craft
How to use ?
Go
go install github.com/kilianpaquier/craft/cmd/craft@latest
Linux
OS="linux" # change it depending on your case
ARCH="amd64" # change it depending on your case
INSTALL_DIR="$HOME/.local/bin" # change it depending on your case
new_version=$(curl -fsSL "https://api.github.com/repos/kilianpaquier/craft/releases/latest" | jq -r '.tag_name')
url="https://github.com/kilianpaquier/craft/releases/download/$new_version/craft_${OS}_${ARCH}.tar.gz"
curl -fsSL "$url" | (mkdir -p "/tmp/craft/$new_version" && cd "/tmp/craft/$new_version" && tar -xz)
cp "/tmp/craft/$new_version/craft" "$INSTALL_DIR/craft"
Commands
Craft initializes or generates craft projects. Craft projects are only defined by a .craft file
and multiple files automatically generated to avoid multiple hours to setup Continuous Integration, coverage, security analyses, helm chart, etc.
Craft generation can be done with 'craft' command or 'craft generate' command.
Additional generation command are available to generate only subparts of craft layout (like 'craft chart').
Usage:
craft [flags]
craft [command]
Available Commands:
completion Generate the autocompletion script for the specified shell
generate Generate project layout
help Help about any command
init Initialize craft project
version Show current craft version
Flags:
-d, --dir string set directory where generation will be made (default is current directory)
-h, --help help for craft
--log-format string set logging format (either "text" or "json") (default "text")
--log-level string set logging level (default "info")
Use "craft [command] --help" for more information about a command.
Init
Initialize new craft project
Usage:
craft init [flags]
Flags:
-h, --help help for init
Global Flags:
-d, --dir string set directory where generation will be made (default is current directory)
--log-format string set logging format (either "text" or "json") (default "text")
--log-level string set logging level (default "info")
Generate
Generate project layout
Usage:
craft generate [flags]
Flags:
-h, --help help for generate
Global Flags:
-d, --dir string set directory where generation will be made (default is current directory)
--log-format string set logging format (either "text" or "json") (default "text")
--log-level string set logging level (default "info")
Craft file
Craft project generation is based on root's .craft
file, it can contain the following configurations:
# bot in charge of keeping dependencies up to date
bot: dependabot | renovate
# project's CI (optional)
# providing it will create the appropriate ci files (.gitlab-ci.yml, .github/actions, .github/workflows)
ci:
# auth configurations for various features in CI
auth:
# maintenance auth strategy for the specified maintenance bot (just above)
maintenance: github-app | github-token | personal-token
# release auth for github only (how should the release token be retrieved)
# will stay empty when using gitlab CICD
release: github-app | github-token | personal-token
docker:
# target path for docker push (i.e. kilianpaquier/craft)
# (optional, by default will be provisionned with <owner>/<repository>)
path: kilianpaquier/craft
# specific exposed port (optional, default is 3000)
# used in various places like helm values.yml service port
# Dockerfile exposed port
port: 3000
# specific docker registry to push images on
# (optional, by default the CI platform decides, both GitHub and GitLab CI/CD points to their own registry)
# used in various places like helm values.yml images registry
# github release workflow to push images
registry: ghcr.io
deployment:
# deployment automatisation on default branch
auto: true | false
# deployment platform name
platform: kubernetes | netlify | pages
helm:
# target path for helm cm-push (i.e. kilianpaquier/craft)
# (optional, by default will be provisionned with <owner>/<repository>)
path: kilianpaquier/craft
# whether to push the project helm chart on an helm repository or not
publish: auto | manual | none
# specific helm repository to push the generated chart on
# (optional, by default the CI platform decides, both GitHub and GitLab CI/CD points to their own registry)
# OCI based repositories can be used
registry: (oci://)?ghcr.io
# ci name - self-explaining what each value will generate - (required when ci section is given)
name: github | gitlab
# ci global options, providing one or multiple options with tune the ci generation (optional)
options:
- codecov
- codeql
- labeler
- pre-commit:auto-commit
- sonar
# release specific options
release:
# whether the release should run automatically
auto: true | false
# whether backmerging should be configured for main, staging and develop branches
backmerge: true | false
# project's description (optional)
# used in various places like helm Chart.yml description
# Dockerfile description label
description: some useful description
# list of parts to not generate.
exclude:
- goreleaser
- makefile
- pre-commit
- shell
# list of optional parts to generate.
include:
- pre-commit:gomod-tidy
- renovate:post-upgrade
# project's license (optional)
# providing it will download the appropriate license
# used in various places like goreleaser executables license
# github release workflow license addition to releases
license: agpl-3.0 | apache-2.0 | bsd-2-clause | bsd-3-clause | bsl-1.0 | cc0-1.0 | epl-2.0 | gpl-2.0 | gpl-3.0 | lgpl-2.1 | mit | mpl-2.0 | unlicense
# project's maintainers (at least one must be provided)
# the first maintainer will be referenced in various places like in goreleaser configuration
# Dockerfile maintainer / authors label
# sonar.properties organization and project key prefix
# helm values.yml for images owner (e.g ghcr.io/maintainer/app_name)
# all maintainers will be referenced in dependabot assignees and reviewers
# helm Chart.yml maintainers
maintainers:
- name: maintainer
email: maintainer@example.com
url: maintainer.example.com
# platform override in case of gitlab on premise, bitbucket on premise, etc.
# by default, an on premise gitlab will be matched if the host contains "gitlab"
# by default, an on premise bitbucket will be matched if the host contains "bitbucket" or "stash"
# when not overridden, the platform is matched based on "git config --get remote.origin.url" on the returned host (github.com, gitlab.com, ...)
platform: bitbucket | gitea | github | gitlab
VSCode association and schema
When working on vscode, feel free to use craft's schemas to help setup your project:
{
"files.associations": {
".craft": "yaml"
}
}
It's only creating the association between yaml files and .craft
, however combined with vscode extension redhat.vscode-yaml,
it will load the schema fine since a header is added in all .craft
when written.
Generations
Craft generation is based on separated parsers and separated generators.
- Each
parser
checks from .craft
configuration and project's files to add specific behaviors in a shared structure.
- Each
generator
generates a part of a project layout (for instance, a generator
is in charge of generating the helm chart).
Generation process first runs over all parsers to make sure a project is fully known with its specificities, then it runs over all generators.
Multiple examples:
- A
go.mod
is detected with Golang
parser, combined with ci
configuration, then the appropriate CI will be generated.
- A
go.mod
is detected with Golang
parser and a hugo.(toml|yaml|...)
or theme.(toml|yaml|...)
is detected too, combined with the ci
and deployment
options,
then the appropriate Netlify or Pages (it can be GitLab or GitHub) deployment will be generated in CI files.
- If
exclude: [chart]
is not given, a custom craft helm chart will be generated.
This helm chart can deploy cronjobs, jobs and workers easily from values.yaml
file.
- A
package.json
is detected with Node
parser, combined with ci
configuration, then the appropriate CI will be generated
(codecov analysis, sonar analysis, lint, tests, build if needed).
Who is using craft ?
Craft as an SDK
Craft can also be used as an SDK, for that you may check the official documentation on pkg.go.dev.