[Feature Request] Support for dependency injection (pinject)

[grizzlybearer]

In order to use Temporal in our systems, we need the ability to inject dependencies such as grpc connections, db connections, kafka connections, auth data into temporal workflows and activities, using pinject, that are scoped at different levels (global, customer specific).

Here's an example:
Say you have a python class to fetch customer scoped data from redis:

from pinject import new_object_graph
import redis

class CustomerDataAccess:
    def __init__(self, customer_partition):
       # this is simplified. IRL there would be nested dependencies injected
        self.customer_partition = customer_partition
        self.redis_client = redis.Redis()

    def get_data(self):
        # Access data from the specified customer partition in Redis
        data = self.redis_client.get(self.customer_partition)
        return data

and then you have a workflow and an activity that needs this dependency:

# Define the Workflow interface
class CustomerWorkflow:
    @workflow_method(task_queue="customer-data")
    def get_customer_data(self) -> str:
        pass

# Implement the Workflow
class CustomerWorkflowImpl(CustomerWorkflow):
    def __init__(self, customer_data_access: CustomerDataAccess):
        self.data_access = customer_data_access

    def get_customer_data(self) -> str:
        # Access data from a specific customer partition using the Activity
        data = self.data_access.get_data("customer1")
        # do business logic on this data
        return result

Describe the solution you'd like

Proposed solution: Add support for pinject, similar to how you folks added support for Spring Boot in Java: temporalio/sdk-java#745

Additional context

[cretz]

Closing as question/duplicate. Duplicate post at https://community.temporal.io/t/python-support-for-dependency-injection-pinject/7538 is a better place to discuss.