The manifest is a small CSV under the data root that records the provenance of every source and derived asset — what it is, where it came from, when it was fetched, its checksum and row count. It is what makes a run reproducible: documentation can point at exact vintages, and a colleague can verify they hold the same bytes.
Usage
np_manifest(root = np_data_root())
np_manifest_add(
asset,
tier = c("raw", "normalized", "results"),
source = NA,
url = NA,
path = NULL,
md5 = NA,
row_count = NA,
note = NA,
root = np_data_root()
)Arguments
- root
Data root. Defaults to
np_data_root().- asset
Short asset name (also the file's basename), e.g.
"bmf-2026-01".- tier
One of
"raw","normalized","results".- source
Human label of the origin (e.g.
"NCCS BMF catalog").- url
Download / origin URL.
- path
Optional path to the file, used to auto-fill
md5androw_count.- md5, row_count
Optional explicit checksum / row count (override
path).- note
Free-text note.
Value
np_manifest() returns the manifest data frame; np_manifest_add()
returns it invisibly after writing.
Details
np_manifest_add() appends (or replaces, by asset name) one row; the md5
is computed with tools::md5sum() when path is given and md5 is not.