Python path
+ENV PYTHONPATH="/app:${PYTHONPATH}"
 
-CMD ["help"]
+# Set the entry point
+ENTRYPOINT ["snakemake"]

Snakefile

@@ -1,20 +1,13 @@
-# TODO - Refactor to input args to the Snakemake file
-WORKFLOW_IDS = range(1,11)
+from snakemake.io import directory
+
 VERSIONS = ['1']
 OUTPUT_DIRS = "data"
 MERGED_FILE = "merged.ttl"
-
-
-def list_expected_files():
-    files = []
-    for wf_id in WORKFLOW_IDS:
-        for ver in VERSIONS:
-            files.append(f"{OUTPUT_DIRS}/{wf_id}_{ver}_ro-crate-metadata.json")
-    return files
+ro_crate_metadata_dir = "ro-crate-metadata/"
 
 rule all:
     input:
-        MERGED_FILE
+        "ro-crate-metadata"
 
 rule source_ro_crates:
     output:
@@ -23,12 +16,20 @@ rule source_ro_crates:
         """
         # Create the output directory if it doesn't exist:
         mkdir -p {OUTPUT_DIRS}
+        
+        # Add the current directory to PYTHONPATH, creating it if it doesn't exist
+        export PYTHONPATH="${{PYTHONPATH:+$PYTHONPATH:}}$(pwd)"
 
-        # Run the source_crates script to download the RO Crate metadata:
-        python workflowhub_graph/source_crates.py  --workflow-ids 1-10 --prod --all-versions
-
-        # After sourcing, check which files were actually created:
-        python workflowhub_graph/check_outputs.py --workflow-ids 1-10 --versions {VERSIONS} --output-dir {OUTPUT_DIRS}
+        # Run the source_crates script to download the RO Crate metadata, 
+        # then check the output files and generate created_files.json:
+        
+        # - all versions of all workflows:
+        # python workflowhub_graph/source_crates.py --prod --all-versions
+        # python workflowhub_graph/check_outputs.py --versions {VERSIONS} --output-dir {OUTPUT_DIRS}
+        
+        # - all versions of first 10 workflows:
+        python workflowhub_graph/source_crates.py --workflow-ids 1-20 --prod --all-versions
+        python workflowhub_graph/check_outputs.py --workflow-ids 1-20 --versions {VERSIONS} --output-dir {OUTPUT_DIRS}
         """
 
 rule report_created_files:
@@ -65,3 +66,36 @@ rule merge_files:
         shell(f"""
             python workflowhub_graph/merge.py {output[0]} -p "data/*.json"
         """)
+
+rule create_ro_crate:
+    input:
+        MERGED_FILE
+    params:
+        workflow_file = "Snakefile"
+    output:
+        directory("ro-crate-metadata/")
+    shell:
+        """
+        # Create a new virtual environment
+        python -m venv rocrate_env
+
+        # Activate the virtual environment
+        source rocrate_env/bin/activate
+
+        # Upgrade pip to avoid any potential issues
+        pip install --upgrade pip
+        
+        # pip uninstall urllib3
+
+        # Install required packages
+        pip install requests urllib3 rocrate rocrate-zenodo
+
+        # Run the create_ro_crate script
+        python workflowhub_graph/create_ro_crate.py {input} {params.workflow_file} {output}
+    
+        # Deactivate the virtual environment
+        deactivate
+
+        # Remove the virtual environment to clean up
+        rm -rf rocrate_env
+        """