Skip to content

Latest commit

 

History

History
47 lines (30 loc) · 2.64 KB

README.md

File metadata and controls

47 lines (30 loc) · 2.64 KB
Raw

Overview

The scripts here are to be used for building all, or any combination of dockers that are listed in dockerfiles directory (and have their dependencies specified in build_docker.py).

We support both building from your local git repo, and remote GitHub repo.

Dependencies

This set of python script depends on Python 3.

Conda packages termcolor and pprint are assumed to be available.

Usage

  1. Checkout your working branch:

    $ cd ${REPO_ROOT}
$ git checkout my_branch

  2. Run the docker build script:

    $ cd scripts/docker
$ python build_docker.py --targets sv-pipeline --image-tag my-branch-a3fd92 --docker-repo myrepo

    This example will build and push myrepo/sv-pipeline:my-branch-a3fd92. All required base images will also be included automatically. Additional options can be viewed by running python build_docker.py.

Notes

  • Multiple target images can be specified like so: --targets sv-pipeline wham manta or --targets all

  • To build docker(s) that you desire, we recommend you look at build_docker_template.ipynb, which consists of example use cases when building from local and remote git resources

  • A full build (e.g. using --targets all) may take a few hours

  • Due to license restrictions, we cannot provide the binaries required for running the MELT docker. Users must obtain a license and download MELTv2.0.5_patch.tar.gz to dockerfiles/melt in order to build the MELT docker for module 00a. Alternatively, users may run MELT locally, and upload the VCFs to a GCS bucket to use as input to modules 00b and 00c.

  • When building from local files, we cautiously refuse to build when there are uncommitted changes, and/or un-tracked files, unless you specifically turn off that protection with --disable-git-protect.

  • When building from GitHub tag/hash values, we assume you'd provide an existing temporary locations where we'll pull the files to. When we are done, the pulled files will be cleaned up, but not the temporary location you provided. When you'd like to pull from GitHub using SSH instead of https, simply turn on flag --use_ssh.

  • test_build_docker.ipynb is for debugging and testing purposes

  • Use update_json_docker.sh to update dockers tags across all json parameter files

  • When building on Apple Silicon macs (newer macs with e.g. the M1 or M2 chipset), ensure that Rosetta is installed: softwareupdate --install-rosetta. Note that docker software is still somewhat unstable on these devices, so we cannot fully support them. We have had the most success by restricting docker desktop to 1 CPU (otherwise builds often churn forever without progress).