Dockseed

Dockseed is a deterministic Python developer tool that inspects an existing Python project and generates a project-specific, best-practice Dockerfile and .dockerignore automatically.

Key Features

• Evidence-Based Detection: Inspects pyproject.toml, requirements.txt, Pipfile, .python-version, .tool-versions, and runtime.txt to resolve compatible Python versions (3.10–3.14).
• Modern Package Manager Support: Automatically configures layer-cached dependency installation for pip, uv, Poetry, PDM, Hatch, and Pipenv.
• Safe AST Static Analysis: Discovers application entry points (FastAPI, Flask, Django, Streamlit, Gradio, Celery, CLI scripts, and standalone scripts) using Python's built-in ast module—never importing or executing project code.
• Native System Dependency Heuristics: Detects packages requiring C extensions or shared libraries (psycopg2, mysqlclient, opencv-python, pyaudio, git+https, etc.) and injects clean apt-get layers.
• Secure & Production-Minded Defaults: Generates single-stage python:3.x-slim images running as an unprivileged appuser, using JSON exec-form CMD arrays for proper SIGTERM signal handling, and excluding .env secrets and virtual environments.
• CI/CD & Headless Ready: Deterministic output, explicit CLI override flags, and strict exit codes.

Installation

pip install dockseed

Or run directly with uvx or pipx:

uvx dockseed scan

CLI Usage

1. Analyze a Project (dockseed scan)

Inspect a project without modifying any files:

dockseed scan
dockseed scan ./my-project
dockseed scan --format json

2. Generate Configuration (dockseed init)

Generate a Dockerfile and .dockerignore:

dockseed init
dockseed init --dry-run
dockseed init --yes
dockseed init --overwrite
dockseed init --no-comments

Bypass ambiguity prompts in CI/CD pipelines using explicit override flags:

dockseed init --yes \
  --python 3.12 \
  --package-manager uv \
  --entrypoint "app.main:app" \
  --port 8000

3. Validate Configuration (dockseed check)

Validate an existing Dockerfile and .dockerignore statically, or perform an opt-in container build check:

dockseed check
dockseed check --build
dockseed check --build --runtime podman

Python API

Dockseed's scanning, generation, and validation engine can be used directly from Python:

from pathlib import Path
from dockseed import (
    scan_project,
    generate_dockerfile,
    generate_dockerignore,
    validate_dockerfile,
)

model = scan_project(Path("."))
dockerfile = generate_dockerfile(model, include_comments=True)
dockerignore = generate_dockerignore(model)

report = validate_dockerfile(dockerfile.content, model, dockerignore.content)
if report.is_valid:
    print(dockerfile.content)

Exit Codes

License

MIT