archeus/src/std/hashtable.c

#include "arc/std/hashtable.h"

#include "arc/std/errno.h"
#include <stdlib.h>
#include <stdio.h>

//a private struct to hold the keys and values of the hashtable
typedef struct ARC_HashtableNode ARC_HashtableNode;
struct ARC_HashtableNode {
    void *key;
    void *value;

    uint32_t hashvalue;

    //will be set if next slot is searched for, to be used to remove elements faster
    uint32_t nextIndex;
};

struct ARC_Hashtable {
    uint32_t currentCapacity;
    uint32_t currentSize;

    ARC_HashtableNode *nodes;

    ARC_Hashtable_HashFn             hashFn;
    ARC_Hashtable_KeyCompareFn       keyCompareFn;
    ARC_Hashtable_DestroyKeyValueFn *destroyKeyValueFn;
};

//copied from here: https://en.wikipedia.org/wiki/Computation_of_cyclic_redundancy_checks#CRC-32_example
uint32_t CRC32Fn(void *key){
    uint32_t hashvalue = 0xffffffff;

    for(uint32_t i = 0; *(((char *)key) + i) != '\0'; i++){
        uint8_t value = *(((uint8_t *)key) + i);

        for(uint8_t j = 0; j < 8; j++){
            uint8_t flag = (uint8_t)((value ^ hashvalue) & 1);
            hashvalue >>= 1;

            if(flag){
                hashvalue ^= 0xEDB888320;
            }

            value >>= 1;
        }
    }

    hashvalue = ~hashvalue;

    return hashvalue;
}

ARC_Bool ARC_Hashtable_DefaultKeyCompareFn(void *key1, void *key2){
    return (ARC_Bool)(key1 == key2);
}

void ARC_Hashtable_Create(ARC_Hashtable **hashtable, ARC_Hashtable_HashFn *hashFn, ARC_Hashtable_KeyCompareFn *keyCompareFn, ARC_Hashtable_DestroyKeyValueFn *destroyKeyValueFn){
    //clear the hashtable
    *hashtable = (ARC_Hashtable *) malloc(sizeof(ARC_Hashtable));

    //set current capacity and size to start
    (*hashtable)->currentCapacity = 1;
    (*hashtable)->currentSize     = 0;

    //reserve enough memory for one node
    (*hashtable)->nodes = (ARC_HashtableNode *)malloc(sizeof(ARC_HashtableNode));

    //set first and only key to null
    (*hashtable)->nodes[0] = (ARC_HashtableNode){ NULL, NULL, 0, 0 };

    //default to CRC32, then override if hashFn exists
    (*hashtable)->hashFn = CRC32Fn;
    if(hashFn != NULL){
        (*hashtable)->hashFn = *hashFn;
    }

    //default to comparing pointers, then override if keyCompareFn exists
    (*hashtable)->keyCompareFn = ARC_Hashtable_DefaultKeyCompareFn;
    if(keyCompareFn != NULL){
        (*hashtable)->keyCompareFn = *keyCompareFn;
    }

    //default to NULL, then create and copy destroyKeyValueFn if it exists
    (*hashtable)->destroyKeyValueFn = NULL;
    if(destroyKeyValueFn != NULL){
        (*hashtable)->destroyKeyValueFn = (ARC_Hashtable_DestroyKeyValueFn *)malloc(sizeof(ARC_Hashtable_DestroyKeyValueFn));
        *((*hashtable)->destroyKeyValueFn) = *destroyKeyValueFn;
    }
}

void ARC_Hashtable_Destroy(ARC_Hashtable *hashtable){
    //remove all the contents before clearing the 
    ARC_Hashtable_Clear(hashtable);

    //free the destroyKeyValueFn if it exists
    if(hashtable->destroyKeyValueFn != NULL){
        free(hashtable->destroyKeyValueFn);
    }

    //free the empty nodes container
    free(hashtable->nodes);

    //free the hashtable
    free(hashtable);
}

void ARC_HashtableNode_SetNearestNodeToArray(ARC_HashtableNode *nodes, uint32_t capacity, ARC_HashtableNode node){
    //get the first possible index based on the node's hashvalue
    uint32_t index = node.hashvalue % capacity;

    //go to last added
    while(nodes[index].nextIndex != index){
        index = nodes[index].nextIndex;
    }

    //init variable for found node
    uint32_t nextIndex = index;

    //check each available node for a free slot
    while(nodes[nextIndex].key != NULL){
        //up the current index by one
        nextIndex++;

        //cycle back to the first index if it is above the array's capacity
        if(nextIndex >= capacity){
            nextIndex = 0;
        }

        //check if the loop has circled back to the starting index to stop checking
        if(index == nextIndex){
            break;
        }
    }

    //set the next index of the last added index
    nodes[index].nextIndex = nextIndex;

    //set the foundNode and next index
    nodes[nextIndex]           = node;
    nodes[nextIndex].nextIndex = nextIndex;
}

void ARC_Hashtable_Add(ARC_Hashtable *hashtable, void *key, void *value){
    //check to see if the current size is the same as a max uint32_t and if so it will overflow so throw an error
    if(hashtable->currentSize == ~((uint32_t)0)){
        arc_errno = ARC_ERRNO_OVERFLOW;
        ARC_DEBUG_LOG_ERROR("ARC_Hashtable_Add(hashtable, key, value), hashtable at max capacity tried adding another value");
        return;
    }

    //check to make sure key is not NULL
    if(key == NULL){
        arc_errno = ARC_ERRNO_NULL;
        ARC_DEBUG_LOG_ERROR("ARC_Hashtable_Add(hashtable, key, value), NULL was passed in for the key, this function cannot handle that");
        return;
    }

    //check if we are at the max of the current capacity
    if(hashtable->currentSize == hashtable->currentCapacity){
        //move the current nodes into a temporary variable to move into a resized array
        uint64_t oldCapacity = hashtable->currentCapacity;
        ARC_HashtableNode *oldNodes = hashtable->nodes;

        //increase the current capacity by double
        hashtable->currentCapacity <<= 1;

        //if for some reason the capacity is 0, we should set it to one so we do not error on realloc
        if(hashtable->currentCapacity != 0){
            hashtable->currentCapacity++;
        }

        //resize the hashtable's array and copy the contents at the same time
        hashtable->nodes = (ARC_HashtableNode *)malloc(sizeof(ARC_HashtableNode) * hashtable->currentCapacity);

        //set nodes to null
        for(uint32_t index = 0; index < hashtable->currentCapacity; index++){
            hashtable->nodes[index] = (ARC_HashtableNode){ NULL, NULL, 0, index };
        }

        //add the old nodes into the new array
        for(uint32_t index = 0; index < oldCapacity; index++){
            ARC_HashtableNode_SetNearestNodeToArray(hashtable->nodes, hashtable->currentCapacity, oldNodes[index]);
        }

        //free the old array
        free(oldNodes);
    }

    //get the hashvalue
    uint32_t hashvalue = hashtable->hashFn(key);

    //add to the vectors array and increase its current size
    ARC_HashtableNode_SetNearestNodeToArray(hashtable->nodes, hashtable->currentCapacity, (ARC_HashtableNode){ key, value, hashvalue, 0 });
    hashtable->currentSize++;
}

void ARC_Hashtable_Remove(ARC_Hashtable *hashtable, void *key){
    //get the index from a hashvalue
    uint32_t initialIndex = hashtable->hashFn(key) % hashtable->currentCapacity;
    uint32_t index = initialIndex;

    //get the first possible node
    ARC_HashtableNode node = hashtable->nodes[index];

    //check each available node for a match and break if the current nodes doesn't hold anything
    ARC_Bool nodeFound = ARC_False;
    while(node.key != NULL){
        if(hashtable->keyCompareFn(node.key, key) == ARC_True){
            nodeFound = ARC_True;
            break;
        }

        //up the current index by one
        index++;

        //cycle back to the first index if it is above the array's capacity
        if(index >= hashtable->currentCapacity){
            index = 0;
        }

        //check if the loop has circled back to the starting index to stop checking
        if(index == initialIndex){
            break;
        }

        //get the next possible node
        node = hashtable->nodes[index];
    }

    //error if the node was not found
    if(nodeFound == ARC_False){
        arc_errno = ARC_ERRNO_DATA;
        ARC_DEBUG_LOG_ERROR("ARC_Hashtable_Remove(hashtable, key), key was not found in hashtable, could not remove");
        return;
    }

    //call delete data to clean up item if delete data function exists
    if(hashtable->destroyKeyValueFn != NULL){
        (*(hashtable->destroyKeyValueFn))(node.key, node.value);
    }

    //cycle back to the first index if it is above the array's capacity
    if(index >= hashtable->currentCapacity){
        index = 0;
    }

    //while the current node needs to be moved back becuase it is offset to the initial index
    while(hashtable->nodes[index].nextIndex != index){
        //get the currently used node
        node = hashtable->nodes[index];

        //move the current node back one
        hashtable->nodes[index] = hashtable->nodes[node.nextIndex];

        //get the next index to move back
        index = hashtable->nodes[index].nextIndex;

        //moves the next index into the next used slot
        hashtable->nodes[index].nextIndex = node.nextIndex;
    }

    //set the current value to an empty node
    hashtable->nodes[index] = (ARC_HashtableNode){ NULL, NULL, 0, index };

    //we have removed the item so we can decrease the current size
    hashtable->currentSize--;

    //if the current size is half the current capacity or the current capacity is at the smallest limit, we do not need to do anything else
    if(hashtable->currentSize != hashtable->currentCapacity >> 1 || hashtable->currentCapacity == 1){
        return;
    }

    //move the current nodes into a temporary variable to move into a resized array
    uint64_t oldCapacity = hashtable->currentCapacity;
    ARC_HashtableNode *oldNodes = hashtable->nodes;

    //half the capacity and copy it into a smaller array
    hashtable->currentCapacity >>= 1;

    //resize the hashtable's array and copy the contents at the same time
    hashtable->nodes = (ARC_HashtableNode *)malloc(sizeof(ARC_HashtableNode) * hashtable->currentCapacity);

    //set keys to null
    for(uint32_t index = 0; index < hashtable->currentCapacity; index++){
        hashtable->nodes[index] = (ARC_HashtableNode){ NULL, NULL, 0, index };
    }

    //add the old nodes into the new array
    for(uint32_t index = 0; index < oldCapacity; index++){
        ARC_HashtableNode_SetNearestNodeToArray(hashtable->nodes, hashtable->currentCapacity, oldNodes[index]);
    }

    //free the old array
    free(oldNodes);
}

void ARC_Hashtable_Clear(ARC_Hashtable *hashtable){
    //delete the array holding all the nodes
    free(hashtable->nodes);

    //set current capacity and size to start
    hashtable->currentCapacity = 1;
    hashtable->currentSize     = 0;

    //reserve enough memory for one node
    hashtable->nodes = (ARC_HashtableNode *)malloc(sizeof(ARC_HashtableNode));

    //set first and only key to null
    hashtable->nodes[0].key = NULL;
}

void *ARC_Hashtable_Get(ARC_Hashtable *hashtable, void *key){
    //check to make sure key is not NULL
    if(key == NULL){
        arc_errno = ARC_ERRNO_NULL;
        ARC_DEBUG_LOG_ERROR("ARC_Hashtable_Get(hashtable, key), NULL was passed in for the key, this function cannot handle that");
        return NULL;
    }

    //get the index from a hashvalue
    uint32_t index = hashtable->hashFn(key) % hashtable->currentCapacity;

    //get the first possible node
    ARC_HashtableNode node = hashtable->nodes[index];

    //check each available node for a match
    for(uint32_t nextIndex = index; node.key != NULL; node = hashtable->nodes[nextIndex]){
        //if the key is found, return its value
        if(hashtable->keyCompareFn(node.key, key) == ARC_True){
            return node.value;
        }

        //up the current index by one
        nextIndex++;

        //cycle back to the first index if it is above the array's capacity
        if(nextIndex >= hashtable->currentCapacity){
            nextIndex = 0;
        }

        //check if the loop has circled back to the starting index to stop checking
        if(index == nextIndex){
            break;
        }
    }

    //if the key is found, return its value
    if(node.key != NULL && hashtable->keyCompareFn(node.key, key) == ARC_True){
        return node.value;
    }

    //could not find node, so return NULL
    return NULL;
}

void ARC_Hashtable_RunIteration(ARC_Hashtable *hashtable, ARC_Hashtable_IteratorFn iteratorFn){
    //pass each non NULL nodes into an iteratorFn callback
    for(uint32_t index = 0; index < hashtable->currentCapacity; index++){
        //get the current node
        ARC_HashtableNode node = hashtable->nodes[index];

        //skip past NULL keys
        if(node.key == NULL){
            continue;
        }

        //passes current iteration into the callback function
        iteratorFn(node.key, node.value, index, node.nextIndex);
    }
}