[misc] user-agent tagging

[bkmartinjr]

On Python/R API-initiated HTTP requests, set User-Agent tags:

• R/Py environment
• Package version
• Means to add additional "application name" tag (allows discrimination of CI, WMG builder, etc)

This should propagate through:

• HTTP requests to CDN, eg, fetch of release.json, mirrors.json
• TileDB requests to S3
• Other requests, e.g., H5AD fetch via API which is currently using fsspec

To do:

• Python API - TileDB requests. ([python] set user-agent #1193)
  - [ ] R API - TiledDB requests
• Python API - all other HTTP requests ([python] set user-agent #1193)
  - [ ] R API - all other HTTP requests
  - [ ] R unit tests set app flag
• Python unit tests set app flag ([python] set user-agent #1193)
• Python download_source_h5ad ([python] set user-agent #1193)
  - [ ] R download dataset file (equivalent of download_source_h5ad)
• Builder fetch of datasets manifest needs to set header
  - [ ] CI and other pipelines set custom "internal user" header

For hints on setting the TileDB header config: TileDB-Inc/TileDB#4400
For common User-Agent header examples, see here

[ivirshup]

We've merged #1193, which checks a few of the boxes here. I've edited the top comment to show this.

[pablo-gar]

Other non-Census usage is now tracked here

chanzuckerberg/single-cell-data-portal#7298

[pablo-gar]

R work is now tracked here
#1249

[pablo-gar]

@ivirshup if this has been completed

Builder fetch of datasets manifest needs to set header

Then we can close this ticket