Run the full PTI calculation pipeline

Runs the seven-step PTI calculation pipeline (pivot -> weight -> standardise -> expand -> merge -> aggregate -> label -> structure) as a single callable, mirroring the reactive chain in `mod_calc_pti2_server()` but free of any Shiny dependency. Designed for testing and headless reproduction of the calculation a deployed app would perform for a given weight scheme.

Usage

run_pti_pipeline(
  weights_clean,
  inp_dta,
  shp_dta,
  indicators_list = get_indicators_list(inp_dta),
  na_rm = FALSE
)

Arguments

weights_clean

Named list of weight tibbles, one per weight scheme. Each element must be a tibble with character column `var_code` and numeric column `weight`. Compatible with the output of [get_rand_weights()] and [get_all_weights_combs()].

inp_dta

The metadata list as returned by [fct_template_reader()] – contains admin-level sheets (`admin<N>_*`), `metadata`, and `general`.

shp_dta

Named list of `sf` tibbles, one per admin level (e.g. `admin1_Oblast`, `admin2_Rayon`).

indicators_list

Indicator definitions as a tibble. Defaults to `get_indicators_list(inp_dta)`.

na_rm

Logical. Forwarded to `agg_pti_scores()` as `na_rm_pti2` – if `TRUE`, missing indicator values are ignored when row-summing standardised contributions to the PTI score.

Value

A named list with one element per admin level; each element is a list with `pti_data` (an `sf` tibble), `pti_codes` (named character), and `admin_level` (named character). This is the same structure that `mod_calc_pti2_server()` returns to its callers via reactive value.

Determinism

This function is deterministic given fixed inputs. It does not consult Golem options. Pass `na_rm` explicitly rather than relying on `golem::get_golem_options("na_rm_pti")` to keep tests hermetic.

See also

Other pti-pipeline: compile_pti_data(), generic_pti_glue(), label_generic_pti()

Examples

data(rwa_shp)
data(rwa_mtdt_full)

# The indicators_list default fires from inside the package namespace,
# so the example does not call get_indicators_list() (now internal).
weights <- get_rand_weights(rwa_mtdt_full$metadata)

result <- run_pti_pipeline(
  weights_clean = weights,
  inp_dta       = rwa_mtdt_full,
  shp_dta       = rwa_shp
)
names(result)
#> [1] "admin0" "admin1" "admin2" "admin9"
names(result[[1]])
#> [1] "pti_data"    "pti_codes"   "admin_level"