Issue with multipage PDFs on s3 without extension

Hello, first of all thanks for the awesome package.

I am currently having an issue trying to run textractor on my PDFs that are stored in s3.
The issue stems from the fact that all my files (for security, and other reasons, which I think are pretty common practice at larger enterprises) are stored as UUIDs instead of their actual filename so when `call_textract` is called, it goes through the entire process without actually hitting any of the if statements and just returns an empty dict.

Is there any way that maybe this use case could be supported?