Skip to contents

Run the whole pipeline

Source: R/pipeline.R
pf_run.Rd

pf_prepare() -> pf_reference() -> pf_link() -> pf_report().

Usage

pf_run(
  grants,
  bmf = NULL,
  pattern = "grants.*[.]csv$",
  dedup = c("fuzzy", "exact", "none"),
  ref_cache = "data-dev/BMF-REFERENCE-BUNDLE.rds",
  dir = "data-dev/batches",
  batch_size = 25000L,
  dyads = NULL,
  report = TRUE,
  verbose = TRUE,
  ref_source = c("auto", "matchdb", "local"),
  tier = getOption("pfmatch.matchdb_tier", "active-observed"),
  ...
)

Arguments

grants, pattern, dedup

Passed to pf_prepare().

bmf, ref_cache

Passed to pf_reference(). With no bmf, the reference is the matchdb tier, falling back to the bundle at ref_cache.

dir, batch_size, dyads

Passed to pf_link(). dyads may be a table or a path to a cached one.

report

Print the outcome tables when done.

verbose

Print progress.

ref_source, tier

Passed to pf_reference() as source and tier.

...

Further arguments to pf_link().

Value

A list of class pf_run with grants, recipients, status, and timing. The reference bundle is deliberately not returned — it is ~1.3 GB and is on disk (at ref_cache, or in matchdb's cache).