Support bitmap PDFs

[mrlerner]

This PDF is primarily bitmaps and we're not getting the text out of the bitmaps so the summary is based only on a little bit of text contained in the PDF.

https://seattle.legistar.com/View.ashx?M=F&ID=11900375&GUID=ABE1CCF3-84B8-4645-B4C7-1078C12B7B72

[davepeck]

extract_pipeline_v2 will probably want to use the GCP Cloud Vision API in PDF mode in order to get maximal content from summarized PDFs.

[davepeck]

I've decided this is out of band for the "wrap up" work I want to do this week. Yes, it'd be essential in the next rev, but it doesn't push the needle enough to matter (for now).

[mrlerner]

Good call. Agreed.