bluelib/object/hashmap.c

#include "hashmap.h"

#include <blue/core/misc.h>
#include <blue/core/status.h>
#include <blue/object/hashmap.h>
#include <blue/object/string.h>
#include <blue/object/type.h>
#include <stdbool.h>
#include <stdlib.h>

#define HASH_OFFSET_BASIS 0xcbf29ce484222325
#define HASH_PRIME        0x100000001b3

/* clang-format off */
static B_BTREE_DEFINE_SIMPLE_GET(
	struct b_hashmap_bucket,
	uint64_t,
	bk_node,
	bk_hash,
	get_bucket)
static B_BTREE_DEFINE_SIMPLE_INSERT(
	struct b_hashmap_bucket,
	bk_node,
	bk_hash,
	put_bucket)

static uint64_t hash_data(const void *p, size_t size)
{
	/* clang-format on */
	const unsigned char *s = p;
	uint64_t hash = HASH_OFFSET_BASIS;

	for (size_t i = 0; s[i]; i++) {
		hash *= HASH_PRIME;
		hash ^= s[i];
	}

	return hash;
}

static void hashmap_release(struct b_object *obj);

static struct b_object_type hashmap_type = {
	.t_name = "corelib::hashmap",
	.t_flags = B_OBJECT_FUNDAMENTAL,
	.t_id = B_OBJECT_TYPE_HASHMAP,
	.t_instance_size = sizeof(struct b_hashmap),
	.t_release = hashmap_release,
};

struct b_hashmap *b_hashmap_create(
	b_hashmap_key_destructor key_dtor, b_hashmap_value_destructor value_dtor)
{
	struct b_hashmap *hashmap
		= (struct b_hashmap *)b_object_type_instantiate(&hashmap_type);
	if (!hashmap) {
		return NULL;
	}

	return hashmap;
}

struct b_hashmap *b_hashmap_create_with_items(const b_hashmap_item *items)
{
	struct b_hashmap *hashmap = b_hashmap_create(NULL, NULL);
	if (!hashmap) {
		return NULL;
	}

	for (size_t i = 0; items[i].key.key_data && items[i].key.key_size; i++) {
		b_hashmap_put(hashmap, &items[i].key, &items[i].value);
	}

	return hashmap;
}

static struct b_hashmap_bucket *create_bucket(void)
{
	/* clang-format on */
	struct b_hashmap_bucket *bucket = malloc(sizeof *bucket);
	if (!bucket) {
		return NULL;
	}

	memset(bucket, 0x0, sizeof *bucket);
	return bucket;
}

static struct b_hashmap_bucket_item *create_bucket_item(void)
{
	struct b_hashmap_bucket_item *item = malloc(sizeof *item);
	if (!item) {
		return NULL;
	}

	memset(item, 0x0, sizeof *item);
	return item;
}

b_status b_hashmap_put(
	struct b_hashmap *hashmap, const b_hashmap_key *key,
	const b_hashmap_value *value)
{
	uint64_t hash = hash_data(key->key_data, key->key_size);
	struct b_hashmap_bucket *bucket = get_bucket(&hashmap->h_buckets, hash);

	if (!bucket) {
		bucket = create_bucket();
		if (!bucket) {
			return B_ERR_NO_MEMORY;
		}

		bucket->bk_hash = hash;
		put_bucket(&hashmap->h_buckets, bucket);
	}

	b_queue_iterator it;
	b_queue_foreach (&it, &bucket->bk_items) {
		struct b_hashmap_bucket_item *item = b_unbox(
			struct b_hashmap_bucket_item, it.entry, bi_entry);

		if (item->bi_key.key_size != key->key_size) {
			continue;
		}

		if (!memcmp(item->bi_key.key_data, key->key_data, key->key_size)) {
			return B_ERR_NAME_EXISTS;
		}
	}

	struct b_hashmap_bucket_item *item = create_bucket_item();
	if (!item) {
		return B_ERR_NO_MEMORY;
	}

	memcpy(&item->bi_key, key, sizeof *key);
	memcpy(&item->bi_value, value, sizeof *value);

	b_queue_push_back(&bucket->bk_items, &item->bi_entry);

	return B_SUCCESS;
}

const struct b_hashmap_value *b_hashmap_get(
	const struct b_hashmap *hashmap, const struct b_hashmap_key *key)
{
	uint64_t hash = hash_data(key->key_data, key->key_size);
	struct b_hashmap_bucket *bucket = get_bucket(&hashmap->h_buckets, hash);
	if (!bucket) {
		return NULL;
	}

	b_queue_iterator it;
	b_queue_foreach (&it, &bucket->bk_items) {
		struct b_hashmap_bucket_item *item = b_unbox(
			struct b_hashmap_bucket_item, it.entry, bi_entry);

		if (item->bi_key.key_size != key->key_size) {
			continue;
		}

		if (!memcmp(item->bi_key.key_data, key->key_data, key->key_size)) {
			return &item->bi_value;
		}
	}

	return NULL;
}

bool b_hashmap_has_key(const struct b_hashmap *hashmap, const b_hashmap_key *key)
{
	uint64_t hash = hash_data(key->key_data, key->key_size);
	struct b_hashmap_bucket *bucket = get_bucket(&hashmap->h_buckets, hash);
	if (!bucket) {
		return false;
	}

	b_queue_iterator it;
	b_queue_foreach (&it, &bucket->bk_items) {
		struct b_hashmap_bucket_item *item = b_unbox(
			struct b_hashmap_bucket_item, it.entry, bi_entry);

		if (item->bi_key.key_size != key->key_size) {
			continue;
		}

		if (!memcmp(item->bi_key.key_data, key->key_data, key->key_size)) {
			return true;
		}
	}

	return false;
}

size_t b_hashmap_get_size(const struct b_hashmap *hashmap)
{
	size_t count = 0;
	b_btree_iterator it1;
	b_queue_iterator it2;
	b_btree_foreach (&it1, &hashmap->h_buckets) {
		struct b_hashmap_bucket *bucket
			= b_unbox(struct b_hashmap_bucket, it1.node, bk_node);

		b_queue_foreach (&it2, &bucket->bk_items) {
			count++;
		}
	}

	return count;
}

bool b_hashmap_is_empty(const b_hashmap *hashmap)
{
	b_btree_node *first_node = b_btree_first(&hashmap->h_buckets);
	struct b_hashmap_bucket *first_bucket
		= b_unbox(struct b_hashmap_bucket, first_node, bk_node);
	if (!first_bucket) {
		return true;
	}

	b_queue_entry *first_entry = b_queue_first(&first_bucket->bk_items);
	struct b_hashmap_bucket_item *first_item
		= b_unbox(struct b_hashmap_bucket_item, first_entry, bi_entry);
	if (!first_item) {
		return true;
	}

	return false;
}

static b_status delete_item(
	struct b_hashmap *hashmap, struct b_hashmap_bucket *bucket,
	struct b_hashmap_bucket_item *item)
{
	b_queue_delete(&bucket->bk_items, &item->bi_entry);

	if (hashmap->h_key_dtor) {
		hashmap->h_key_dtor((void *)item->bi_key.key_data);
	}

	if (hashmap->h_value_dtor) {
		hashmap->h_value_dtor((void *)item->bi_value.value_data);
	}

	free(item);

	if (b_queue_empty(&bucket->bk_items)) {
		b_btree_delete(&hashmap->h_buckets, &bucket->bk_node);
		free(bucket);
	}

	return B_SUCCESS;
}

static bool hashmap_iterator_next(struct b_iterator *it)
{
	return b_hashmap_iterator_next((struct b_hashmap_iterator *)it);
}

static b_status hashmap_iterator_erase(struct b_iterator *it)
{
	return b_hashmap_iterator_erase((struct b_hashmap_iterator *)it);
}

static bool hashmap_iterator_is_valid(const struct b_iterator *it)
{
	return b_hashmap_iterator_is_valid((struct b_hashmap_iterator *)it);
}

static struct b_iterator_ops it_ops = {
	.it_next = hashmap_iterator_next,
	.it_erase = hashmap_iterator_erase,
	.it_close = NULL,
	.it_is_valid = hashmap_iterator_is_valid,
};

int b_hashmap_iterator_begin(
	struct b_hashmap *hashmap, struct b_hashmap_iterator *it)
{
	it->_h = hashmap;
	it->_base.it_ops = &it_ops;

	it->i = 0;
	if (b_hashmap_is_empty(hashmap)) {
		it->key = NULL;
		it->value = NULL;
		return -1;
	}

	struct b_btree_node *first_node = b_btree_first(&hashmap->h_buckets);
	struct b_hashmap_bucket *first_bucket
		= b_unbox(struct b_hashmap_bucket, first_node, bk_node);
	if (!first_bucket) {
		it->key = NULL;
		it->value = NULL;
		return -1;
	}

	struct b_queue_entry *first_entry = b_queue_first(&first_bucket->bk_items);
	struct b_hashmap_bucket_item *first_item
		= b_unbox(struct b_hashmap_bucket_item, first_entry, bi_entry);
	if (!first_item) {
		it->key = NULL;
		it->value = NULL;
		return -1;
	}

	it->key = &first_item->bi_key;
	it->value = &first_item->bi_value;

	it->_cbn = first_node;
	it->_cqe = first_entry;

	return 0;
}

static bool get_next_node(
	struct b_btree_node *cur_node, struct b_queue_entry *cur_entry,
	struct b_btree_node **out_next_node, struct b_queue_entry **out_next_entry)
{
	struct b_hashmap_bucket *cur_bucket
		= b_unbox(struct b_hashmap_bucket, cur_node, bk_node);
	if (!cur_bucket) {
		return false;
	}

	struct b_hashmap_bucket_item *cur_item
		= b_unbox(struct b_hashmap_bucket_item, cur_entry, bi_entry);
	if (!cur_item) {
		return false;
	}

	struct b_btree_node *next_node = cur_node;
	struct b_queue_entry *next_entry = b_queue_next(cur_entry);
	if (!next_entry) {
		next_node = b_btree_next(cur_node);
		if (!next_node) {
			return false;
		}

		struct b_hashmap_bucket *next_bucket
			= b_unbox(struct b_hashmap_bucket, next_node, bk_node);
		if (!next_bucket) {
			return false;
		}

		next_entry = b_queue_first(&next_bucket->bk_items);
		if (!next_entry) {
			return false;
		}
	}

	struct b_hashmap_bucket_item *next_item
		= b_unbox(struct b_hashmap_bucket_item, next_entry, bi_entry);
	if (!next_item) {
		return false;
	}

	*out_next_node = next_node;
	*out_next_entry = next_entry;

	return true;
}

bool b_hashmap_iterator_next(struct b_hashmap_iterator *it)
{
	struct b_btree_node *next_node;
	struct b_queue_entry *next_entry;
	if (!get_next_node(it->_cbn, it->_cqe, &next_node, &next_entry)) {
		it->key = NULL;
		it->value = NULL;
		return false;
	}

	struct b_hashmap_bucket_item *next_item
		= b_unbox(struct b_hashmap_bucket_item, next_entry, bi_entry);

	if (!next_item) {
		it->key = NULL;
		it->value = NULL;
		return false;
	}

	it->i++;
	it->key = &next_item->bi_key;
	it->value = &next_item->bi_value;

	it->_cbn = next_node;
	it->_cqe = next_entry;

	return true;
}

b_status b_hashmap_iterator_erase(struct b_hashmap_iterator *it)
{
	if ((it->key || it->value) && !(it->_cbn && it->_cqe)) {
		return B_ERR_BAD_STATE;
	}

	if (!it->key || !it->_cqe) {
		return B_ERR_OUT_OF_BOUNDS;
	}

	struct b_btree_node *next_node;
	struct b_queue_entry *next_entry;
	if (!get_next_node(it->_cbn, it->_cqe, &next_node, &next_entry)) {
		it->key = NULL;
		it->value = NULL;
		return false;
	}

	struct b_hashmap_bucket *cur_bucket
		= b_unbox(struct b_hashmap_bucket, it->_cbn, bk_node);
	struct b_hashmap_bucket_item *cur_item
		= b_unbox(struct b_hashmap_bucket_item, it->_cqe, bi_entry);

	struct b_hashmap_bucket_item *next_item
		= b_unbox(struct b_hashmap_bucket_item, next_entry, bi_entry);

	b_status status = delete_item(it->_h, cur_bucket, cur_item);
	if (B_ERR(status)) {
		return status;
	}

	if (next_item) {
		it->key = &next_item->bi_key;
		it->value = &next_item->bi_value;

		it->_cbn = next_node;
		it->_cqe = next_entry;
	} else {
		it->key = NULL;
		it->value = NULL;

		it->_cbn = NULL;
		it->_cqe = NULL;
	}

	return B_SUCCESS;
}

bool b_hashmap_iterator_is_valid(const struct b_hashmap_iterator *it)
{
	return it->key != NULL;
}

static void hashmap_release(struct b_object *obj)
{
	struct b_hashmap *map = B_HASHMAP(obj);

	b_hashmap_iterator it;
	b_hashmap_iterator_begin(map, &it);
	while (b_hashmap_iterator_is_valid(&it)) {
		b_hashmap_iterator_erase(&it);
	}
}

b_object_type_id b_hashmap_type_id(void)
{
	return (b_object_type_id)&hashmap_type;
}