remove functionality

- Sample from arbitrary cdf: unused complexity
- Sample confidence interval directly from cdf: makes sampling
function/array function boundary unclear.
This commit is contained in:
NunoSempere 2024-02-01 20:26:37 +01:00
parent b4c50996cd
commit 0975512412

View File

@ -194,6 +194,17 @@ double array_get_median(double xs[], int n){
return quickselect(median_k, xs, n);
}
/* array print: potentially useful for debugging */
void array_print(double xs[], int n)
{
printf("[");
for (int i = 0; i < n - 1; i++) {
printf("%f, ", xs[i]);
}
printf("%f", xs[n - 1]);
printf("]\n");
}
void array_print_stats(double xs[], int n){
ci ci_90 = array_get_ci((ci) { .low = 0.05, .high = 0.95 }, xs, n);
ci ci_80 = array_get_ci((ci) { .low = 0.1, .high = 0.9 }, xs, n);
@ -391,23 +402,6 @@ void array_print_90_ci_histogram(double* xs, int n_samples, int n_bins){
}
// Replicate some of the above functions over samplers
// However, in the future I'll delete this
// There should be a clear boundary between working with samplers and working with an array of samples
ci sampler_get_ci(ci interval, double (*sampler)(uint64_t*), int n, uint64_t* seed)
{
UNUSED(seed); // don't want to use it right now, but want to preserve ability to do so (e.g., remove parallelism from internals). Also nicer for consistency.
double* xs = malloc((size_t)n * sizeof(double));
sampler_parallel(sampler, xs, 16, n);
ci result = array_get_ci(interval, xs, n);
free(xs);
return result;
}
ci sampler_get_90_ci(double (*sampler)(uint64_t*), int n, uint64_t* seed)
{
return sampler_get_ci((ci) { .low = 0.05, .high = 0.95 }, sampler, n, seed);
}
/* Algebra manipulations */
#define NORMAL90CONFIDENCE 1.6448536269514727
@ -458,216 +452,3 @@ ci convert_lognormal_params_to_ci(lognormal_params y)
ci result = { .low = exp(loglow), .high = exp(loghigh) };
return result;
}
/* Scaffolding to handle errors */
// We will sample from an arbitrary cdf
// and that operation might fail
// so we build some scaffolding here
#define MAX_ERROR_LENGTH 500
#define EXIT_ON_ERROR 0
#define PROCESS_ERROR(error_msg) process_error(error_msg, EXIT_ON_ERROR, __FILE__, __LINE__)
typedef struct box_t {
int empty;
double content;
char* error_msg;
} box;
box process_error(const char* error_msg, int should_exit, char* file, int line)
{
if (should_exit) {
printf("%s, @, in %s (%d)", error_msg, file, line);
exit(1);
} else {
char error_msg[MAX_ERROR_LENGTH];
snprintf(error_msg, MAX_ERROR_LENGTH, "@, in %s (%d)", file, line); // NOLINT: We are being carefull here by considering MAX_ERROR_LENGTH explicitly.
box error = { .empty = 1, .error_msg = error_msg };
return error;
}
}
/* Invert an arbitrary cdf at a point */
// Version #1:
// - input: (cdf: double => double, p)
// - output: Box(number|error)
box inverse_cdf_double(double cdf(double), double p)
{
// given a cdf: [-Inf, Inf] => [0,1]
// returns a box with either
// x such that cdf(x) = p
// or an error
// if EXIT_ON_ERROR is set to 1, it exits instead of providing an error
double low = -1.0;
double high = 1.0;
// 1. Make sure that cdf(low) < p < cdf(high)
int interval_found = 0;
while ((!interval_found) && (low > -DBL_MAX / 4) && (high < DBL_MAX / 4)) {
// for floats, use FLT_MAX instead
// Note that this approach is overkill
// but it's also the *correct* thing to do.
int low_condition = (cdf(low) < p);
int high_condition = (p < cdf(high));
if (low_condition && high_condition) {
interval_found = 1;
} else if (!low_condition) {
low = low * 2;
} else if (!high_condition) {
high = high * 2;
}
}
if (!interval_found) {
return PROCESS_ERROR("Interval containing the target value not found, in function inverse_cdf");
} else {
int convergence_condition = 0;
int count = 0;
while (!convergence_condition && (count < (INT_MAX / 2))) {
double mid = (high + low) / 2;
int mid_not_new = (mid == low) || (mid == high);
// double width = high - low;
// if ((width < 1e-8) || mid_not_new){
if (mid_not_new) {
convergence_condition = 1;
} else {
double mid_sign = cdf(mid) - p;
if (mid_sign < 0) {
low = mid;
} else if (mid_sign > 0) {
high = mid;
} else if (mid_sign == 0) {
low = mid;
high = mid;
}
}
}
if (convergence_condition) {
box result = { .empty = 0, .content = low };
return result;
} else {
return PROCESS_ERROR("Search process did not converge, in function inverse_cdf");
}
}
}
// Version #2:
// - input: (cdf: double => Box(number|error), p)
// - output: Box(number|error)
box inverse_cdf_box(box cdf_box(double), double p)
{
// given a cdf: [-Inf, Inf] => Box([0,1])
// returns a box with either
// x such that cdf(x) = p
// or an error
// if EXIT_ON_ERROR is set to 1, it exits instead of providing an error
double low = -1.0;
double high = 1.0;
// 1. Make sure that cdf(low) < p < cdf(high)
int interval_found = 0;
while ((!interval_found) && (low > -DBL_MAX / 4) && (high < DBL_MAX / 4)) {
// for floats, use FLT_MAX instead
// Note that this approach is overkill
// but it's also the *correct* thing to do.
box cdf_low = cdf_box(low);
if (cdf_low.empty) {
return PROCESS_ERROR(cdf_low.error_msg);
}
box cdf_high = cdf_box(high);
if (cdf_high.empty) {
return PROCESS_ERROR(cdf_low.error_msg);
}
int low_condition = (cdf_low.content < p);
int high_condition = (p < cdf_high.content);
if (low_condition && high_condition) {
interval_found = 1;
} else if (!low_condition) {
low = low * 2;
} else if (!high_condition) {
high = high * 2;
}
}
if (!interval_found) {
return PROCESS_ERROR("Interval containing the target value not found, in function inverse_cdf");
} else {
int convergence_condition = 0;
int count = 0;
while (!convergence_condition && (count < (INT_MAX / 2))) {
double mid = (high + low) / 2;
int mid_not_new = (mid == low) || (mid == high);
// double width = high - low;
if (mid_not_new) {
// if ((width < 1e-8) || mid_not_new){
convergence_condition = 1;
} else {
box cdf_mid = cdf_box(mid);
if (cdf_mid.empty) {
return PROCESS_ERROR(cdf_mid.error_msg);
}
double mid_sign = cdf_mid.content - p;
if (mid_sign < 0) {
low = mid;
} else if (mid_sign > 0) {
high = mid;
} else if (mid_sign == 0) {
low = mid;
high = mid;
}
}
}
if (convergence_condition) {
box result = { .empty = 0, .content = low };
return result;
} else {
return PROCESS_ERROR("Search process did not converge, in function inverse_cdf");
}
}
}
/* Sample from an arbitrary cdf */
// Before: invert an arbitrary cdf at a point
// Now: from an arbitrary cdf, get a sample
box sampler_cdf_box(box cdf(double), uint64_t* seed)
{
double p = sample_unit_uniform(seed);
box result = inverse_cdf_box(cdf, p);
return result;
}
box sampler_cdf_double(double cdf(double), uint64_t* seed)
{
double p = sample_unit_uniform(seed);
box result = inverse_cdf_double(cdf, p);
return result;
}
double sampler_cdf_danger(box cdf(double), uint64_t* seed)
{
double p = sample_unit_uniform(seed);
box result = inverse_cdf_box(cdf, p);
if (result.empty) {
exit(1);
} else {
return result.content;
}
}
/* array print: potentially useful for debugging */
void array_print(double xs[], int n)
{
printf("[");
for (int i = 0; i < n - 1; i++) {
printf("%f, ", xs[i]);
}
printf("%f", xs[n - 1]);
printf("]\n");
}