Skip to content

Docker Usage

This page describes how to build and run PhosKinTime inside a Docker container.

Dockerfile Overview

The included Dockerfile uses python:3.10-slim as the base image and installs dependencies via Poetry.

FROM python:3.10-slim

RUN pip install poetry

WORKDIR /app
COPY pyproject.toml poetry.lock* /app/
RUN poetry install --no-root

COPY . /app

CMD ["poetry", "run", "phoskintime"]

Default command: poetry run phoskintime — runs the phoskintime CLI entry point.

Build

docker build -t phoskintime .

Run

Run the default CLI

The Dockerfile uses CMD (not ENTRYPOINT), so you must include the full poetry run phoskintime invocation when overriding the command:

docker run --rm phoskintime poetry run phoskintime --help

Run with your data mounted

Mount your local data/ directory and a results output directory into the container:

docker run --rm \
  -v "$PWD/data:/app/data" \
  -v "$PWD/results:/app/results_model" \
  phoskintime poetry run phoskintime all

Run a specific pipeline stage

# Preprocessing only
docker run --rm \
  -v "$PWD/data:/app/data" \
  -v "$PWD/results:/app/results_model" \
  phoskintime poetry run phoskintime prep

# Kinase optimization (local solver)
docker run --rm \
  -v "$PWD/data:/app/data" \
  -v "$PWD/results:/app/results_model" \
  phoskintime poetry run phoskintime kinopt --mode local

Run the global model

docker run --rm \
  -v "$PWD/data:/app/data" \
  -v "$PWD/results:/app/results_model_global" \
  -v "$PWD/config.toml:/app/config.toml" \
  --entrypoint poetry \
  phoskintime run phoskintime-global

Pass a custom config file

docker run --rm \
  -v "$PWD/my_config.toml:/app/config.toml" \
  -v "$PWD/data:/app/data" \
  -v "$PWD/results:/app/results_model" \
  phoskintime poetry run phoskintime all

The container reads config.toml from /app/ (the WORKDIR). Mount your own config file to override the defaults.

Launch the Dashboard

The Streamlit dashboard requires exposing a port:

docker run --rm \
  -v "$PWD/results_model_global:/app/results_model_global" \
  -p 8501:8501 \
  --entrypoint poetry \
  phoskintime run streamlit run run_dashboard.py -- --output-dir results_model_global

Then open http://localhost:8501 in your browser.

Limitations

  • The Docker image uses python:3.10-slim, which satisfies the >=3.10,<3.14 requirement.
  • Multi-core parallelism inside Docker is limited by the host's CPU allocation. Use --cpus to set limits, and set cores in config.toml accordingly.
  • The container does not persist results by default — always mount a host volume for output directories to avoid losing results when the container exits.
  • Numba JIT compilation occurs on first run inside the container; subsequent runs reuse the compiled cache if the container filesystem is not discarded.
  • The Dockerfile installs --no-root initially for caching, then copies the full source. If you modify source files locally, rebuild the image.