So you can do something like this instead of needing to instantiate a DocumentTranslationInput:
poller = client.begin_translation(source, target, "fr", storage_type="file")
Need to add support to pass storage_type, prefix, suffix, source_storage_source, target_storage_source, source_language_code, category_id, and glossaries as keyword arguments into the single translation overloaded version of begin_translation.
The kwargs above account for both source and target options. They match the name of the properties found on DocumentTranslationInput except in the case of source_storage_source and target_storage_source where they are prefixed with "source" and "target" to differentiate the types of storage sources. These currently only support "AzureBlob" so might not be worth even exposing?