Validation should fail on unrecognized file type

[konklone]

Right now, validation will fail if the file_type wasn't detected (the URL has no file extension) but will not fail if the detected file_type is unknown.

Since we only have text processors for HTML and PDF files, the file_type should be either auto-detected, or set by a scraper, to html or pdf. If it's not, it should choke and force the scraper to pick one -- and if we come across a report format that isn't HTML or PDF, then it's time to extend the system to process text from that format.

[konklone]

This can build on @divergentdave's work in 1fa8f5d, but that only patches the problem -- the file_type field should be html, for a report whose URL ends in .aspx, and the saved file should be report.html.

[audiodude]

👍 I agree this is the more correct way to do it.