Add frs_params() for loading habitat model parameter sets

[NewGraphEnvironment]

Problem

bcfishpass species parameters (gradient thresholds, channel width ranges, observation count thresholds) are scattered across SQL scripts. Need a single way to load parameter sets from any source for the frs_break() / frs_classify() pipeline.

Proposed Solution

frs_params() — load parameter sets from pg table, custom table, or local CSV. Returns a list ready for purrr::walk() over the pipeline.

# Default: bcfishpass parameter tables already in pg
params <- frs_params(conn)

# Custom: your own table with experimental thresholds
params <- frs_params(conn, table = "working.my_thresholds")

# Offline: CSV for rapid iteration without touching the database
params <- frs_params(csv = "my_experimental_thresholds.csv")

Design decisions

• Returns a list of per-species param sets, each containing gradient thresholds, channel width ranges, species codes, observation thresholds
• CSV enables scenario testing without DB access: edit locally, run both, compare
• Schema for the default bcfishpass tables is configurable (same options() pattern as other fresh functions)
• No opinion on what the params mean — just loads and normalizes them. The consuming functions (frs_break(), frs_classify()) interpret the values

Depends on

• None (can be built first)

[NewGraphEnvironment]

Implemented in R/frs_params.R with tests. Loads from DB table or CSV, returns named list keyed by species_code with ranges sub-list for frs_classify().