This repository contains the DOWNLOAD service for OpenCitations.

The service requires the following environment variables. These values take precedence over the ones defined in conf.json:

• BASE_URL: Base URL for the DOWNLOAD endpoint
• LOG_DIR: Directory path where log files will be stored
• SYNC_ENABLED: Enable/disable static files synchronization (default: false)

For instance:

BASE_URL=download.opencitations.net
LOG_DIR=/home/dir/log/
SYNC_ENABLED=true

Note: When running with Docker, environment variables always override the corresponding values in conf.json. If an environment variable is not set, the application will fall back to the values defined in conf.json.

The application can synchronize static files from a GitHub repository. This configuration is managed in conf.json:

{
  [...]
  "oc_services_templates": "https://github.com/opencitations/oc_services_templates",
  "sync": {
    "folders": [
      "static",
      "html-template/common"
    ],
    "files": [
      "test.txt"
    ]
  }
}

• oc_services_templates: The GitHub repository URL to sync files from
• sync.folders: List of folders to synchronize
• sync.files: List of individual files to synchronize

When static sync is enabled (via --sync-static or SYNC_ENABLED=true), the application will:

1. Clone the specified repository
2. Copy the specified folders and files
3. Keep the local static files up to date

Note: Make sure the specified folders and files exist in the source repository.

For local development and testing, the application uses the built-in web.py HTTP server:

Examples:

# Run with default settings
python3 download_oc.py

# Run with static sync enabled
python3 download_oc.py --sync-static

# Run on custom port
python3 download_oc.py --port 8085

# Run with both options
python3 download_oc.py --sync-static --port 8085

The application supports the following command line arguments:

• --sync-static: Synchronize static files at startup and enable periodic sync (every 30 minutes)
• --port PORT: Specify the port to run the application on (default: 8080)

When running in Docker/Kubernetes, the application uses Gunicorn as the WSGI HTTP server for better performance and concurrency handling:

• Server: Gunicorn with gevent workers
• Workers: 4 concurrent worker processes
• Worker Type: gevent (async) for handling thousands of simultaneous requests
• Timeout: 1200 seconds (to handle long-running SPARQL queries)
• Connections per worker: 400 simultaneous connections

The Docker container automatically uses Gunicorn and is configured with static sync enabled by default.

Note: The application code automatically detects the execution environment. When run with python3 download_oc.py, it uses the built-in web.py server. When run with Gunicorn (as in Docker), it uses the WSGI interface. You can customize the Gunicorn server configuration by modifying the gunicorn.conf.py file.

You can change these variables in the Dockerfile:

# Base image: Python slim for a lightweight container
FROM python:3.11-slim

# Define environment variables with default values
# These can be overridden during container runtime
ENV BASE_URL="download.opencitations.net" \
    SYNC_ENABLED="true" \
    LOG_DIR="/mnt/log_dir/oc_download"


# Ensure Python output is unbuffered
ENV PYTHONUNBUFFERED=1

# Install system dependencies required for Python package compilation
# We clean up apt cache after installation to reduce image size
RUN apt-get update && \
    apt-get install -y \
    git \
    python3-dev \
    build-essential

# Set the working directory for our application
WORKDIR /website

# Clone the specific branch (download) from the repository
# The dot at the end means clone into current directory
RUN git clone --single-branch --branch main https://github.com/opencitations/oc_download .

# Install Python dependencies from requirements.txt
RUN pip install -r requirements.txt

# Expose the port that our service will listen on
EXPOSE 8080

# Start the application with gunicorn for production
CMD ["gunicorn", "-c", "gunicorn.conf.py", "download_oc:application"]