How to Use Dockerfiles

GitHub actions are triggered for Pull Requests to the "develop" and "main_v*" branches and for pushes to the branches "feature_", "bugfix_", "develop", "develop-ref", and "main_v*" created in this repository to automatically build images from these Dockerfiles and push them to DockerHub repositories. The MET Docker image is used in Dockerfiles in the METplus wrappers component repository for testing.

Descriptions of the Dockerfiles in this directory are provided below. Sample commands for building these Docker images locally are also provided. Please note that these commands should be run from the top-level directory of the MET repository.

1. The first example below shows how to manually build a MET Docker image using code that is cloned from the MET GitHub repository⁠ and using a Docker image for the base compilation environment for MET from the dtcenter/met-base DockerHub repository. It uses Dockerfile, which checks out MET from GitHub, compiles the specified branch or tag from source, sets a working directory, downloads and installs GhostScript fonts, and deletes the MET source code for tagged releases matching "v"*. Tagged versions of the MET Docker image are available in the dtcenter/met DockerHub repository.

docker build -t dtcenter/met:${TAG_NAME} --build-arg SOURCE_BRANCH=${BRANCH_NAME} internal/scripts/docker
docker push dtcenter/met:${TAG_NAME}

where:

• ${TAG_NAME} is the name of the DockerHub tag to create
• ${BRANCH_NAME} is the MET branch to checkout

1. The second example below shows how to manually build a MET Docker image using a specified branch or tag from local source code and using a local Docker image for the base compilation environment for MET from the METbaseimage⁠ GitHub repository. (See the METbaseimage README.md⁠ file for more information on manually creating the Docker image for the base compilation environment.) This example uses Dockerfile.copy, which compiles MET using the specified branch or tag from local source code, sets a working directory, downloads and installs GhostScript fonts, and deletes the MET source code for tagged releases matching "v"*.

docker build -t dtcenter/met:${TAG_NAME} --build-arg SOURCE_BRANCH=${BRANCH_NAME} -f internal/scripts/docker/Dockerfile.copy .
docker push dtcenter/met:${TAG_NAME}

where:

• ${TAG_NAME} is the name of the DockerHub tag to create
• ${BRANCH_NAME} is the identifier to use for $MET_GIT_NAME inside image

1. The third example below shows how to manually build a MET Docker image using a specified branch or tag from local source code and using a local Docker image for the base compilation environment along with the additional packages required for running the MET unit tests from the METbaseimage⁠ GitHub repository. (See the METbaseimage README.md⁠ file for more information on manually creating the Docker image with the base compilation environment and with the additional packages required for running the MET unit tests.) This example also uses Dockerfile.copy, which is described above.

docker build -t dtcenter/met:${TAG_NAME} --build-arg SOURCE_BRANCH=${BRANCH_NAME} --build-arg MET_BASE_TAG=${MET_BASE_TAG} -f internal/scripts/docker/Dockerfile.copy .
docker push dtcenter/met:${TAG_NAME}

where:

• ${TAG_NAME} is the name of the DockerHub tag to create
• ${BRANCH_NAME} is the identifier to use for $MET_GIT_NAME inside image
• ${MET_BASE_TAG} is the version of dtcenter/met-base-unit-test to be used