Derives each axis's specification from what a build under cfg would
actually read, and matches it against lnk_vocabulary_read():
Arguments
- cfg
An
lnk_configobject fromlnk_config().- loaded
Named list from
lnk_load_overrides(). Loaded fromcfgwhenNULL.
Details
cut: the break sources in the order applied (
pipeline: break_order, or its default), the observation breaks (allortop_per_species, and a digest of the record rule that decides which records cut, resolved as if every species were present; only whenobservationsis a source) and the gradient classes streams are broken at (pipeline: gradient_classes, or bcfishpass's 15 / 20 / 25 / 30 %), and which of the definite barrier, barrier control and habitat classification files the bundle declares (their rows are data, which the build record carries).acc: a sha256 digest of the access rules, which decide where gradient barriers cut streams as well as what is accessible: the species the rules model, each species'
access_gradient_maxandobservation_*lift parameters, the species presence table, which fish records make code 2 (pipeline: observation_upstream, resolved), the record rule resolved for every group in the presence table (which recorded codes count as which species, with their date floors and year limits: the pooling tracker, its species groups, the region lookup andpipeline: observation_record_date, link#470), and the access rules held elsewhere: the gradient class each access label maps to (in code) and the presence groups (lnk_presence()'s groups as the bundle'spresence_groups.csvgives them, link#292). The digest is over the parsed values, so a change in quoting or column order is not a new rule set.grad: how a segment's gradient is read (
pipeline: gradient:):gradient=segment, the segment's own (the value when absent), or forsource: parentgradient=parent;parent_cut=<cut token>, the gradient of the segment ofminimal_config's network it lies in, where<cut token>is the tokenminimal_configresolves to. Aminimal_configwhose acc or flow spec differs fromcfg's is refused.flow:
pipeline: discharge_sourceandpipeline: discharge_fill.
A config whose specification matches no token, or only a retired one,
is refused: the new output gets a name (a row in
inst/extdata/vocabulary.csv) before anything is built.
See also
Other vocabulary:
lnk_vocabulary_read()
Examples
lnk_vocabulary_resolve(lnk_config("default"))
#> cut acc grad flow
#> "cut-full-top-2" "acc-v6" "grad-seg" "q-fwapg"
lnk_vocabulary_resolve(lnk_config("bcfishpass"))
#> cut acc grad
#> "cut-full-ssf-all-2" "acc-bcfp-5" "grad-seg"
#> flow
#> "q-fwapg"
