salty-dog is a tool to validate JSON and YAML data using JSON schema rules. It can filter elements and validate
select parts of the document, supports multiple documents in the same stream or file, and can insert defaults during
validation.
Check out the getting started guide.
salty-dog is written in Typescript and requires make, node, and yarn to build. It can be built locally or in
a container.
Please see the build docs for more details.
salty-dog is distributed as a docker container and an npm package.
While the container is the preferred way of running salty-dog, it has one limitation: docker run combines
stdout and stderr, making it impossible to separate logs and the output document. Writing either the logs or dest
to a file works around this.
To run the Docker container: docker run --rm ssube/salty-dog:master
The latest semi-stable image is published to ssube/salty-dog:master. Containers are published based on both Alpine
Linux and Debian (currently Stretch). All of the available tags are listed here.
Rules are provided in the image at /salty-dog/rules. To use custom rules in the container, mount them with
-v $(pwd)/rules:/salty-dog/rules:ro and load them with --rules /rules/foo.yml.
The ssube/salty-dog container image can be run normally or interactively.
To validate a file or input normally:
> docker run --rm ssube/salty-dog:master --helpYou can also launch a shell within the container, using local rules:
> docker run \
--rm \
-it \
--entrypoint bash \
ssube/salty-dog:mastersalty-dog is also published as an npm package with a binary, so it can
be used as a CLI command or programmatically.
To install salty-dog for the current project:
> yarn add -D salty-dog
> $(yarn bin)/salty-dog --helpIt is also possible to install salty-dog globally, rather than within a project. However, this is
not recommended.
> yarn global add salty-dog
> export PATH="${PATH}:$(yarn global bin)"
> salty-dog --helpsalty-dog can run in a few different modes: check mode will report errors, fix mode will attempt to modify the
input document, and list mode will print the active set of rules.
By default, salty-dog will validate the structure and contents of the --source document. If all rules pass, the
document will be printed to --dest.
> cat examples/kubernetes-resources-pass.yml | salty-dog \
--rules rules/kubernetes.yml \
--tag kubernetes
...
[2019-06-15T23:53:34.223Z] INFO: salty-dog/19839 on cerberus: all rules passed
> cat examples/kubernetes-resources-fail.yml | salty-dog \
--rules rules/kubernetes.yml \
--tag kubernetes
...
[2019-06-15T23:56:04.764Z] ERROR: salty-dog/22211 on cerberus: some rules failed (errors=1)
The --source and --dest default to stdin and stdout, respectively, but a path may be provided:
> salty-dog \
--rules rules/kubernetes.yml \
--tag kubernetes \
--source examples/kubernetes-resources-pass.yml \
--dest /tmp/kubernetes-resource.yml
...
[2019-06-15T23:53:34.223Z] INFO: salty-dog/19839 on cerberus: all rules passedsalty-dog can also add default values to missing properties in fix mode. If a rule does not immediately pass
with the --source document, but defaults are provided in the schema, the defaults will be inserted before printing to
--dest.
> salty-dog fix \
--source examples/kubernetes-resources-some.yml \
--rules rules/kubernetes.yml \
--tag kubernetesProperties that appear in the schema with a default provided will be added to each element as it is checked. Rules
apply in order, as do their defaults.
Properties that appear in the document with a different type than they have in the schema may be coerced, if the
value is compatible with the schema type. The full matrix of valid type coercions
is documented by Ajv.
salty-dog can list the active set of rules, to help debug tags and inclusion. Both --source and --dest are
ignored in list mode.
> salty-dog list \
--rules rules/kubernetes.yml \
--tag kubernetes
...
[2019-06-30T18:39:11.930Z] INFO: salty-dog/26330 on cerberus: listing active rules
rules: [
{
"desc": "resource limits are too low",
"level": "debug",
"name": "kubernetes-resources-minimum-cpu",
...
]salty-dog uses node-bunyan for logging and prints structured JSON output.
Logs can be pretty-printed by redirecting stderr through bunyan itself or jq, both of which are installed in
the salty-dog container:
> cat resource.yml | salty-dog --rules rules/kubernetes.yml --tag kubernetes 2> >(bunyan)
...
[2019-06-15T23:53:34.223Z] INFO: salty-dog/19839 on cerberus: all rules passed
> cat resource.yml | salty-dog --rules rules/kubernetes.yml --tag kubernetes 2> >(jq)
...
{
"name": "salty-dog",
"hostname": "cerberus",
"pid": 19839,
"level": 30,
"msg": "all rules passed",
"time": "2019-06-15T23:53:34.223Z",
"v": 0
}Using jq allows for additional filtering and formatting. For example, jq 'select(.level > 30)' will only print
warnings and errors (the minimum log level to print can be set in the configuration file).
To print the last line's message and error messages: tail -1 | jq '[.msg, try (.errors[] | .msg)]'
> cat test/examples/kubernetes-resources-high.yml | salty-dog \
--rules rules/kubernetes.yml \
--tag kubernetes 2> >(tail -1 | jq '[.msg, try (.errors[] | .msg)]')
[
"all rules passed"
]
> cat test/examples/kubernetes-resources-some.yml | salty-dog \
--rules rules/kubernetes.yml \
--tag kubernetes 2> >(tail -1 | jq '[.msg, try (.errors[] | .msg)]')
[
"some rules failed",
".resources.limits should have required property 'memory' at $.spec.template.spec.containers[*] for kubernetes-resources",
".metadata should have required property 'labels' at $ for kubernetes-labels"
]Rules combine a jsonpath expression and JSON schema to select and validate the document.
The rule's select expression is used to select nodes that should be validated, which are filtered, then checked.
The structure of rule files and the rules within them are documented here.
Rules can be loaded from a file, module, or path.
To load a file by name, --rule-file foo.yml. This will accept any extension.
To load a module, --rule-module foo. The required module exports are documented here.
To load a path, --rule-path foo/. This will recursively load any files matching *.+(json|yaml|yml).
All rules are disabled by default and must be enabled by name, level, or tag.
To enable a single rule by name, --include-name foo-rule.
To enable a group of rules by level, --include-level warn.
To enable a group of rules by tag, --include-tag foo.
To validate the rules in the rules/ directory using the meta-rules:
> make test-rules
...
{"name":"salty-dog","hostname":"cerberus","pid":29403,"level":30,"msg":"all rules passed","time":"2019-06-16T00:56:55.132Z","v":0}
![dependabot[bot] avatar](https://avatars.githubusercontent.com/in/29110?v=4 "dependabot[bot]")
![isolex[bot] avatar](https://avatars.githubusercontent.com/u/451959?v=4 "isolex[bot]")
![renovate[bot] avatar](https://avatars.githubusercontent.com/in/2740?v=4 "renovate[bot]")
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent