gelman

-module(gelman@transform).
-compile([no_auto_import, nowarn_unused_vars, nowarn_unused_function, nowarn_nomatch, inline]).
-define(FILEPATH, "src/gelman/transform.gleam").
-export([winsorize/3]).

-if(?OTP_RELEASE >= 27).
-define(MODULEDOC(Str), -moduledoc(Str)).
-define(DOC(Str), -doc(Str)).
-else.
-define(MODULEDOC(Str), -compile([])).
-define(DOC(Str), -compile([])).
-endif.

?MODULEDOC(
    " This module defines functions that perform\n"
    " statistical transformations over a dataset.\n"
    "\n"
    " A statistical transformation is a function that modifies\n"
    " the dataset based on some statistically desirable qualities.\n"
    "\n"
    " A statistical function could preserve the size of the dataset\n"
    " or it could reduce the size of the dataset by excluding some\n"
    " elements.\n"
    " Unlike the functions in summarize or discretize, these do not\n"
    " error out if the dataset has no elements.\n"
).

-file("src/gelman/transform.gleam", 22).
-spec winsorize(list(float()), float(), float()) -> {ok, list(float())} |
    {error, gelman@error:stats_error()}.
winsorize(Dataset, Left, Right) ->
    case {Dataset,
        (Left > +0.0) andalso (Left < 1.0),
        (Right > +0.0) andalso (Right < 1.0)} of
        {_, false, _} ->
            {error,
                {invalid_parameter,
                    <<"Winsorize bounds between 0.0 and 0.1"/utf8>>}};

        {_, _, false} ->
            {error,
                {invalid_parameter,
                    <<"Winsorize bounds between 0.0 and 0.1"/utf8>>}};

        {[], _, _} ->
            {error, empty_dataset};

        {[X], _, _} ->
            {ok, [X]};

        {[X@1, Y], _, _} ->
            {ok, [X@1, Y]};

        {_, _, _} ->
            _pipe = Dataset,
            _pipe@1 = gelman@internal@nonempty@transform:winsorize(
                _pipe,
                Left,
                Right
            ),
            {ok, _pipe@1}
    end.