voice tests

[jamesmcneill]

Currently chatpickle supports postText. Could this be expanded to postContent as well, this would allow posting voice to lex.

[arsena]

Good idea.
https://docs.aws.amazon.com/lex/latest/dg/API_runtime_PostContent.html

Two ways that this could become part of chatpickle.

1. Create a Lex client that uses postContent for all conversational steps like User: This would get converted to audio before posted to lex
2. Add a step definition like Given its a voice channel, and that might change how we test what was returned, if a string is returned, great chatpickle already asserts the expected string, but if its audio that comes back, that would be a fun area to explore how we would verify an audio file is accurate.

[arsena]

@Taylor-Lapointe fyi

[reddyharsha]

I forked the repo and am working on this issue. Pull request will follow soon.

[reddyharsha]

Here's the pull request. #46.

[seckinarsoy]

I'm working on testing this locally.