Skip to content

Validation

Overview

Dynaconf allows the validation of settings parameters, for cases where you want to validate the settings before starting the program.

Usage

To define validation rules, you must create Validator objects, which are constructed using keys (positional arguments) and rules (kwargs). For example:

Validator("AGE", gte=20, lte=80) # multiple rules
Validator("NAME", "VERSION", "PORT", must_exist=True) # multiple keys
Validator("DB.PORT", eq=8080) # keys with dot-path notation

In the first sections, we'll see some different ways to use Validator objects. For an extensive description of available validators, see the Reference section.

With Python

On instantiation

When you instantiate your settings, the declared validators won't be immediately run. They'll be first triggered upon trying to access some setting, and only the first exception will be raised:

from dynaconf import Dynaconf, Validator

settings = Dynaconf(
    settings_file=["settings.toml"],
    environments=True,
    validators=[
        # Ensure some parameter meets a condition
        Validator("AGE", lte=30, gte=10),
        # Only the first ValidationError will be raised
        Validator("NAME", eq="John"),
    ],
)

# The following three lines will trigger a validation error:
print(settings.age)
print(settings.validators)
print(settings.as_dict())
[default]
version = "1.0.0"
age = 35
name = "Bruno"
DEV_SERVERS = ['127.0.0.1', 'localhost', 'development.com']
PORT = 8001
JAVA_BIN = "/usr/bin/java"

[production]
PROJECT = "This is not hello_world"
$ python app.py
(...)
dynaconf.validator.ValidationError: AGE must lte 30 but it is 35 in env DEVELOPMENT

Warning

Sometimes an external library might implicitly access your settings object through module inspection. This may may trigger an undesired validation, even without an explicit access call in your code.

If you are running into early validation triggering problems, consider if this is not the case.

Lazy validation

Instead of passing validators= argument to Dynaconf class you can register validators after the instance is created and trigger validation manually.

Register

First, register some validators. This won't trigger the validation yet.

settings = Dynaconf()

settings.validators.register(
    Validator("MYSQL_HOST", eq="development.com", env="DEVELOPMENT"),
    Validator("MYSQL_HOST", ne="development.com", env="PRODUCTION"),
)
Trigger manually

You may choose two strategies for the validation:

  • validate: raises ValidationError on the first error found.
  • validate_all: raises ValidationError at the end. Accumulative error data is stored at details.
# raises on first error found
settings.validators.validate()

# raises after all possible errors are evaluated
try:
    settings.validators.validate_all()
except dynaconf.ValidationError as e:
    accumulative_errors = e.details
    print(accumulative_errors)
Trigger on data update

By default, if the data of an instance is updated with update, set or load_file methods, no validation will be triggered.

You can override this globally with the option validate_on_update or set this on a per-call basis.

# validate_on_update=False (default)
settings.update({"NEW_VALUE": 123}, validate=True) # triggers validators.validate()
settings.update({"NEW_VALUE": 123}, validate="all") # triggers validators.validate_all()

# validate_on_update=True or "all"
settings.update({"NEW_VALUE": 123}) # will trigger with the global strategy

With CLI

It is possible to define validators in a TOML file called dynaconf_validators.toml placed in the same folder as your settings files. For more information, see the CLI section.

dynaconf_validators.toml is equivalent to the program below:

[default]

version = {must_exist=true}
name = {must_exist=true}
password = {must_exist=false}

# dot notation is also supported
'a_big_dict.nested_1.nested_2.nested_3.nested_4' = {must_exist=true, eq=1}

  [default.age]
  must_exist = true
  lte = 30
  gte = 10

[production]
project = {eq="hello_world"}

Then to fire the validation use:

$ dynaconf validate

This returns code 0 (success) if validation is ok.

Info

All values in dynaconf are parsed using toml format, TOML tries to be smart and infer the type of the settings variables, some variables will be automatically converted to integer:

FOO = "0x..."  # hexadecimal
FOO = "0o..."  # Octal
FOO = "0b..."  # Binary

All cases are on toml specs https://github.com/toml-lang/toml/blob/master/toml.abnf

If you need to force a specific type casting there are 2 options.

  1. Use double quoted for strings ex: `FOO = "'0x...'" will be string.
  2. Specify the type using @ ex: FOO = "@str 0x..." (available converters are @int, @float, @bool, @json)

Features

Multiple Validators

A single validator can have multiple conditions.

Validator(
  "NAME",
  ne="john",
  len_min=4,
  must_exist=True, # redundant but allowed
  startswith="user_",
  cast=str,
  condition=lambda v: v not in FORBIDDEN_USERS,
  ...
)

But it can also be expressed in separate validators. Notice that order matters because validators are evaluated in the given order.

validators = [
  Validator("NAME", ne="john"),
  Validator("NAME", len_min=4),
  Validator("NAME", must_exist=True),
  Validator("NAME", startswith="user_"),
]

Custom messages

Messages can be customized by passing a messages argument to the Validator constructor. This argument must be passed a dict with one of the valid keys, shown in the dict below which contains the default messages:

{
    "must_exist_true": "{name} is required in env {env}",
    "must_exist_false": "{name} cannot exists in env {env}",
    "condition": "{name} invalid for {function}({value}) in env {env}",
    "operations": (
        "{name} must {operation} {op_value} "
        "but it is {value} in env {env}"
    ),
    "combined": "combined validators failed {errors}",
}

Note that these default messages also show all the variables that can be interpolated to each message.

Example

Validator(
    "VERSION",
    must_exist=True,
    messages={"must_exist_true": "You forgot to set {name} in your settings."}
)

Default values

Validators can be used to provide default values, which can be either static or computed.

Static default values

If Dynaconf fails to load a value for the given setting, it'll give it the exact value provided in default.

Validator("FOO", default="A default value for foo")

Warning

YAML reads empty keys as None and in that case defaults are not applied, if you want to change it set apply_default_on_none=True either globally to Dynaconf class or individually on a Validator.

Computed default values

Sometimes you need some values to be computed by calling functions.

Eager evaluation

If you want the function to be run during validation time, define it with this signature:

def my_function(settings, validator):
    return "this is computed during validation time"

and pass it directly to the default kwarg:

Validator("FOO", default=my_function)
Lazy evaluation

If you want the default to be evaluated when the value is first accessed, define my_function with this signature:

def my_lazy_function(value, **context):
    """
    value: Default value passed to the validator, defaults to `empty`
    context: A dictionary containing
            env: All the environment variables
            this: The settings instance
    """
    return "When the first value is accessed, then the my_lazy_function will be called"

and pass it to the default kwarg like so:

from dynaconf.utils.functional import empty
from dynaconf.utils.parse_conf import Lazy

Validator("FOO", default=Lazy(empty, formatter=my_lazy_function))

Casting / Transform

Validators can be used to postprocess your settings after being loaded from the files by passing a Callable to the cast argument. For example, they can be used to cast their type.

This Callable will get called with the setting's value as its single argument (note that this could potentially be a list or dict depending on your setting path) and its return value will be assigned to the provided setting path.

Note that you can pass any Callable such as regular function or a class/type.

Example

Given this settings.toml

name = 'Bruno'
colors = ['red', 'green', 'blue']

Validators can be passed a cast attribute

settings = Dynaconf(
    validators=[
        # Order matters here
        Validator("name", len_eq=5),
        Validator("name", len_min=1),
        Validator("name", len_max=5),
        # This will cast the str to list
        Validator("name", cast=list),
        # From this point on Validation pipeline
        # `name` will be a list of chars
        # and this will affect the settings.NAME

        Validator("colors", len_eq=3),
        Validator("colors", len_eq=3),
        # this will cast the list to str
        Validator("colors", len_eq=24, cast=str),
        # From this point on Validation pipeline
        # `colors` will be a str of 24 chars
        # and this will affect the settings.COLORS

    ],
)

assert settings.name == ['B', 'r', 'u', 'n', 'o']
assert type(settings.name ) == list
assert settings.colors == '["red", "green", "blue"]'
assert type(settings.colors) == str

Custom conditional expressions

The condition argument expects a Callable that receives the setting's value as its only argument and returns a bool.

To pass the validation, the condition function must return True (or a truthy type, such that bool(x) == True). If the returned value is False (or a falsy type, such that bool(x) == False) then the condition fails, and a ValidationError is risen.

Example

Validator("VERSION", condition=lambda v: v.startswith("1."))


def user_must_be_chuck_norris(value):
    return value == "Chuck Norris"

Validator("USER", condition=user_must_be_chuck_norris)

Conditional Validation

In some cases you might want to perform a validation only when another validator passes. To do this, use the when parameter.

Simple Example

Say you want to ensure that the DATABASE.HOST is set only when DATABASE.USER is set. To do so, pass another Validator to when:

Validator(
    "DATABASE.HOST",
    must_exist=True,
    when=Validator("DATABASE.USER", must_exist=True)
)

Complex Example

Say you want to validate that DATABASE.CONNECTION_ARGS is set only when DATABASE.URI starts with "sqlite://". This will do that for you:

Validator(
    "DATABASE.CONNECTION_ARGS",
    must_exist=True,
    when=Validator("DATABASE.URI", condition=lambda v: v.startswith("sqlite://")),
    messages={"must_exist_true": "{name} is required when DATABASE is SQLite"}
)

Combination Operators

Validators can be combined using | or &.

or operator (|)

Validator('DATABASE.USER', must_exist=True) | Validator('DATABASE.KEY', must_exist=True)

which generates a single Validator which succeeds if any of these pass.

and operator (&)

Validator('DATABASE.HOST', must_exist=True) & Validator('DATABASE.CONN', must_exist=True)

which generates a single Validator which succeeds only if both of these pass.

Selective Validation

New in 3.1.6

You can also choose what sections of the settings you do or don't want to validate. This is achieved by using the validate_only and validate_exclude kwargs to Dynaconf (read more in the Configuration page) or by passing only or exclude to the validate method of the Dynaconf.validators list.

This is useful when:

  • You want to add additional validators after the settings object is created.
  • You want settings validated only when certain sections of your project are loaded.
  • You want to offer incremental configuration levels, validating only what is needed.

Note that exclusions are applied after selections. Therefore, if you pass a settings path in only that also matches a path passed in exclude, it'll end up excluded.

A settings path starts at the top level element and can be specified down to the lowest component. For example: my_settings.server.user.password can have the following settings paths passed in server, server.user, server.user.password.

Note: Selective validation matches the passed in value(s) to settings paths that start with that value. This means that passing exclude="FOO" will exclude not only paths that start with FOO but also FOOBAR.

Example

In this example, the behavior is as follows:

  1. Dynaconf will only validate any setting that starts with server upon instantiation.
  2. Only settings that start with module1 will be validated when module1.py is first run.
  3. When module2.py is first run, only settings that start with module2 (excluding those that start with module2.bad) will be validated.

config.py

# create a settings object, validating only settings under settings.server
settings = Dynaconf(
    validators=[
        Validator(
            "server.hostname",
            "server.port",
            "server.auth",
            must_exist=True
        ),
        Validator(
            "module1.value1",
            "module1.value2",
            "module1.value3",
            must_exist=True
        ),
        Validator(
            "module2.value1",
            "module2.value2",
            "module2.bad",
            must_exist=True
        )
    ],
    validate_only="server"
)

module1.py

# call validation on module1 settings
settings.validators.validate(only=["module1"])

module2.py

# call validation on module2 settings
# ignore validation for a subsection of module2's settings
settings.validators.validate(
    only=["module2"],
    exclude=["module2.bad"]
)

Validating only the current_env

You can specify if you want to validate all environments defined for a validator (default behavior) or only the current environment. To do so, you may use the only_current_env argument of single Validators (i.e., in Validator.validate), the same argument on ValidatorLists (such as Dynaconf.validators.validate) or by passing the validate_only_current_env kwarg to Dynaconf (see Configuration).

In the first case, the validators will run on all possible settings defined in their list of environments, while in the latter the validators with environments different from the current environment will be skipped.

This is useful when your configuration for different environments (let's say production and development) comes from different files you don't necessarily have access to during development. You would want to write different validators for your development and production environments, and only run the right validator for the current environment.

It's also useful if certain settings are only required in production.

Example

Using:

settings.toml

[development]
version = "dev"
age = 35
name = "Bruno"
servers = ['127.0.0.1', 'localhost', 'development.com']
PORT = 80

[production]
version = "1.0.0"
age = 35
name = "Bruno"
servers = ['production.com']
PORT = 443

.secrets.toml

[production]
api_key = 'secret_api_key'

You could then have these validators:

from dynaconf import Dynaconf, Validator

settings = Dynaconf(
    settings_files=['setting.toml', '.secrets.toml'],
    environments=True,
    validators=[
        # Ensure some parameters exist for both envs
        Validator('VERSION', 'NAME', 'SERVERS', env=['development', 'production'], must_exist=True),

        # Ensure some parameter validate certain condition in dev env
        Validator("SERVERS", env='development', cont='localhost'),

        # Ensure some parameter exists in production env
        Validator('API_KEY', env='production', must_exist=True),
    ]
)

And suppose during development, when settings.current_env == 'development', you don't have the file .secrets.toml.

Running settings.validators.validate() will fail even if settings.current_env == 'development', because by default all validators will run on all of their environments, whether or not it is the current env.

If you instantiate your settings with the parameter validate_only_current_env=True, no errors will be raised if settings.current_env == 'development', but a ValidationError will raise if settings.current_env == 'production'. This forces you to have the .secrets.toml file in production but not during development.

Reference

Validators can be created by passing the following arguments:

# names: list[str]
# can be a single or multiple positional strings
Validator('VERSION', 'AGE', 'NAME'),
# can also use dot notation
Validator('DATABASE.HOST', 'DATABASE.PORT'),
Validator('DATABASE.HOST', 'DATABASE.PORT'),


# must_exist: bool (alias: required)
# Check whether variable must or not exist
Validator('VERSION', must_exist=True),
Validator('PASSWORD', must_exist=False),
# there is an alias for must_exist called `required`
Validator('VERSION', required=True),

# condition: callable
# A function or any other callable that accepts `value` and
# must return a boolean
Validator('VERSION', condition=lambda v: v.startswith("1.")),

# when: Validator
# Conditionally runs the validator only when the passed validator passes
Validator(
    'VERSION',
    condition=lambda v: v.endswith("-dev"),
    when=Validator('ENV_FOR_DYNACONF', eq='development')
),

# env: str
# Runs the validator against the specified env, only for
# settings that are loaded from files with environments=True
Validator('VERSION', must_exist=True, env='production'),

# messages: dict[str, str]
# A dictionary with custom messages for each validation type
Validator(
    "VERSION",
    must_exist=True,
    condition=lambda v: v.startswith("1."),
    messages={
        "must_exist_true": "You forgot to set {name} in your settings.",
        "condition": "The allowed version must start with 1., you passed {value}"
    }
),

# cast: callable/class
# A type or a callable to transform the type of the passed object
# can also be used to apply transformations/normalizations
Validator("VERSION", cast=str),
Validator("VERSION", cast=lambda v: v.replace("1.", "2.")),
Validator("STATIC_FOLDER", cast=Path)
# Cast will be called for default values and also for values defined on
# settings via files or envvars

# default: any value or a callable
# If the value is not found it will be set to the default value
# if the default is a callable it will be called with the
# settings and instance of validator as arguments.
def default_connection_args(settings, validator):
    if settings.DATABASE.uri.startswith("sqlite://"):
        return {"echo": True}
    else:
        return {}

Validator("DATABASE.CONNECTION_ARGS", default=default_connection_args),
# default will be called only if the value is not explicitly set on settings
# via files or env vars.

# description: str
# As of 3.1.12 dynaconf doesn't use this for anything
# but there are plugins and external tools that uses it.
# this value to generate documentation
Validator("VERSION", description="The version of the app"),

# apply_default_on_none: bool
# YAML parser parses empty values as `None` so in this case
# you might want to force the application of default when the
# value in settings is `None`
Validator("VERSION", default="1.0.0", apply_default_on_none=True),


# Operations: comparison operations
# - eq: value == other
# - ne: value != other
# - gt: value > other
# - lt: value < other
# - gte: value >= other
# - lte: value <= other
# - is_type_of: isinstance(value, type)
# - is_in:  value in sequence
# - is_not_in: value not in sequence
# - identity: value is other
# - cont: contain value in
# - len_eq: len(value) == other
# - len_ne: len(value) != other
# - len_min: len(value) > other
# - len_max: len(value) < other
# - startswith: value.startswith(other)
# - endswith: value.endswith(other)
# Examples:
Validator("VERSION", eq="1.0.0"),
Validator("VERSION", ne="1.0.0"),
Validator("AGE", gt=18),
Validator("AGE", lt=18),
Validator("AGE", gte=18),
Validator("AGE", lte=18),
Validator("AGE", is_type_of=int),
Validator("AGE", is_in=[18, 19, 20]),
Validator("AGE", is_not_in=[18, 19, 20]),
Validator("THING", identity=thing),  # settings.THING is thing
Validator("THING", cont="hello"),  # "hello" in settings.THING
Validator("THING", len_eq=3),  # len(settings.THING) == 3
Validator("THING", len_ne=3),  # len(settings.THING) != 3
Validator("THING", len_min=3),  # len(settings.THING) > 3
Validator("THING", len_max=3),  # len(settings.THING) < 3
Validator("THING", startswith="hello"),  # settings.THING.startswith("hello")
Validator("THING", endswith="world"),  # settings.THING.endswith("world")