// Binomial PDF
//
// binomial_pdf - binomial probability density
//
// P(X=k) = C(n, k) * p^k * (1-p)^(n-k)
//
// Inputs:    1 (k values)
// Outputs:   1 (PDF value)
// Constants: 1 (n trials, p probability)

#include <math.h>
#include "ts.h"

typedef struct {
    uint_t n_trials;
    fp_t p;
} ts_binomial_pdf_params_t;

ts_parameter_t const ts_binomial_pdf_p = { .n_inputs = 1, .n_outputs = 1, .n_constants = 1, .params_size = sizeof(ts_binomial_pdf_params_t) };

uint_t ts_binomial_pdf_offset( ts_constant_t const *c )
{
	( void )c;
	return ( 0 );
}

int ts_binomial_pdf(
	fp_t * const *y, uint_t y_n,
	fp_t const * const *x, uint_t x_n,
	uint_t n, ts_constant_t const *c
)
{
	TS_PROLOGUE( ts_binomial_pdf_p );
	if ( !c ) return -1;
	uint_t n_trials = ((ts_binomial_pdf_params_t *)c)->n_trials;
	fp_t p = ((ts_binomial_pdf_params_t *)c)->p;
	if (p < 0.0 || p > 1.0 || isnan(p) || isinf(p)) return -1;
	for( uint_t i = 0; i < n; ++i )
	{
		uint_t k = (uint_t)x[0][i];
		if( k > n_trials )
		{
			y[0][i] = 0.0;
		}
		else if( p == 0.0 )
		{
			y[0][i] = k == 0 ? 1.0 : 0.0;
		}
		else if( p == 1.0 )
		{
			y[0][i] = k == n_trials ? 1.0 : 0.0;
		}
		else
		{
			fp_t log_comb = lgamma( n_trials + 1.0 ) - lgamma( k + 1.0 ) - lgamma( n_trials - k + 1.0 );
			y[0][i] = exp( log_comb + k * log( p ) + ( n_trials - k ) * log( 1.0 - p ) );
		}
	}
	return 0;
}