Functions and Interpolators

The top-level Functions block defines reusable helper functions — typically interpolators built from a table — once, so you can call them by name inside any expression without duplicating logic.

Example

Functions:
  - name: FASER2                          # call it as FASER2(...) in any expr
    source:
      type: csv
      path: ./Fns/FASER2_620.csv
      x: "np.log(y)"                      # expression for the table's x values
      y: "x"                              # expression for the table's y values
      sort_by: "y"
      drop_duplicates: "y"
    method: interp1d
    options:
      kind: linear                        # linear | nearest | cubic | ...
      bounds: clamp                       # behavior outside the table range

Once defined, use it like any built-in function:

coordinates:
  y: {expr: "FASER2(np.log10(mass))"}

Fields

Field	Purpose
name	The callable name exposed to expressions
source.type	Table format, e.g. csv
source.path	Path to the table (resolved like any DataSet path)
source.x / source.y	Expressions selecting the table's input/output columns
source.sort_by	Column/expression to sort the table by before building the interpolator
source.drop_duplicates	Drop duplicate rows on this key
method	Interpolator kind, e.g. interp1d
options	Interpolator options (kind, bounds, …)

How it is used

At load time these specs are compiled into callables and injected into the expression runtime. After that, filter, add_column, and any coordinates expression can call them directly — the same way Jarvis-Operas-registered operators become available by name.