Use cases

The Wave CLI enables you to build Docker and Singularity containers from various sources, including Dockerfiles, Singularity definition files, file system directories, and Conda packages. The following sections describe several common use cases.

tip

To get started with an example Nextflow pipeline that uses Wave CLI, see Wave CLI.

Augment a container with a directory

The Wave CLI supports container augmentation with a specified directory. You can use container augmentation to dynamically add layers, such as scripts or configuration files, to your container.

Augment a container with a directory

Related CLI arguments

Directory builds support the following arguments:

• --layer: Specifies a directory containing layer content.
• --image, -i: Specifies an existing container image (default: docker.io). Accepts image names (e.g., alpine:latest) or image URLs (e.g., public.ecr.aws/docker/library/busybox).

Limitations

The following limitations apply:

• A base image must be specified.
• Each file must be no larger than 1 MB.
• A directory must be no larger than 10 MB, inclusive of all files.

Example usage

Create a custom layer with a shell script and add it to an Alpine base image:

1. Create a new context directory:

   mkdir -p new-layer/usr/local/bin
   printf 'echo Hello world!' > new-layer/usr/local/bin/hello.sh
   chmod +x new-layer/usr/local/bin/hello.sh

2. Build and run the image:

   docker run $(wave -i alpine --layer new-layer) sh -c hello.sh

Build a container from Conda packages

The Wave CLI supports building a container from a list of Conda packages.

Build a container from Conda packages

Related CLI arguments

Conda builds support the following arguments:

• --conda-base-image: Specifies the base image for installing Conda packages (default: mambaorg/micromamba:1.5.10-noble).
• --conda-channels: Specifies one or more comma-separated channels (default: seqera,bioconda,conda-forge,defaults).
• --conda-file: Specifies a Conda lock file path or URL.
• --conda-package: Specifies Conda packages to install. Supports expressions such as bioconda::samtools=1.17 or samtools>=1.0,<1.17. Accepts a comma-separated list or can be specified multiple times.
• --conda-run-command: Specifies a Docker RUN command to execute during the build. Can be specified multiple times.

Example usage

Build a container with specific versions of the samtools and bamtools packages:

wave \
  --conda-package bamtools=2.5.2 \
  --conda-package samtools=1.17

Build a container from a Dockerfile

The Wave CLI supports building a container from a Dockerfile. Specifying an optional build context allows the use of ADD and COPY commands in a Dockerfile.

note

Dockerfile builds that require --build-arg for build time variables aren't currently supported.

Build a container from a Dockerfile

Related CLI arguments

Container builds support the following arguments:

• --containerfile, -f: A Dockerfile to build. Build arguments aren't currently supported.
• --context: A directory that contains the context for the build.

Example usage

Build a container that installs several packages from a Dockerfile:

1. Create a Dockerfile:

   cat << EOF > ./Dockerfile
   FROM alpine
   
   RUN apk update && apk add bash cowsay \
           --update-cache \
           --repository https://alpine.global.ssl.fastly.net/alpine/edge/community \
           --repository https://alpine.global.ssl.fastly.net/alpine/edge/main \
           --repository https://dl-3.alpinelinux.org/alpine/edge/testing
   EOF

2. Build and run the container:

   container=$(wave --containerfile ./Dockerfile)
   docker run --rm $container cowsay "Hello world"

Build a container from a Dockerfile with a local build context:

1. Create a Dockerfile that references a local file:

   cat << EOF > ./Dockerfile
   FROM alpine
   ADD hello.sh /usr/local/bin/
   EOF

2. Create the shell script in the build context directory:

   mkdir -p build-context/
   printf 'echo Hello world!' > build-context/hello.sh
   chmod +x build-context/hello.sh

3. Build and run the container with the build context:

   docker run $(wave -f Dockerfile --context build-context) sh -c hello.sh

Build a Singularity container

The Wave CLI supports building Singularity containers. A target build repository, specified with the --build-repo argument, is required to build Singularity containers.

Build a Singularity container

Related CLI arguments

Singularity container builds support the following arguments:

• --build-repo: Specifies the target repository to save the built container.
• --freeze: Enables container freeze mode.
• --singularity, -s: Enables Singularity container builds.
• --tower-token: Specifies a Seqera access token to access private registry credentials stored in Platform (not required if the TOWER_ACCESS_TOKEN environment variable is set).
• --tower-workspace-id: Specifies a Seqera workspace ID (e.g., 1234567890) where credentials are stored. Requires --tower-token flag or TOWER_ACCESS_TOKEN environment variable to be set.

Limitations

The following limitations apply:

• The linux/arm64 platform is not currently supported.

Example usage

Augment a Docker base image and save it as a Singularity container:

wave -i alpine --layer context-dir/ --build-repo docker.io/user/repo

Build a Singularity container from a SingularityCE definition (.def) file:

wave -f hello-world.def --singularity --freeze --build-repo docker.io/user/repo

Build a Singularity container from Conda packages:

wave --conda-package bamtools=2.5.2 --conda-package samtools=1.17 --freeze --singularity --build-repo docker.io/user/repo

Build a container and freeze to a container registry

The Wave CLI supports building a container and persisting the container to a container registry, such as DockerHub. You can refer to this frozen container image in a Dockerfile or Nextflow pipeline in the same way as any other container.

Build a container and freeze to a container registry

Prerequisites

Ensure the following conditions are met:

• You created a Seqera access token
• You specified the destination container registry credentials in Seqera
• You specified the Seqera access token via either the TOWER_ACCESS_TOKEN environment variable or the --tower-token Wave command-line option

Related CLI arguments

Container freeze builds support the following arguments:

• --build-repo: Specifies the target repository to save the built container.
• --freeze: Enables container freeze mode.
• --tower-token: Specifies a Seqera access token for accessing private registry credentials (not required if the TOWER_ACCESS_TOKEN environment variable is set).
• --tower-workspace-id: Specifies a Seqera workspace ID (e.g., 1234567890) where credentials are stored.

Example usage

Freeze the alpine container image to a private Docker Hub registry:

wave -i alpine --freeze \
  --build-repo docker.io/<USER>/repo --tower-token <TOWER_TOKEN>

Mirror a container image to another registry

The Wave CLI supports mirroring, i.e., copying containers to a container registry of your choice.

Mirror a container image to another registry

Prerequisites

Ensure the following conditions are met:

• You created a Seqera access token.
• You specified the destination container registry credentials in Seqera.
• You specified the Seqera access token via either the TOWER_ACCESS_TOKEN environment variable or the --tower-token Wave command-line option.

Related CLI arguments

Container mirroring supports the following arguments:

• --mirror: Enables container mirror mode.
• --build-repo: Specifies the target repository to save the mirrored container.
• --tower-token: Specifies a Seqera access token to access private registry credentials stored in Platform (not required if the TOWER_ACCESS_TOKEN environment variable is set).

Example usage

Mirror the samtools:0.1.16--2 container image to a private Docker Hub registry:

wave -i quay.io/biocontainers/samtools:0.1.16--2 --mirror \
  --build-repo docker.io/<USER>/containers --tower-token <TOWER_TOKEN>