aws-lambda-decorators

A set of Python decorators to ease the development of AWS lambda functions.

Installation

The easiest way to use these AWS Lambda Decorators is to install them through Pip:

pip install aws-lambda-decorators

Logging

The Logging level of the decorators can be controlled by setting a LOG_LEVEL environment variable. In python:

os.environ["LOG_LEVEL"] = "INFO"

The default value is "INFO"

Package Contents

Decorators

The current list of AWS Lambda Python Decorators includes:

• extract: a decorator to extract and validate specific keys of a dictionary parameter passed to a AWS Lambda function.
• extract_from_event: a facade of extract to extract and validate keys from an AWS API Gateway lambda function event parameter.
• extract_from_context: a facade of extract to extract and validate keys from an AWS API Gateway lambda function context parameter.
• extract_from_ssm: a decorator to extract from AWS SSM the values of a set of parameter keys.
• validate: a decorator to validate a list of function parameters.
• log: a decorator to log the parameters passed to the lambda function and/or the response of the lambda function.
• handle_exceptions: a decorator to handle any type of declared exception generated by the lambda function.
• response_body_as_json: a decorator to transform a response dictionary body to a json string.
• handle_all_exceptions: a decorator to handle all exceptions thrown by the lambda function.
• cors: a decorator to add cors headers to a lambda function.

Validators

Currently, the package offers 5 validators:

• Mandatory: Checks if a parameter has a value.
• RegexValidator: Checks a parameter against a regular expression.
• SchemaValidator: Checks if an object adheres to the schema. Uses schema library.
• Minimum: Checks if an optional numerical value is greater than a minimum value.
• Maximum: Checks if an optional numerical value is less than a maximum value.
• MinLength: Checks if an optional string value is longer than a minimum length.
• MaxLength: Checks if an optional string value is shorter than a maximum length.
• Type: Checks if an optional object value is of a given python type.
• EnumValidator: Checks if an optional object value is in a list of valid values.
• NonEmpty: Checks if an optional object value is not an empty value.

Decoders

The package offers functions to decode from JSON and JWT.

• decode_json: decodes/converts a json string to a python dictionary
• decode_jwt: decodes/converts a JWT string to a python dictionary

Examples

extract

This decorator extracts and validates values from dictionary parameters passed to a Lambda Function.

• The decorator takes a list of Parameter objects.
• Each Parameter object requires a non-empty path to the parameter in the dictionary, and the name of the dictionary (func_param_name)
• The parameter value is extracted and added as a kwarg to the lambda handler (or any other decorated function/method).
• You can add the parameter to the handler signature, or access it in the handler through kwargs.
• The name of the extracted parameter is defaulted to the last element of the path name, but can be changed by passing a (valid pythonic variable name) var_name
• You can define a default value for the parameter in the Parameter or in the lambda handler itself.
• A 400 exception is raised when the parameter cannot be extracted or when it does not validate.
• A variable path (e.g. '/headers/Authorization[jwt]/sub') can be annotated to specify a decoding. In the example, Authorization might contain a JWT, which needs to be decoded before accessing the "sub" element.

Example:

@extract(parameters=[
    Parameter(path='/parent/my_param', func_param_name='a_dictionary'),  # extracts a non mandatory my_param from a_dictionary
    Parameter(path='/parent/missing_non_mandatory', func_param_name='a_dictionary', default='I am missing'),  # extracts a non mandatory missing_non_mandatory from a_dictionary
    Parameter(path='/parent/missing_mandatory', func_param_name='a_dictionary'),  # does not fail as the parameter is not validated as mandatory
    Parameter(path='/parent/child/id', validators=[Mandatory], var_name='user_id', func_param_name='another_dictionary')  # extracts a mandatory id as "user_id" from another_dictionary
])
def extract_example(a_dictionary, another_dictionary, my_param='aDefaultValue', missing_non_mandatory='I am missing', missing_mandatory=None, user_id=None):
    """
        Given these two dictionaries:

        a_dictionary = { 
            'parent': { 
                'my_param': 'Hello!' 
            }, 
            'other': 'other value' 
        }

        another_dictionary = { 
            'parent': { 
                'child': { 
                    'id': '123' 
                } 
            } 
        }

        you can now access the extracted parameters directly: 
    """
    return my_param, missing_non_mandatory, missing_mandatory, user_id

Or you can use kwargs instead of specific parameter names:

Example:

@extract(parameters=[
    Parameter(path='/parent/my_param', func_param_name='a_dictionary')  # extracts a non mandatory my_param from a_dictionary
])
def extract_to_kwargs_example(a_dictionary, **kwargs):
    """
        a_dictionary = { 
            'parent': { 
                'my_param': 'Hello!' 
            }, 
            'other': 'other value' 
        }
    """
    return kwargs['my_param']  # returns 'Hello!'

A missing mandatory parameter, or a parameter that fails validation, will raise an exception:

Example:

@extract(parameters=[
    Parameter(path='/parent/mandatory_param', func_param_name='a_dictionary', validators=[Mandatory])  # extracts a mandatory mandatory_param from a_dictionary
])
def extract_mandatory_param_example(a_dictionary, mandatory_param=None):
    return 'Here!'  # this part will never be reached, if the mandatory_param is missing

response = extract_mandatory_param_example({'parent': {'my_param': 'Hello!'}, 'other': 'other value'} )

print(response)  # prints { 'statusCode': 400, 'body': '{"message": [{"mandatory_param": ["Missing mandatory value"]}]}' } and logs a more detailed error

You can add custom error messages to all validators, and incorporate to those error messages the validated value and the validation condition:

Example:

@extract(parameters=[
    Parameter(path='/parent/an_int', func_param_name='a_dictionary', validators=[Minimum(100, 'Bad value {value}: should be at least {condition}')])  # extracts a mandatory mandatory_param from a_dictionary
])
def extract_minimum_param_with_custom_error_example(a_dictionary, mandatory_param=None):
    return 'Here!'  # this part will never be reached, if the an_int param is less than 100

response = extract_minimum_param_with_custom_error_example({'parent': {'an_int': 10}})

print(response)  # prints { 'statusCode': 400, 'body': '{"message": [{"an_int": ["Bad value 10: should be at least 100"]}]}' } and logs a more detailed error

You can group the validation errors together (instead of exiting on first error).

Example:

@extract(parameters=[
    Parameter(path='/parent/mandatory_param', func_param_name='a_dictionary', validators=[Mandatory]),  # extracts two mandatory parameters from a_dictionary
    Parameter(path='/parent/another_mandatory_param', func_param_name='a_dictionary', validators=[Mandatory]),
    Parameter(path='/parent/an_int', func_param_name='a_dictionary', validators=[Maximum(10), Minimum(5)])
], group_errors=True)  # groups both errors together
def extract_multiple_param_example(a_dictionary, mandatory_param=None, another_mandatory_param=None, an_int=0):
    return 'Here!'  # this part will never be reached, if the mandatory_param is missing

response = extract_multiple_param_example({'parent': {'my_param': 'Hello!', 'an_int': 20}, 'other': 'other value'})

print(response)  # prints {'statusCode': 400, 'body': '{"message": [{"mandatory_param": ["Missing mandatory value"]}, {"another_mandatory_param": ["Missing mandatory value"]}, {"an_int": ["\'20\' is greater than maximum value \'10\'"]}]}'}

You can decode any part of the parameter path from json or any other existing annotation.

Example:

@extract(parameters=[
    Parameter(path='/parent[json]/my_param', func_param_name='a_dictionary')  # extracts a non mandatory my_param from a_dictionary
])
def extract_from_json_example(a_dictionary, my_param=None):
    """
        a_dictionary = { 
            'parent': '{"my_param": "Hello!" }', 
            'other': 'other value' 
        }
    """
    return my_param  # returns 'Hello!'

You can also use an integer annotation to access an specific list element by index.

Example:

@extract(parameters=[
    Parameter(path='/parent[1]/my_param', func_param_name='a_dictionary')  # extracts a non mandatory my_param from a_dictionary
])
def extract_from_list_example(a_dictionary, my_param=None):
    """
        a_dictionary = { 
            'parent': [
                {'my_param': 'Hello!'},
                {'my_param': 'Bye!'}
            ]
        }
    """
    return my_param  # returns 'Bye!'

You can extract all parameters into a dictionary

Example:

@extract(parameters=[
    Parameter(path='/params/my_param_1', func_param_name='a_dictionary'),  # extracts a non mandatory my_param_1 from a_dictionary
    Parameter(path='/params/my_param_2', func_param_name='a_dictionary')  # extracts a non mandatory my_param_2 from a_dictionary
])
def extract_dictionary_example(a_dictionary, **kwargs):
    """
        a_dictionary = { 
            'params': {
                'my_param_1': 'Hello!',
                'my_param_2': 'Bye!'
            }
        }
    """
    return kwargs  # returns {'my_param_1': 'Hello!', 'my_param_2': 'Bye!'}

extract_from_event

This decorator is just a facade to the extract method to be used in AWS Api Gateway Lambdas. It automatically extracts from the event lambda parameter.

Example:

@extract_from_event(parameters=[
    Parameter(path='/body[json]/my_param', validators=[Mandatory]),  # extracts a mandatory my_param from the json body of the event
    Parameter(path='/headers/Authorization[jwt]/sub', validators=[Mandatory], var_name='user_id')  # extract the mandatory sub value as user_id from the authorization JWT
])
def extract_from_event_example(event, context, my_param=None, user_id=None):
    """
        event = { 
            'body': '{"my_param": "Hello!"}', 
            'headers': { 
                'Authorization': 'eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiIxMjM0NTY3ODkwIiwibmFtZSI6IkpvaG4gRG9lIiwiaWF0IjoxNTE2MjM5MDIyfQ.SflKxwRJSMeKKF2QT4fwpMeJf36POk6yJV_adQssw5c' 
            } 
        }
    """
    return my_param, user_id  # returns ('Hello!', '1234567890')

extract_from_context

This decorator is just a facade to the extract method to be used in AWS Api Gateway Lambdas. It automatically extracts from the context lambda parameter.

Example:

@extract_from_context(parameters=[
    Parameter(path='/parent/my_param', validators=[Mandatory])  # extracts a mandatory my_param from the parent element in context
])
def extract_from_context_example(event, context, my_param=None):
    """
        context = {
            'parent': {
                'my_param': 'Hello!'
            }
        }
    """    
    return my_param  # returns 'Hello!'

extract_from_ssm

This decorator extracts a parameter from AWS SSM and passes the parameter down to your function as a kwarg.

• The decorator takes a list of SSMParameter objects.
• Each SSMParameter object requires the name of the SSM parameter (ssm_name)
• If no var_name is passed in, the extracted value is passed to the function with the ssm_name name

Example:

@extract_from_ssm(ssm_parameters=[
    SSMParameter(ssm_name='one_key'),  # extracts the value of one_key from SSM as a kwarg named "one_key"
    SSMParameter(ssm_name='another_key', var_name="another")  # extracts another_key as a kwarg named "another"
])
def extract_from_ssm_example(your_func_params, one_key=None, another=None):
    return your_func_params, one_key, another

validate

This decorator validates a list of non dictionary parameters from your lambda function.

• The decorator takes a list of ValidatedParameter objects.
• Each parameter object needs the name of the lambda function parameter that it is going to be validated, and the list of rules to validate.
• A 400 exception is raised when the parameter does not validate.

Example:

@validate(parameters=[
    ValidatedParameter(func_param_name='a_param', validators=[Mandatory]),  # validates a_param as mandatory
    ValidatedParameter(func_param_name='another_param', validators=[Mandatory, RegexValidator(r'\d+')])  # validates another_param as mandatory and containing only digits
    ValidatedParameter(func_param_name='param_with_schema', validators=[SchemaValidator(Schema({'a': Or(str, dict)}))])  # validates param_with_schema as an object with specified schema
])
def validate_example(a_param, another_param, param_with_schema):
    return a_param, another_param, param_with_schema  # returns 'Hello!', '123456', {'a': {'b': 'c'}}

validate_example('Hello!', '123456', {'a': {'b': 'c'}})

Given the same function validate_example, a 400 exception is returned if at least one parameter does not validate (as per the extract decorator, you can group errors with the group_errors flag):

validate_example('Hello!', 'ABCD')  # returns a 400 status code and an error message

log

This decorator allows for logging the function arguments and/or the response.

Example:

@log(parameters=True, response=True)
def log_example(parameters): 
    return 'Done!'

log_example('Hello!')  # logs 'Hello!' and 'Done!'

handle_exceptions

This decorator handles a list of exceptions, returning a 400 response containing the specified friendly message to the caller.

• The decorator takes a list of ExceptionHandler objects.
• Each ExceptionHandler requires the type of exception to check, and an optional friendly message to return to the caller.

Example:

@handle_exceptions(handlers=[
    ExceptionHandler(ClientError, "Your message when a client error happens.")
])
def handle_exceptions_example():
    dynamodb = boto3.resource('dynamodb')
    table = dynamodb.Table('non_existing_table')
    table.query(KeyConditionExpression=Key('user_id').eq(user_id))
    # ...

handle_exceptions_example()  # returns {'body': '{"message": "Your message when a client error happens."}', 'statusCode': 400}

handle_all_exceptions

This decorator handles all exceptions thrown by a lambda, returning a 400 response and the exception's message.

Example:

@handle_all_exceptions()
def handle_exceptions_example():
    test_list = [1, 2, 3]
    invalid_value = test_list[5]
    # ...    

handle_all_exceptions_example()  # returns {'body': '{"message": "list index out of range"}, 'statusCode': 400}

response_body_as_json

This decorator ensures that, if the response contains a body, the body is dumped as json.

• Returns a 500 error if the response body cannot be dumped as json.

Example:

@response_body_as_json
def response_body_as_json_example():
    return {'statusCode': 400, 'body': {'param': 'hello!'}}

response_body_as_json_example()  # returns { 'statusCode': 400, 'body': "{'param': 'hello!'}" }

cors

This decorator adds your defined CORS headers to the decorated function response.

• Returns a 500 error if one or more of the CORS headers have an invalid type

Example:

@cors(allow_origin='*', allow_methods='POST', allow_headers='Content-Type', max_age=86400)
def cors_example():
    return {'statusCode': 200}

cors_example()  # returns {'statusCode': 200, 'headers': {'access-control-allow-origin': '*', 'access-control-allow-methods': 'POST', 'access-control-allow-headers': 'Content-Type', 'access-control-max-age': 86400}}

Writing your own validators

You can create your own validators by inheriting from the Validator class.

Fix length validator example:

class FixLength(Validator):
    ERROR_MESSAGE = "'{value}' length should be '{condition}'"

    def __init__(self, fix_length: int, error_message=None):
        super().__init__(error_message=error_message, condition=fix_length)

    def validate(self, value=None):
        if value is None:
            return True

        return len(str(value)) == self._condition

Documentation

You can get the docstring help by running:

>>> from aws_lambda_decorators.decorators import extract
>>> help(extract)

Links

• PyPi
• Test PyPi
• Github