Comma-joined `signals`/`geo_type` with a space after the comma fails with "not an available signal" — the split doesn't trim whitespace

[docxology]

Summary

While using the documented comma-joined style for v5 signals, I hit a confusing failure: epidata_snapshot("nhsn", "pct_visits, urgent_care", "state") errors with "not an available signal" because the comma-split at R/endpoints.R:1576-1577 (and :1688-1689 in epidata_archive()) uses strsplit(..., fixed = TRUE) without trimws(), so the second signal arrives as " urgent_care" with a leading space. The comma-joined input style is the documented one (R/endpoints.R:1461-1462), so this is a small but real papercut; a trimws() before unique() fixes it.

Evidence (dev tip 87b10e1)

# R/endpoints.R:1576-1577 (epidata_snapshot; same at :1688-1689 in epidata_archive)
signals <- unique(unlist(strsplit(signals, ",", fixed = TRUE)))
geo_type <- unique(unlist(strsplit(geo_type, ",", fixed = TRUE)))

strsplit("pct_visits, urgent_care", ",", fixed = TRUE)
# [[1]] "pct_visits" " urgent_care"   <- leading space preserved

Note: the in-flight PR #380 (send signals comma-joined in one request per geo_type) reworks this fan-out but keeps the same untrimmed split, so the papercut survives it.

Maintainer-runnable repro

epidata_snapshot("nhsn", "pct_visits, urgent_care", "state")
# error: " urgent_care" is not an available signal for source nhsn...

Suggested fix

signals <- unique(trimws(unlist(strsplit(signals, ",", fixed = TRUE))))
geo_type <- unique(trimws(unlist(strsplit(geo_type, ",", fixed = TRUE))))

(applied in both epidata_snapshot() and epidata_archive()).

I checked for prior reports (open/closed issues about signal names or the comma style — #380's PR is the in-flight rework) and found none covering the trimming.