jinja2-jsonschema 0.4.0

JSON/YAML schema validation within Jinja2 templates

JSON Schema Validation within Jinja2 Templates

A Jinja2 extension providing a Jinja2 filter and a Jinja2 test for validating data against a JSON/YAML schema within Jinja2 templates.

Installation

• With pip:

  pip install jinja2-jsonschema
  
  # ... or with YAML support
  pip install jinja2-jsonschema[yaml]
  

• With poetry:

  poetry add jinja2-jsonschema
  
  # ... or with YAML support
  poetry add jinja2-jsonschema -E yaml
  

• With pdm:

  pdm add jinja2-jsonschema
  
  # ... or with YAML support
  pdm add jinja2-jsonschema[yaml]
  

• With pipx (injected into the pipx-managed virtual env of a package):

  pipx inject PACKAGE jinja2-jsonschema
  
  # ... or with YAML support
  pipx inject PACKAGE jinja2-jsonschema[yaml]
  

Usage

The extension provides:

• A Jinja2 filter which receives a schema file path or schema object as input and returns a jsonschema.ValidationError object when validation fails and an empty string ("") otherwise.
• A Jinja2 test which receives a schema file path or schema object as input and returns False when validation fails and True otherwise.

The JSON Schema dialect is inferred from the $schema field in the JSON Schema document and, when omitted, defaults to the latest dialect supported by the installed jsonschema library. Both local and remote schemas are supported including schema references and JSON Pointers.

Local schema files are loaded via a Jinja2 loader in which case configuring the Jinja2 environment with a loader is mandatory.

Some example usage of the JSON Schema validation filter and test is this:

from jinja2 import Environment
from jinja2 import FileSystemLoader


env = Environment(
    # Register a loader (only necessary when using local schema files).
    loader=FileSystemLoader("/path/to/templates"),
    # Register the extension.
    extensions=["jinja2_jsonschema.JsonSchemaExtension"],
)

# Example using an inline schema object.
template = env.from_string("{{ age | jsonschema({'type': 'integer', 'minimum': 0}) }}")
template.render(age=30)  # OK
template.render(age=-1)  # ERROR
template = env.from_string("{{ age is jsonschema({'type': 'integer', 'minimum': 0}) }}")
template.render(age=30)  # --> `True`
template.render(age=-1)  # --> `False`

# Example using a local schema file.
template = env.from_string("{{ age | jsonschema('age.json') }}")
template.render(age=30)  # OK
template.render(age=-1)  # ERROR
template = env.from_string("{{ age is jsonschema('age.json') }}")
template.render(age=30)  # --> `True`
template.render(age=-1)  # --> `False`

# Example using a remote schema file.
template = env.from_string("{{ age | jsonschema('https://example.com/age.json') }}")
template.render(age=30)  # OK
template.render(age=-1)  # ERROR
template = env.from_string("{{ age is jsonschema('https://example.com/age.json') }}")
template.render(age=30)  # --> `True`
template.render(age=-1)  # --> `False`

Usage with Copier

The extension integrates nicely with Copier, e.g. for validating complex JSON/YAML answers in the Copier questionnaire. For this, add the extension as a Jinja2 extension in copier.yml and use the Jinja2 filter in the validator field of a Copier question. For instance:

_jinja_extensions:
  - jinja2_jsonschema.JsonSchemaExtension

complex_question:
  type: json # or `yaml`
  validator: "{{ complex_question | jsonschema('schemas/complex.json') }}"

In this example, a local schema file schemas/complex.json is used whose path is relative to the template root. To prevent copying schema files to the generated project, they should be either excluded

+_exclude:
+  - schemas/
 _jinja_extensions:
   - jinja2_jsonschema.JsonSchemaExtension

or the project template should be located in a subdirectory such as template/:

+_subdirectory_: template
 _jinja_extensions:
   - jinja2_jsonschema.JsonSchemaExtension

Finally, template consumers need to install the extension along with Copier. For instance with pipx:

pipx install copier
pipx inject copier jinja2-jsonschema

Contributions

Contributions are always welcome via filing issues or submitting pull requests. Please check the contribution guide for more details.