From d77a813fd26da1aaff96e47cec44b56d74ef8685 Mon Sep 17 00:00:00 2001
From: Giovanni La Mura <giovanni.lamura@inaf.it>
Date: Wed, 5 Mar 2025 17:33:59 +0100
Subject: [PATCH] Update README instructions for reference data

---
 ref_data/README.md | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)

diff --git a/ref_data/README.md b/ref_data/README.md
index a2e4d676..d58e3fe9 100644
--- a/ref_data/README.md
+++ b/ref_data/README.md
@@ -1,18 +1,12 @@
 # Folder instructions
 
-This directory collects the input and output data files to test the code.
+This directory contains reference data for model configuration, including, in particular, the dielectric constant files.
 
 ## Instructions
 
-The execution of the original code can be controlled through a set of configuration files that define the characteristics of the problem and affect the type of output. In order to allow for consistency tests between the results of the legacy _FORTRAN_ implementation and the new _C++_ version, a set or pre-computed test cases is provided. The test cases are arranged in folders named after the code branch they are intended to test. Namely:
+Dielectric constant files must be formatted as comma separated values, following the scheme provided by the available example files. `NPtm_code` can extract the necessary dielectric constants from these files, either by interpolating them on an arbitrary set of wavelengths, or just computing the scattering process exactly on the grid of available constants.
 
-- the `cluster` folder contains input and output data for test cases using clusters of spheres;
-- the `sphere` folder contains input and output data for test cases based on a single sphere;
-- the `trapping` folder contains input and output data for test cases applying to the `trapping` code.
-
-The legacy _FORTRAN_ code is distributed together with the `np_tmcode` software package, so that the user can verrify the correspondence between results obtained by different code builds. However, some of the _FORTRAN_ based calculations are time consuming. For this reason, the set of pre-computed output files should serve as a first choice for reference.
-
-In general, input data files are identified by names that start with the `D` character, while output files have names that start with the `O` character. The _C++_ implementation of the code prepends a `c_` prefix to all of its output files (while still using the same input ones).
+If the calculation uses the exact grid matching, then all the dielectric constants *must* be provided on the same set of wavelengths. This constraint does not apply to the interpolation mode, because the dielectric constants will be internally interpolated on the same grid of wavelengths. In this second case, however, it is user's responsibility that the requested wavelength range is covered by all the necessary dielectric constant data.
 
 # License
 
-- 
GitLab