Provide generic DockerContainer with init attributes

First I want to thank you for this very nice and convenient library.

Background

I am using the library for smoke tests for microservices. In these tests I spin up a container and make sure I can e.g., fetch an endpoint and see that the container is built and starts correctly.

The library provides all convenient methods, except for a way to set e.g., ports and env variables in the context manager. To solve this I have subclassed testcontainers.core.container.DockerContainer into a class as below:

from typing import Any

import testcontainers.core.container

class ContainerContext(testcontainers.core.container.DockerContainer):
    def __init__(
        self, image: str, ports: list[int], env_vars: dict[str, str], **kwargs: Any
    ) -> None:
        super().__init__(image, **kwargs)

        self.with_exposed_ports(*ports)
        for env, value in env_vars.items():
            self.with_env(env, value)

and use it as

import requests

def test_smoke():
    with container_context(
        "my_service_image",
        ports=[8000],
        env_vars={"SOME_ENV_SETTINGS": "some_value", "SOME_ENV_SETTINGS2": "other_value"},
    ) as con:
        r = requests.get(
            f"http://{con.get_container_host_ip()}:{con.get_exposed_port(8000)}"
        )
        assert r

Proposal

• Include a generic container class where all name, ports, env, volumes, … can be set as init variables.

I saw #114 that removed the previous GenericContainer but I don't see a motivation why it was done and why it could not be extended.

Review testcontainers.core.container.DockerContainer and the previous GenericContainer removal discussed in #114. Compare the requested init attributes—name, ports, env, and volumes—with the existing container API, and consider the proposal complete when a generic container can configure these values at initialization without subclassing.