azure/functions/extension/func_extension_base.py (55 lines of code) (raw):
# Copyright (c) Microsoft Corporation. All rights reserved.
# Licensed under the MIT License.
import abc
import os
import typing
from logging import Logger
from .extension_meta import ExtensionMeta
from .extension_scope import ExtensionScope
from .function_extension_exception import FunctionExtensionException
from .._abc import Context
class FuncExtensionBase(metaclass=ExtensionMeta):
"""An abstract class defines the life-cycle hooks which to be implemented
by customer's extension.
Everytime when a new extension is initialized in customer function scripts,
the ExtensionManager._func_exts field records the extension to this
specific function name.
"""
_scope = ExtensionScope.FUNCTION
@abc.abstractmethod
def __init__(self, file_path: str):
"""Constructor for extension. This needs to be implemented and ensure
super().__init__(file_path) is called.
The initializer serializes the extension to a tree. This speeds
up the worker lookup and reduce the overhead on each invocation.
_func_exts[<trigger_name>].<hook_name>.(ext_name, ext_impl)
(e.g. _func_exts['HttpTrigger'].pre_invocation.ext_impl)
Parameters
----------
file_path: str
The name of trigger the extension attaches to (e.g. __file__).
"""
script_root = os.getenv('AzureWebJobsScriptRoot')
if script_root is None:
raise FunctionExtensionException(
'AzureWebJobsScriptRoot environment variable is not defined. '
'Please ensure the extension is running in Azure Functions.'
)
# Split will always return ('') in if no folder exist in the path
relpath_to_project_root = os.path.relpath(
os.path.normpath(file_path),
os.path.normpath(script_root)
)
trigger_name = (relpath_to_project_root.split(os.sep) or [''])[0]
if not trigger_name or trigger_name.startswith(('.', '..')):
raise FunctionExtensionException(
'Failed to parse trigger name from filename. '
'Function extension should bind to a trigger script, '
'not share folder. Please ensure extension is create inside a'
'trigger while __file__ is passed into the argument. '
'The trigger name is resolved from os.path.relpath(file_path,'
'project_root).'
)
# This is used in ExtensionMeta._register_function_extension
self._trigger_name = trigger_name
# DO NOT decorate this with @abc.abstractmethod
# since implementation by subclass is not mandatory
def post_function_load(self,
function_name: str,
function_directory: str,
*args, **kwargs) -> None:
"""This hook will be called right after a customer's function loaded.
In this stage, the customer's logger is not fully initialized, so it
is not provided. Please use print() to emit message if necessary.
Parameters
----------
function_name: str
The name of customer's function (e.g. HttpTrigger)
function_directory: str
The path to customer's function directory
(e.g. /home/site/wwwroot/HttpTrigger)
"""
pass
# DO NOT decorate this with @abc.abstractmethod
# since implementation by subclass is not mandatory
def pre_invocation(
self,
logger: Logger,
context: Context,
func_args: typing.Optional[typing.Dict[str, object]] = None,
*args,
**kwargs) -> None:
"""This hook will be called right before customer's function
is being executed.
Parameters
----------
logger: logging.Logger
A logger provided by Python worker. Extension developer should
use this logger to emit telemetry to Azure Functions customers.
context: azure.functions.Context
This will include the function_name, function_directory and an
invocation_id of this specific invocation.
func_args: typing.Dict[str, object]
Arguments that are passed into the Azure Functions. The name of
each parameter is defined in function.json. Extension developers
may also want to do isinstance() check if you want to apply
operations to specific trigger types or input binding types.
"""
pass
# DO NOT decorate this with @abc.abstractmethod
# since implementation by subclass is not mandatory
def post_invocation(
self,
logger: Logger,
context: Context,
func_args: typing.Optional[typing.Dict[str, object]] = None,
func_ret: typing.Optional[object] = None,
*args,
**kwargs) -> None:
"""This hook will be called right after a customer's function
is executed.
Parameters
----------
logger: logging.Logger
A logger provided by Python worker. Extension developer should
use this logger to emit telemetry to Azure Functions customers.
context: azure.functions.Context
This will include the function_name, function_directory and an
invocation_id of this specific invocation.
func_args: typing.Dict[str, object]
Arguments that are passed into the Azure Functions. The name of
each parameter is defined in function.json. Extension developers
may also want to do isinstance() check if you want to apply
operations to specific trigger types or input binding types.
func_ret: typing.Optional[object]
Return value from Azure Functions. This is usually the value
defined in function.json $return section. Extension developers
may also want to do isinstance() check if you want to apply
operations to specific types or input binding types.
"""
pass