This commit is contained in:
bronku 2025-11-13 14:46:52 +01:00
parent 758f733c84
commit 69bc9646e4
11 changed files with 6 additions and 830 deletions

129
buffer.c
View file

@ -1,129 +0,0 @@
#include "buffer.h"
#include "record.h"
#include "status_codes.h"
#include <stdbool.h>
#include <string.h>
buffer* create_buffer(int capacity)
{
buffer* out = malloc(sizeof(buffer));
out->length = 0;
out->capacity = capacity;
out->location = malloc(sizeof(record) * capacity);
out->original = true;
return out;
}
void destroy_buffer(buffer* buff)
{
if (buff->original) {
free(buff->location);
}
free(buff);
}
buffer** split_buffer(buffer* original, int pieces)
{
int size = original->capacity;
if (size <= 0) {
return NULL;
}
buffer** out = malloc(sizeof(buffer) * pieces);
for (int i = 0; i < pieces; i++) {
out[i]->length = 0;
out[i]->capacity = size;
out[i]->original = 0;
out[i]->location = original->location;
out[i]->location += i * size * sizeof(record);
}
return out;
}
int read_buffer(buffer* buff, FILE* in)
{
int status;
record* current = buff->location;
buff->length = 0;
for (int i = 0; i < buff->capacity; i++) {
status = read_record(in, current);
if (status != SUCCESS) {
return status;
}
buff->length++;
current++;
}
return SUCCESS;
}
int write_buffer(buffer* buff, FILE* out)
{
int status;
record* current = buff->location;
for (int i = 0; i < buff->length; i++) {
status = print_record(out, current);
if (status < 0) {
return status;
}
fprintf(out, "\n");
current++;
}
return SUCCESS;
}
int write_buffer_debug(buffer* buff, FILE* out)
{
int status;
record* current = buff->location;
for (int i = 0; i < buff->length; i++) {
fprintf(out, "%d\t", g(current));
status = print_record(out, current);
if (status < 0) {
return status;
}
fprintf(out, "\n");
current++;
}
return SUCCESS;
}
int partition(record* arr, int start, int end)
{
int i = start;
record* pivot = arr;
pivot += end;
for (int j = start; j < end; j++) {
record* current = arr;
current += j;
if (compare_records(current, pivot) > 0) {
continue;
}
record* i_rec = arr;
i_rec += i;
record tmp = *i_rec;
*i_rec = *current;
*current = tmp;
i++;
}
record* i_rec = arr;
i_rec += i;
record tmp = *i_rec;
*i_rec = *pivot;
*pivot = tmp;
return i;
}
void quicksort(record* arr, int start, int end, int depth)
{
if (start >= end || start < 0) {
return;
}
int p = partition(arr, start, end);
quicksort(arr, start, p - 1, depth + 1);
quicksort(arr, p + 1, end, depth + 1);
}
void sort_buffer(buffer* buff)
{
quicksort(buff->location, 0, buff->length - 1, 0);
}

View file

@ -1,19 +0,0 @@
#pragma once
#include "record.h"
#include <stdbool.h>
typedef struct {
record* location;
int length;
int capacity;
bool original;
} buffer;
buffer* create_buffer(int capacity);
void destroy_buffer(buffer* buff);
buffer** split_buffer(buffer* original, int pieces);
int read_buffer(buffer* buff, FILE* in);
int write_buffer(buffer* buff, FILE* in);
int write_buffer_debug(buffer* buff, FILE* in);
void sort_buffer(buffer* buff);

View file

@ -1,52 +0,0 @@
#include "config.h"
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
int process_args(int argc, char** argv, Configuration* opts)
{
opts->input_file = DEFAULT_IN;
opts->output_file = DEFAULT_OUT;
opts->directory = DEFAULT_DIR;
opts->generate_data = false;
opts->N = DEFAULT_NN;
opts->n = DEFAULT_N;
opts->b = DEFAULT_B;
int opt;
bool input_set = false;
while ((opt = getopt(argc, argv, "i:o:d:gN:n:b:")) != -1) {
switch (opt) {
case 'i':
opts->input_file = optarg;
input_set = true;
break;
case 'o':
opts->output_file = optarg;
break;
case 'd':
opts->directory = optarg;
break;
case 'g':
opts->generate_data = true;
break;
case 'N':
opts->N = atoi(optarg);
break;
case 'n':
opts->n = atoi(optarg);
break;
case 'b':
opts->b = atoi(optarg);
break;
case '?':
printf("wrong argument\n");
return ERROR;
}
}
if (opts->generate_data && input_set) {
printf("incompatible arguments\n");
return ERROR;
}
return SUCCESS;
}

View file

@ -1,23 +0,0 @@
#pragma once
#include "status_codes.h"
#include <stdbool.h>
#define DEFAULT_IN "1.in"
#define DEFAULT_OUT "1.in"
#define DEFAULT_DIR "tmp"
#define DEFAULT_NN 100000
#define DEFAULT_N 101
#define DEFAULT_B 10
typedef struct
{
const char* input_file;
const char* output_file;
const char* directory;
bool generate_data;
int N; // number of records in a file
int b; // blocking factor
int n; // number of buffers
} Configuration;
int process_args(int argc, char** argv, Configuration* opts);

78
heap.c
View file

@ -1,78 +0,0 @@
#include "heap.h"
#include "status_codes.h"
#include <stdlib.h>
heap* create_heap(int size, int (*comparison)(const void* a, const void* b))
{
heap* out = malloc(sizeof(heap));
out->data = malloc(sizeof(void*) * size);
out->length = 0;
out->capacity = size;
out->compare = comparison;
return out;
}
void destroy_heap(heap* h)
{
free(h->data);
free(h);
}
void swap(void** a, void** b)
{
void* tmp = *a;
*a = *b;
*b = tmp;
}
int heap_push(heap* h, void* in)
{
int current = h->length;
void** d = h->data;
h->length++;
d[current] = in;
if (current == 0) {
return SUCCESS;
}
int parent = (current - 1) / 2;
while (h->compare(d[current], d[parent]) == -1) {
swap(&d[current], &d[parent]);
current = parent;
if (current == 0) {
break;
}
parent = (current - 1) / 2;
}
return SUCCESS;
}
int heap_pop(heap* h, void** out)
{
void** d = h->data;
*out = d[0];
h->length--;
if (h->length == 0) {
return SUCCESS;
}
swap(&d[0], &d[h->length]);
int current = 0;
int left = current * 2 + 1;
int right = current * 2 + 2;
while (true) {
int smallest = current;
if (left < h->length && h->compare(d[smallest], d[left]) == 1) {
smallest = left;
}
if (right < h->length && h->compare(d[smallest], d[right]) == 1) {
smallest = right;
}
if (smallest == current) {
break;
}
swap(&d[smallest], &d[current]);
current = smallest;
left = current * 2 + 1;
right = current * 2 + 2;
}
return SUCCESS;
}

19
heap.h
View file

@ -1,19 +0,0 @@
// doesn't ceck for user erros, so no veryfying that the void* supplied are valid
// unsafe for general use, but just fnuctional enough for the purposes of this project
// does not maintain ownership of its elements, they are managed by the caller
#pragma once
#include "stdbool.h"
typedef struct {
void** data;
int length;
int capacity;
int (*compare)(const void* a, const void* b);
} heap;
heap* create_heap(int size, int (*comparison)(const void* a, const void* b));
void destroy_heap(heap* h);
int heap_pop(heap* h, void** out);
int heap_push(heap* h, void* in);

View file

@ -1,302 +0,0 @@
// the tests are ai generated, with prompt based on the heap.h file
#include "heap.h"
#include <assert.h>
#include <stdio.h>
#include <stdlib.h>
// Comparator for integers
int compare_ints(const void* a, const void* b)
{
int ia = *(const int*)a;
int ib = *(const int*)b;
return (ia > ib) - (ia < ib);
}
// Comparator for doubles
int compare_doubles(const void* a, const void* b)
{
double da = *(const double*)a;
double db = *(const double*)b;
if (da < db)
return -1;
if (da > db)
return 1;
return 0;
}
// Test structure
typedef struct {
int priority;
char name[32];
} Task;
// Comparator for tasks (by priority)
int compare_tasks(const void* a, const void* b)
{
const Task* ta = (const Task*)a;
const Task* tb = (const Task*)b;
return (ta->priority > tb->priority) - (ta->priority < tb->priority);
}
void test_basic_push_pop()
{
printf("Test: Basic push and pop... ");
heap* h = create_heap(5, compare_ints);
int values[] = { 5, 3, 7, 1, 9 };
// Push all values
for (int i = 0; i < 5; i++) {
heap_push(h, &values[i]);
}
// Pop and verify they come out in sorted order
void* out;
int expected[] = { 1, 3, 5, 7, 9 };
for (int i = 0; i < 5; i++) {
heap_pop(h, &out);
assert(*(int*)out == expected[i]);
}
destroy_heap(h);
printf("PASSED\n");
}
void test_single_element()
{
printf("Test: Single element... ");
heap* h = create_heap(1, compare_ints);
int val = 42;
heap_push(h, &val);
void* out;
heap_pop(h, &out);
assert(*(int*)out == 42);
destroy_heap(h);
printf("PASSED\n");
}
void test_already_sorted()
{
printf("Test: Already sorted input... ");
heap* h = create_heap(5, compare_ints);
int values[] = { 1, 2, 3, 4, 5 };
for (int i = 0; i < 5; i++) {
heap_push(h, &values[i]);
}
void* out;
for (int i = 0; i < 5; i++) {
heap_pop(h, &out);
assert(*(int*)out == i + 1);
}
destroy_heap(h);
printf("PASSED\n");
}
void test_reverse_sorted()
{
printf("Test: Reverse sorted input... ");
heap* h = create_heap(5, compare_ints);
int values[] = { 5, 4, 3, 2, 1 };
for (int i = 0; i < 5; i++) {
heap_push(h, &values[i]);
}
void* out;
for (int i = 0; i < 5; i++) {
heap_pop(h, &out);
assert(*(int*)out == i + 1);
}
destroy_heap(h);
printf("PASSED\n");
}
void test_duplicates()
{
printf("Test: Duplicate values... ");
heap* h = create_heap(7, compare_ints);
int values[] = { 3, 1, 4, 1, 5, 9, 2 };
for (int i = 0; i < 7; i++) {
heap_push(h, &values[i]);
}
void* out;
int expected[] = { 1, 1, 2, 3, 4, 5, 9 };
for (int i = 0; i < 7; i++) {
heap_pop(h, &out);
assert(*(int*)out == expected[i]);
}
destroy_heap(h);
printf("PASSED\n");
}
void test_with_doubles()
{
printf("Test: Double precision values... ");
heap* h = create_heap(5, compare_doubles);
double values[] = { 3.14, 1.41, 2.71, 0.57, 1.61 };
for (int i = 0; i < 5; i++) {
heap_push(h, &values[i]);
}
void* out;
double expected[] = { 0.57, 1.41, 1.61, 2.71, 3.14 };
for (int i = 0; i < 5; i++) {
heap_pop(h, &out);
assert(*(double*)out == expected[i]);
}
destroy_heap(h);
printf("PASSED\n");
}
void test_with_structs()
{
printf("Test: Custom struct (tasks by priority)... ");
heap* h = create_heap(4, compare_tasks);
Task tasks[] = {
{ 3, "Low priority" },
{ 1, "High priority" },
{ 2, "Medium priority" },
{ 1, "Also high priority" }
};
for (int i = 0; i < 4; i++) {
heap_push(h, &tasks[i]);
}
void* out;
int expected_priorities[] = { 1, 1, 2, 3 };
for (int i = 0; i < 4; i++) {
heap_pop(h, &out);
Task* t = (Task*)out;
assert(t->priority == expected_priorities[i]);
}
destroy_heap(h);
printf("PASSED\n");
}
void test_interleaved_operations()
{
printf("Test: Interleaved push/pop... ");
heap* h = create_heap(10, compare_ints);
int values[] = { 5, 3, 7, 1, 9, 2, 8, 4, 6, 0 };
// Push 5, pop 1, push 3 more, pop 2, etc.
heap_push(h, &values[0]); // 5
heap_push(h, &values[1]); // 3
heap_push(h, &values[2]); // 7
void* out;
heap_pop(h, &out);
assert(*(int*)out == 3);
heap_push(h, &values[3]); // 1
heap_push(h, &values[4]); // 9
heap_push(h, &values[5]); // 2
heap_pop(h, &out);
assert(*(int*)out == 1);
heap_pop(h, &out);
assert(*(int*)out == 2);
heap_push(h, &values[6]); // 8
heap_push(h, &values[7]); // 4
heap_pop(h, &out);
assert(*(int*)out == 4);
destroy_heap(h);
printf("PASSED\n");
}
void test_large_heap()
{
printf("Test: Large heap (1000 elements)... ");
heap* h = create_heap(1000, compare_ints);
int* values = malloc(1000 * sizeof(int));
// Insert values in random-ish order
for (int i = 0; i < 1000; i++) {
values[i] = (i * 7 + 13) % 1000;
heap_push(h, &values[i]);
}
// Pop all and verify sorted
void* out;
int prev = -1;
for (int i = 0; i < 1000; i++) {
heap_pop(h, &out);
int curr = *(int*)out;
assert(curr >= prev);
prev = curr;
}
free(values);
destroy_heap(h);
printf("PASSED\n");
}
void test_all_same_values()
{
printf("Test: All same values... ");
heap* h = create_heap(5, compare_ints);
int val = 42;
int values[5];
for (int i = 0; i < 5; i++) {
values[i] = val;
heap_push(h, &values[i]);
}
void* out;
for (int i = 0; i < 5; i++) {
heap_pop(h, &out);
assert(*(int*)out == 42);
}
destroy_heap(h);
printf("PASSED\n");
}
int main()
{
printf("Running Min Heap Tests\n");
printf("======================\n\n");
test_basic_push_pop();
test_single_element();
test_already_sorted();
test_reverse_sorted();
test_duplicates();
test_with_doubles();
test_with_structs();
test_interleaved_operations();
test_large_heap();
test_all_same_values();
printf("\n======================\n");
printf("All tests PASSED! ✓\n");
return 0;
}

120
main.c
View file

@ -1,120 +0,0 @@
#include "buffer.h"
#include "config.h"
#include "record.h"
#include "status_codes.h"
#include <sys/stat.h>
#include <sys/types.h>
#include <time.h>
#include <unistd.h>
int generate_file(int N, const char* filename)
{
srand(time(0));
FILE* f = fopen(filename, "w");
record rec;
for (int i = 0; i < N; i++) {
random_record(&rec);
print_record(f, &rec);
fprintf(f, "\n");
}
return SUCCESS;
}
int sort_file(Configuration* opts)
{
buffer* buff = create_buffer(opts->b * opts->n);
FILE* in = fopen(opts->input_file, "r");
int runs;
// <stage 1>
for (runs = 0; true; runs++) {
// read buffer
int status = read_buffer(buff, in);
if (status != SUCCESS && status != EOF) {
fclose(in);
destroy_buffer(buff);
}
// sort buffer
sort_buffer(buff);
// write run
char* filename = malloc(256);
sprintf(filename, "%s/%d", opts->directory, runs);
FILE* tmp = fopen(filename, "w");
write_buffer(buff, tmp);
// write_buffer_debug(buff, stdout);
fclose(tmp);
free(filename);
if (status != SUCCESS) {
break;
}
}
printf("created runs: %d\n", runs);
// <stage 2>
while (runs > 1) {
// split buffers
// read buffers
for (int i = 0; i < opts->b - 1; i++) {
}
// b-1 runs are turned into one, so the total is reduced by b-2
runs -= (opts->b - 2);
}
fclose(in);
destroy_buffer(buff);
return SUCCESS;
}
int main(int argc, char** argv)
{
Configuration opts;
if (argc <= 1) {
// #todo: stdin input
fprintf(stderr, "please specify input args\n");
return ERROR;
}
int result = process_args(argc, argv, &opts);
if (result != SUCCESS) {
return result;
}
if (opts.generate_data) {
return generate_file(opts.N, opts.output_file);
}
sort_file(&opts);
}
// int sort_block(tape* t, buffer* buff)
// {
// static int counter = 0;
// int status = read_buffer(buff, t);
// if (status != SUCCESS && status != EOF) {
// fprintf(stderr, "Error reading file: %d\n", status);
// return status;
// }
// sort_buffer(buff);
// printf("sorted the buffer %d, the results:\n", counter);
// print_buffer(buff);
// counter++;
// return status;
// }
// int sort_file(options* opts)
// {
// tape a = { 0, 0, opts->input };
// buffer buff;
// buff.capacity = opts->b * opts->n;
// buff.location = malloc(sizeof(record) * buff.capacity);
// int status = 0;
// while (status == SUCCESS) {
// status = sort_block(&a, &buff);
// }
// if (status != EOF) {
// free(buff.location);
// fclose(a.file);
// return status;
// }
// free(buff.location);
// fclose(a.file);
// return SUCCESS;
// }

View file

@ -1,19 +1,7 @@
CFLAGS = -g
CFLAGS = -g -lc++
build/main: build/main.o build/config.o build/buffer.o build/heap.o
clang $(CFLAGS) build/main.o build/config.o build/buffer.o build/heap.o -o build/main
build/main.o: main.c build
clang $(CFLAGS) -c main.c -o build/main.o
build/config.o: config.c build
clang $(CFLAGS) -c config.c -o build/config.o
build/buffer.o: buffer.c build
clang $(CFLAGS) -c buffer.c -o build/buffer.o
build/heap.o: heap.c build
clang $(CFLAGS) -c heap.c -o build/heap.o
build/main: main.cpp build
clang $(CFLAGS) main.cpp -o build/main
build:
mkdir -p build
@ -27,22 +15,15 @@ build/run/tmp: build/run
clean:
rm -rf build
generate: build/run/1.in
build/run/1.in: build/run build/main
build/main -g -o build/run/1.in
run: build/main build/run/1.in build/run/tmp
build/main -i build/run/1.in -d build/run/tmp | less
build/heap_test.o: heap_test.c build
clang $(CFLAGS) -c heap_test.c -o build/heap_test.o
build/test: build/heap_test.o build/heap.o
clang $(CFLAGS) build/heap_test.o build/heap.o -o build/test
test: build/test
build/test
debug: build/main tests/1.in
lldb -- build/main
.PHONY: clean run debug test
.PHONY: clean run debug test generate

View file

@ -1,57 +0,0 @@
#pragma once
#include <stdio.h>
#include <stdlib.h>
typedef struct {
int a[5];
int x;
} record;
static inline int g(const record* x)
{
int out = 0;
int x_n = 1;
for (int i = 0; i < 5; i++) {
out += x->a[i] * x_n;
x_n *= x->x;
}
return out;
}
static inline int compare_records(const void* a, const void* b)
{
const record* ra = (const record*)a;
const record* rb = (const record*)b;
int ga = g(ra);
int gb = g(rb);
int result = (ga > gb) - (ga < gb);
return result;
}
static inline int print_record(FILE* stream, const record* out)
{
return fprintf(stream, "%d %d %d %d %d %d",
out->a[0], out->a[1], out->a[2], out->a[3], out->a[4], out->x);
}
static inline int read_record(FILE* stream, record* out)
{
record tmp;
int result = fscanf(stream, "%d %d %d %d %d %d", &tmp.a[0], &tmp.a[1], &tmp.a[2], &tmp.a[3], &tmp.a[4], &tmp.x);
if (result == 6) {
*out = tmp;
return 0;
}
if (result == EOF) {
return EOF;
}
return -2;
}
static inline void random_record(record* out)
{
for (int i = 0; i < 5; i++) {
out->a[i] = rand();
}
out->x = rand();
}

View file

@ -1,6 +0,0 @@
#pragma once
#define END_PROGRAM -1
#define SUCCESS 0
#define ERROR -2
#define NO_SPACE -3