python-asyncio: update retry factors for actual exponential retries (#19337)

* python-asyncio: update retry factors for actual exponential retries

As per the `aiohttp-retry` library's code[^1], the timeout is

```python
timeout = self._start_timeout * (self._factor ** attempt)
```

This means the previous setting with "start_timeout=0.0" would have
always just retried right away (0 timeout) regardless of the "factor" value,
and also, "factor=0.0" would never have increased the timeout, rather it
would have resulted in a 0 timeout regardless of the value of "start_timeout".

This double-zeroing effectively rendered exponential backoff to nothing (rather than
"retries" number of retries in quick succession.

The update is a quick fix to set the same default as in `aiohttp-retry`. In
the future this should likely be configurable (through extra Configuration settings perhaps?),
as not all APIs are created equal, but this works as a quick fix for making retries more effective.

[^1]: https://github.com/inyutin/aiohttp_retry/blob/ba2169891f5b32a5c59e48ca185dd8e68e44ded7/aiohttp_retry/retry_options.py#L38-L65

* updated example

Author: imrehg

modules/openapi-generator/src/main/resources/python/asyncio/rest.mustache

@@ -79,8 +79,8 @@ class RESTClientObject:
                 client_session=self.pool_manager,
                 retry_options=aiohttp_retry.ExponentialRetry(
                     attempts=retries,
-                    factor=0.0,
-                    start_timeout=0.0,
+                    factor=2.0,
+                    start_timeout=0.1,
                     max_timeout=120.0
                 )
             )

samples/openapi3/client/petstore/python-aiohttp/petstore_api/rest.py

@@ -89,8 +89,8 @@ def __init__(self, configuration) -> None:
                 client_session=self.pool_manager,
                 retry_options=aiohttp_retry.ExponentialRetry(
                     attempts=retries,
-                    factor=0.0,
-                    start_timeout=0.0,
+                    factor=2.0,
+                    start_timeout=0.1,
                     max_timeout=120.0
                 )
             )