Skip to contents

Retrieve trait data and priors for one PFT from BETYdb

Source: R/get_trait_data_pft.R
get_trait_data_pft.Rd

Core computation extracted from get.trait.data.pft. Queries the database for trait observations and prior distributions, returning them as R objects with no file I/O of any kind.

Usage

get_trait_data_pft(pft_name, modeltype, dbcon, trait_names, constants = list())

Arguments

pft_name

character. PFT name as stored in BETYdb.

modeltype

character or NULL. Disambiguates PFTs that share a name across model types (e.g. "SIPNET", "ED2").

dbcon

database connection from db.open.

trait_names

character vector of trait names to retrieve.

constants

named list from pft$constants in the settings. Traits named here are excluded from the returned priors because their values are fixed rather than sampled by the meta-analysis.

Value

Named list with three elements:

trait_data

Named list of data frames, one per trait that has observations. Column structure matches what meta_analysis_standalone expects. Traits with no observations are omitted from the list.

prior_distns

Data frame with columns distn, parama, paramb, n; rows named by trait. Traits listed in constants are excluded.

pft_info

List with name, pft_id, pft_type, pft_members, pft_member_filename, and posteriorid. pft_members is the data frame of species or cultivar IDs used during the query. pft_member_filename is "species.csv" or "cultivars.csv" depending on PFT type. posteriorid is always NULL — the wrapper sets it after registering outputs in BETYdb.

Details

The wrapper get.trait.data.pft handles directory creation, caching, CSV output, and BETYdb registration via dbfile.insert. This function handles only the query. Choosing between them is the provenance opt-in: calling the wrapper saves artifacts to disk; calling this function never does.

This follows the pattern established by meta_analysis_standalone for the meta-analysis step and get_parameter_samples for parameter sampling — each is a computation core that can be tested in isolation without a filesystem or a settings object.

See also

get.trait.data.pft for the backward-compatible wrapper that handles provenance and caching. meta_analysis_standalone (in PEcAn.MA) for the analogous function in the meta-analysis step. get_parameter_samples for the analogous function in the parameter sampling step.

Author

David LeBauer, Shawn Serbin, Alexey Shiklomanov, Om Kapale

Examples

if (FALSE) { # \dontrun{
dbcon <- PEcAn.DB::db.open(list(
  host = "localhost", user = "bety",
  password = "bety",  dbname = "bety"
))
result <- get_trait_data_pft(
  pft_name    = "temperate.deciduous",
  modeltype   = "SIPNET",
  dbcon       = dbcon,
  trait_names = c("SLA", "Vcmax", "leaf_respiration_rate_m2")
)
str(result$trait_data)
str(result$prior_distns)
PEcAn.DB::db.close(dbcon)
} # }