umccrise: support presigned URL parsing

[pdiakumis]

Mainly adding functionality for parsing umccrise outputs (HRD/Sigs/QCsum) via presigned URLs.

• read_tsvgz uses a gz connection prior to attempting to read the TSVs, in the same vein as described in Handle presigned URLs of gzipped JSONs #74.
• glims_read is a convenience function for parsing the Google LIMS via {googlesheets}.
• The umccrise metadata available from Athena (workflows table) is enough to grab LibraryID and SampleID for tumor and normal, except for those runs prior to 2023-04-07 (tumor SampleID is not available for those). We can instead use the LibraryID from the Google LIMS to map those and fill in the gaps.
• Setting up an Rmd reporter for multiple umccrise results. Ingestion is actually pretty quick, a bit less than 1sec per file (449 successful umccrise runs x 5 files = 2,245 files from 2022-01-24 until 2023-09-03, excluding MultiQC results).

[pdiakumis]

Merging, this looks good for now, added a funky HRD grouped plot.
Will try to grab older bcbio results too this week.