A simple and easy to use TOML validator for Python.
You can install the package from PyPI:
pip install tomlvalThe package is available for Python 3.11 and newer.
Handlers are the validation functions used to validate the value of keys in the input data.
A handler must be one of type, Callable. This means any object of type type is valid, and and Callable, such as lambda functions as well as named functions are valid.
The handler will dynamically be passed either the key and/or value argument based of what parameters are defined.
Examples of valid handlers are:
- Types:
str,int,datetime.datetime, ... - Anonymous functions:
lambda: ...,lambda key: ...,lambda value: ...,lambda key, value: ... - Named functions:
def my_fn(),def my_fn(key),def my_fn(value),def my_fn(key, value)
If a handler accepts any parameters which are not key or value, a TOMLHandlerError will be raised.
A handler returns an error, meaning nullish values tell the validator that the test passes. The reason for this design is that the handler may return error messages or any value your program needs.
A schema is an optional structure used to add functionality to the validator, this includes validation for missing keys and default handlers.
Keys follow the TOML specification, meaning keys must be in either snake_case or SCREAMING_SNAKE_CASE. This project adds some special notation in the form of suffixing a key with ? to make it optional, adding [] to the end to make the key a nested array as well as wildcard regex pattern support. The importance of keys are based of specificity, so my.key would dominate both my.* and *.
This means the following keys are examples of valid keys:
name,user.name: Specific key*_name,user.*,*name*,user.*.name: Wildcard keyslast_name?,user.name?,array?[].key: Optional keysarray[],array?[],array[].key: Nested arrays
All keys can be written in dot-notation, meaning a deeply nested object/array can be written in a simpler form. For example:
{
"very": {
"deeply": {
"nested": {
"object": {
"key": str
}
}
}
}
}can be written as "very.deeply.nested.object.key": str. This notation also supports optionality and arrays. This would work by just suffixing the word with ? and if an array, suffix the ? with [].
In order to define a new schema, you can use the following code as reference:
from tomlval import TOMLSchema
def my_fn(key, value):
return "some-error"
def default_handler() -> str:
""" Default handler for all keys """
return "invalid-key"
schema = TOMLSchema({
"single_type": str,
"multiple_types": (int, float),
"single_handler": lambda: "error-message",
"multiple_handlers": (lambda: "error-message", str, my_fn),
"optional?": str
"list_of_strings": [str],
"nested_dictionary": {
"key": str,
...
},
"nested_array": [
{
"key": str,
...
},
...
],
})Note: When a nested array includes dictionaries with different structures, they will be merged. If the merge fails, a TOMLSchemaMergeError will be raised.
The validator defines the blueprint for how data should be validated. This is defined in the optional schema, or handlers can be manually added using the add_handler(key, fn) method. Handlers, like keys, are prioritized based of the key priority.
This examples includes the most basic use case, where a default handler is defined manually:
from tomlval import TOMLValidator
validator = TOMLValidator()
validator.add_handler("*", lambda: "invalid-key")This example includes a schema, assume the schema is populated with the structure and handlers you require.
from tomlval import TOMLValidator, TOMLSchema
schema = TOMLSchema({...})
validator = TOMLValidator(schema)This example includes a case where you might have defined a shared schema somewhere in your code but you need to customize specific keys:
from tomlval import TOMLValidator
from .schema import schema
def validate_age(value):
if value <= 0:
return "value-to-low"
return None
validator = TOMLValidator(schema)
validator.add_handler("user.age", validate_age)For some people, it might not be the best option to return an error message, and instead some other value might be preferred or you might want a more verbose error message. In this case, the on_missing and on_type_mismatch callbacks can be changed changed:
from tomlval import TOMLValidator
from .schema import schema
def on_missing(key: str):
return f"'{key}' is missing"
def on_type_mismatch(key: str, expected: type, got: type)
return f"The argument '{key}' expected type '{expected.__name__}', got '{got.__name__}'"
validator = TOMLValidator(
schema,
on_missing=on_missing,
on_type_mismatch=on_type_mismatch
)Now that you have defined your schema and validator, the validator is now ready to be used on TOML data.
In order to use the validator, the validate(data) method is used. It accepts any dictionary as an argument and outputs a flat dictionary of all keys in dot-notation with each key's respective error value.
This example shows a use-case where a TOML file is validated.
import tomllib
from datetime import datetime
from pathlib import Path
from tomlval import TOMLSchema, TOMLValidator
# Read file
file_path = Path("example.toml")
with file_path.open("rb") as file:
data = tomllib.load(file)
# Define schema
schema = TOMLSchema({
"*_name": str,
"age": lambda value: "invalid-age" if age <= 0 else None,
"birthday": datetime,
"*": lambda: "invalid-key"
})
# Define validator
validator = TOMLValidator(schema)
# Validate data
errors = validator.validate(data)Instead of loading a file, you might have pre-loaded TOML-data in the form of a dictionary.
import tomllib
from datetime import datetime
from pathlib import Path
from tomlval import TOMLSchema, TOMLValidator
from .data import data
# Define schema
schema = TOMLSchema({
"*_name": str,
"age": lambda value: "invalid-age" if age <= 0 else None,
"birthday": datetime,
"*": lambda: "invalid-key"
})
# Define validator
validator = TOMLValidator(schema)
# Validate data
errors = validator.validate(data)This project is licensed under the MIT License - seea the LICENSE file for details.