Linter for JSON Schema instances.
Install with pip:
pip install jsonschema-lint[yaml]
ℹ️ Leave out the
[yaml]suffix if you don't need YAML support.
The linter is invoked on the command line:
$ jsonschema-lint
By default, the linter will attempt to lint every file matching extension under the currect directory. This means every .json file, plus every .yaml/.yml file if PyYAML is installed. A file will only be linted if a matching schema can be detected (see below).
You can override this behaviour by passing arguments to the linter, e.g.
$ jsonschema-lint **/*.avsc
There are three ways schemas can be selected for a given instance. In order of priority:
- If provided, the
--schemaoption will be used to validate all target instances. - A matching rule in a
.jsonschema-lintfile, in the instance directory or its parents (see below). - If the
--schema-storeflag is provided, then matching rules from Schema Store will be used.
.jsonschema-lint files follow similar logic to .gitignore and Github CODEOWNERS files, in that they apply to all files under their directory tree.
Each line in a .jsonschema-lint file is a rule, matching a filepath pattern to a particular schema. For example:
**/.circleci/config.yml https://json.schemastore.org/circleciconfig.json yaml
There are three parts to each rule, separated by whitespace. These are:
- the pattern/glob, in this case matching any files named
config.ymlin a directory named.circleci - the location of the schema. This can be a remote URL, or a path on the local filesystem. If this is a relative path, it is resolved relative to the
.jsonschema-lintfile. - (optional) the expected file format of any instances. If this is omitted, the linter will attempt to detect the correct type from the file extension. If it cannot be detected, both will be attempted.
Install dependencies:
pyenv shell 3.9.4 # Or other 3.9.x
pre-commit install # Configure commit hooks
poetry install # Install Python dependenciesRun tests:
poetry run inv verifyThis project is distributed under the MIT license.