__init__.py: improve error handling and enhance async processing

[thor4]

I experienced many errors when trying to run the indexer due to async issues. It was difficult at first to discern what was driving the failures. So, I modified the async approach and improved error handling.

Key changes:

1. Improved error handling:

   • Used specific exceptions like KeyError and RequestException for more granular error handling.
   • Logged errors using logging.error() for better visibility and troubleshooting.
   • Raised exceptions using raise_for_status() for 4xx and 5xx status codes from the Text Analytics API.
   • Handled job failure by raising a RuntimeError if the job status is 'failed'.

2. Enhanced asynchronous processing:

   • Implemented a retry mechanism with exponential backoff to handle cases where the Text Analytics job takes longer to complete.
   • Added a maximum number of retries (max_retries) and a retry delay (retry_delay) that doubles with each retry attempt.
   • Raised a TimeoutError if the job doesn't complete within the maximum number of retries.

[ivorb]

There are some additional changes we want to include here, we appreciate these updates and improvements, but since this is a zip I don't want to merge as is. Thanks for your contributions! Update with some of these changes will be coming shortly