refactor: rename TernaryTrie to Trie
ci/woodpecker/push/woodpecker Pipeline was successful Details

trie-skips
Jef Roosens 2022-11-29 11:27:28 +01:00
parent cc8cfaeace
commit 4bcdd5c4d9
Signed by: Jef Roosens
GPG Key ID: B75D4F293C7052DB
5 changed files with 95 additions and 98 deletions

View File

@ -12,4 +12,4 @@ if(CMAKE_BUILD_TYPE STREQUAL Release)
add_compile_options(-O3 -flto)
endif()
add_executable(lander src/main.cpp src/tries/ternarytrie.c)
add_executable(lander src/main.cpp src/trie.c)

View File

@ -29,7 +29,7 @@ static const size_t charset_len = sizeof(charset) - 1;
*
* You can (and should) redefine this in your c-file with the concrete fields.
*/
typedef struct ttrie TernaryTrie;
typedef struct ttrie Trie;
typedef enum entry_type { Redirect, Paste, Unknown } EntryType;
@ -45,7 +45,7 @@ Entry *entry_new(EntryType type, const char *string);
*
* @return a pointer to an empty Trie struct
*/
TernaryTrie *ternarytrie_init();
Trie *trie_init();
/**
* Populate trie with entries stored in the given file.
@ -54,14 +54,14 @@ TernaryTrie *ternarytrie_init();
* @param file_path path to file containing entries
* @return amount of entries added; -1 if an error occured
*/
int ternarytrie_populate(TernaryTrie *trie, const char *file_path);
int trie_populate(Trie *trie, const char *file_path);
/**
* De-allocate a trie by freeing the memory occupied by this trie.
*
* @param trie which should be freed
*/
void ternarytrie_free(TernaryTrie *trie);
void trie_free(Trie *trie);
/**
* Search for an entry in the trie.
@ -70,7 +70,7 @@ void ternarytrie_free(TernaryTrie *trie);
* @param key key representing the entry
* @return pointer to entry; NULL if not found
*/
Entry *ternarytrie_search(TernaryTrie *trie, const char *key);
Entry *trie_search(Trie *trie, const char *key);
/**
* Add a string to this trie.
@ -81,7 +81,7 @@ Entry *ternarytrie_search(TernaryTrie *trie, const char *key);
* @return true if the trie was changed by this operation, false if it was
* already present
*/
bool ternarytrie_add(TernaryTrie *trie, const char *key, Entry *entry);
bool trie_add(Trie *trie, const char *key, Entry *entry);
/**
* Add an entry by generating a random string as the key.
@ -91,7 +91,7 @@ bool ternarytrie_add(TernaryTrie *trie, const char *key, Entry *entry);
* @param secure whether to generate a longer, more secure random key
* @return the generated key
*/
char *ternarytrie_add_random(TernaryTrie *trie, Entry *entry, bool secure);
char *trie_add_random(Trie *trie, Entry *entry, bool secure);
/**
* Remove an entry from this trie given its key.
@ -101,7 +101,7 @@ char *ternarytrie_add_random(TernaryTrie *trie, Entry *entry, bool secure);
* @return true if the entry was present and has been removed, false if it was
* not present
*/
bool ternarytrie_remove(TernaryTrie *trie, const char *key);
bool trie_remove(Trie *trie, const char *key);
/**
* Returns the number of entries in this trie.
@ -109,6 +109,6 @@ bool ternarytrie_remove(TernaryTrie *trie, const char *key);
* @param trie
* @return the number of entries in this trie
*/
size_t ternarytrie_size(TernaryTrie *trie);
size_t trie_size(Trie *trie);
#endif // AD3_TERNARYTRIE

View File

@ -4,7 +4,7 @@
#include "crow.h"
extern "C" {
#include "ternarytrie.h"
#include "trie.h"
}
static const std::string index_page = R"(
@ -31,10 +31,10 @@ static const std::string index_page = R"(
return crow::response(crow::status::UNAUTHORIZED); \
}
crow::response add_redirect(std::string base_url, TernaryTrie *trie,
crow::response add_redirect(std::string base_url, Trie *trie,
const char *url, bool secure) {
Entry *new_entry = entry_new(Redirect, url);
char *key = ternarytrie_add_random(trie, new_entry, secure);
char *key = trie_add_random(trie, new_entry, secure);
if (key == NULL) {
return crow::response(crow::status::INTERNAL_SERVER_ERROR);
@ -61,10 +61,10 @@ bool store_paste(const char *key, const char *body) {
return true;
}
crow::response add_paste(std::string base_url, TernaryTrie *trie,
crow::response add_paste(std::string base_url, Trie *trie,
const char *body, bool secure) {
Entry *new_entry = entry_new(Paste, "");
char *key = ternarytrie_add_random(trie, new_entry, secure);
char *key = trie_add_random(trie, new_entry, secure);
if (key == NULL) {
return crow::response(crow::status::INTERNAL_SERVER_ERROR);
@ -88,14 +88,14 @@ int main() {
ENV(base_url, "LANDER_BASE_URL");
// Initialize trie and populate from data file
TernaryTrie *trie = ternarytrie_init();
Trie *trie = trie_init();
std::string file_path = "lander.data";
std::cout << "Populating trie from file '" << file_path << "'..."
<< std::endl;
int count = ternarytrie_populate(trie, file_path.c_str());
int count = trie_populate(trie, file_path.c_str());
if (count == -1) {
std::cout << "An error occured while populating the trie." << std::endl;
@ -119,7 +119,7 @@ int main() {
CROW_ROUTE(app, "/<string>")
.methods(crow::HTTPMethod::Get)(
[trie](crow::response &res, std::string key) {
Entry *entry = ternarytrie_search(trie, key.c_str());
Entry *entry = trie_search(trie, key.c_str());
if (entry != NULL) {
if (entry->type == Redirect) {
@ -160,7 +160,7 @@ int main() {
Entry *new_entry = entry_new(Redirect, req.body.c_str());
bool added = ternarytrie_add(trie, key.c_str(), new_entry);
bool added = trie_add(trie, key.c_str(), new_entry);
if (!added) {
return crow::response(crow::status::CONFLICT);
@ -194,7 +194,7 @@ int main() {
AUTH();
Entry *new_entry = entry_new(Paste, "");
bool added = ternarytrie_add(trie, key.c_str(), new_entry);
bool added = trie_add(trie, key.c_str(), new_entry);
if (!added) {
return crow::response(crow::status::CONFLICT);

View File

@ -4,23 +4,23 @@
#include <stdlib.h>
#include <string.h>
#include "ternarytrie.h"
#include "ternarytrie_node.c"
#include "trie.h"
#include "trie_node.c"
typedef struct ttrie {
TernaryTrieNode *root;
TrieNode *root;
size_t size;
char *file_path;
pthread_rwlock_t lock;
} TernaryTrie;
} Trie;
/**
* Allocate and initialize an empty TernaryTrie
* Allocate and initialize an empty Trie
*
* @return pointer to the empty TernaryTrie
* @return pointer to the empty Trie
*/
TernaryTrie *ternarytrie_init() {
TernaryTrie *trie = (TernaryTrie *)calloc(1, sizeof(TernaryTrie));
Trie *trie_init() {
Trie *trie = calloc(1, sizeof(Trie));
trie->root = ttnode_init();
pthread_rwlock_init(&trie->lock, NULL);
@ -32,12 +32,12 @@ TernaryTrie *ternarytrie_init() {
*
* @param trie trie to free
*/
void ternarytrie_free(TernaryTrie *trie) {
void trie_free(Trie *trie) {
ttnode_free(trie->root);
free(trie);
}
bool ternarytrie_add_internal(TernaryTrie *trie, const char *key, Entry *entry);
bool trie_add_internal(Trie *trie, const char *key, Entry *entry);
EntryType entry_type_from_char(char c) {
switch (c) {
@ -62,7 +62,7 @@ char entry_type_to_char(EntryType et) {
}
Entry *entry_new(EntryType type, const char *string) {
Entry *entry = (Entry *)malloc(sizeof(Entry));
Entry *entry = malloc(sizeof(Entry));
entry->type = type;
if (string != NULL) {
@ -74,7 +74,7 @@ Entry *entry_new(EntryType type, const char *string) {
return entry;
}
int ternarytrie_populate(TernaryTrie *trie, const char *file_path) {
int trie_populate(Trie *trie, const char *file_path) {
trie->file_path = strdup(file_path);
FILE *fp = fopen(file_path, "r");
@ -116,7 +116,7 @@ int ternarytrie_populate(TernaryTrie *trie, const char *file_path) {
buffer[j] = '\0';
entry = entry_new(type, buffer + i + 3);
ternarytrie_add_internal(trie, buffer, entry);
trie_add_internal(trie, buffer, entry);
entries++;
}
@ -127,11 +127,11 @@ int ternarytrie_populate(TernaryTrie *trie, const char *file_path) {
}
typedef struct searchresult {
TernaryTrieNode *parent;
TernaryTrieNode *child;
TrieNode *parent;
TrieNode *child;
} SearchResult;
SearchResult ternarytrie_search_node(TernaryTrie *trie, const char *key) {
SearchResult trie_search_node(Trie *trie, const char *key) {
SearchResult out = {NULL, NULL};
// Edge case for empty string
@ -144,8 +144,8 @@ SearchResult ternarytrie_search_node(TernaryTrie *trie, const char *key) {
}
size_t i = 0;
TernaryTrieNode **node_ptr = &(trie->root);
TernaryTrieNode **child_ptr;
TrieNode **node_ptr = &(trie->root);
TrieNode **child_ptr;
do {
child_ptr = ttnode_search(*node_ptr, key[i], false);
@ -188,10 +188,10 @@ SearchResult ternarytrie_search_node(TernaryTrie *trie, const char *key) {
* @param string string to look up
* @return true if the string is present in the trie, false otherwise
*/
Entry *ternarytrie_search(TernaryTrie *trie, const char *key) {
Entry *trie_search(Trie *trie, const char *key) {
pthread_rwlock_rdlock(&trie->lock);
SearchResult res = ternarytrie_search_node(trie, key);
SearchResult res = trie_search_node(trie, key);
Entry *return_value = NULL;
@ -205,14 +205,14 @@ Entry *ternarytrie_search(TernaryTrie *trie, const char *key) {
}
/**
* Add the given string to the TernaryTrie.
* Add the given string to the Trie.
*
* @param trie trie to add string to
* @param string string to add
* @return true if the string wasn't present in the trie and thus added, false
* otherwise
*/
bool ternarytrie_add_internal(TernaryTrie *trie, const char *string,
bool trie_add_internal(Trie *trie, const char *string,
Entry *entry) {
// Edge case for empty string
if (string[0] == DELIMITER) {
@ -228,8 +228,8 @@ bool ternarytrie_add_internal(TernaryTrie *trie, const char *string,
}
size_t i = 0;
TernaryTrieNode **node_ptr = &(trie->root);
TernaryTrieNode **new_node_ptr;
TrieNode **node_ptr = &(trie->root);
TrieNode **new_node_ptr;
do {
new_node_ptr = ttnode_search(*node_ptr, string[i], true);
@ -256,7 +256,7 @@ bool ternarytrie_add_internal(TernaryTrie *trie, const char *string,
// The next node in the string's path doesn't exist yet, so we add it to the
// trie
if (*node_ptr == NULL) {
TernaryTrieNode *new_node = ttnode_init();
TrieNode *new_node = ttnode_init();
// If there's a remaining part of the string, we add it to the leaf
if (string[i] != DELIMITER) {
@ -296,7 +296,7 @@ bool ternarytrie_add_internal(TernaryTrie *trie, const char *string,
return true;
}
bool ternarytrie_add_persistent(TernaryTrie *trie, const char *key,
bool trie_add_persistent(Trie *trie, const char *key,
Entry *entry) {
bool return_value = false;
@ -304,7 +304,7 @@ bool ternarytrie_add_persistent(TernaryTrie *trie, const char *key,
// Easiest way to make sure we don't add duplicate entries
// We use an internal function that doesn't require a read lock, as we're
// already inside a write lock
if (ternarytrie_search_node(trie, key).child != NULL) {
if (trie_search_node(trie, key).child != NULL) {
return false;
}
@ -326,26 +326,26 @@ bool ternarytrie_add_persistent(TernaryTrie *trie, const char *key,
// This function *should* always return true. Otherwise, the function would've
// exited because the string was found in the trie.
return ternarytrie_add_internal(trie, key, entry);
return trie_add_internal(trie, key, entry);
}
bool ternarytrie_add(TernaryTrie *trie, const char *key, Entry *entry) {
bool trie_add(Trie *trie, const char *key, Entry *entry) {
pthread_rwlock_wrlock(&trie->lock);
bool return_value = ternarytrie_add_persistent(trie, key, entry);
bool return_value = trie_add_persistent(trie, key, entry);
pthread_rwlock_unlock(&trie->lock);
return return_value;
}
char *ternarytrie_add_random(TernaryTrie *trie, Entry *entry, bool secure) {
char *trie_add_random(Trie *trie, Entry *entry, bool secure) {
pthread_rwlock_wrlock(&trie->lock);
// Generate random key
bool ok = false;
int key_length = secure ? RANDOM_KEY_LENGTH_LONG : RANDOM_KEY_LENGTH_SHORT;
char *key = (char *)malloc(key_length + 1);
char *key = malloc(key_length + 1);
key[key_length] = '\0';
// We naively generate new keys until we find a key that isn't in the trie
@ -356,10 +356,10 @@ char *ternarytrie_add_random(TernaryTrie *trie, Entry *entry, bool secure) {
key[i] = charset[rand() % charset_len];
}
ok = ternarytrie_search_node(trie, key).child == NULL;
ok = trie_search_node(trie, key).child == NULL;
}
bool res = ternarytrie_add_persistent(trie, key, entry);
bool res = trie_add_persistent(trie, key, entry);
char *return_value;
if (res) {
@ -375,18 +375,18 @@ char *ternarytrie_add_random(TernaryTrie *trie, Entry *entry, bool secure) {
}
/**
* Remove the given string from a TernaryTrie.
* Remove the given string from a Trie.
*
* @param trie trie to remove string from
* @param string string to remove
* @return true if the string was in the trie and thus removed, false otherwise
*/
bool ternarytrie_remove(TernaryTrie *trie, const char *string) {
bool trie_remove(Trie *trie, const char *string) {
pthread_rwlock_wrlock(&trie->lock);
bool return_value = false;
SearchResult res = ternarytrie_search_node(trie, string);
SearchResult res = trie_search_node(trie, string);
if (res.child == NULL) {
goto end;
@ -439,4 +439,4 @@ end:
* @param trie trie to return size for
* @return size of the trie
*/
size_t ternarytrie_size(TernaryTrie *trie) { return trie->size; }
size_t trie_size(Trie *trie) { return trie->size; }

View File

@ -2,21 +2,21 @@
#include <stdint.h>
#include <stdlib.h>
#include "ternarytrie.h"
#include "trie.h"
/**
* Represents a node of the binary tree contained within each non-leaf
* TernaryTrieNode.
* TrieNode.
*/
typedef struct ttinode {
struct ttinode *left;
struct ttinode *right;
struct ttnode *next;
char key;
} TernaryTrieInnerNode;
} TrieInnerNode;
/**
* Represents a node inside a TernaryTrie. A node can be in one of three states:
* Represents a node inside a Trie. A node can be in one of three states:
* - Internal node: a node that's part of a path to a leaf node. This node will
* always have a size greater than one, and an initialized root.
* - Leaf: a node solely used to represent a string ending there. Its size is 0,
@ -28,7 +28,7 @@ typedef struct ttinode {
*/
typedef struct ttnode {
union {
TernaryTrieInnerNode *root;
TrieInnerNode *root;
char *string;
} ptr;
Entry *entry;
@ -41,43 +41,40 @@ typedef struct ttnode {
// 0, 1: size of underlying binary tree
// 2: length of string
uint8_t size;
} TernaryTrieNode;
} TrieNode;
// Required for recursively freeing tree structure
void ttnode_free(TernaryTrieNode *node);
void ttnode_free(TrieNode *node);
/**
* Allocate and initialize a new TernaryTrieInnerNode representing a given
* Allocate and initialize a new TrieInnerNode representing a given
* character.
*
* @param c character to represent
* @return pointer to newly allocated struct
*/
TernaryTrieInnerNode *ttinode_init(char c) {
TernaryTrieInnerNode *node =
(TernaryTrieInnerNode *)calloc(1, sizeof(TernaryTrieInnerNode));
TrieInnerNode *ttinode_init(char c) {
TrieInnerNode *node = calloc(1, sizeof(TrieInnerNode));
node->key = c;
return node;
}
/**
* Allocate and initialize a new TernaryTrieNode.
* Allocate and initialize a new TrieNode.
*
* @return pointer to newly allocated struct
*/
TernaryTrieNode *ttnode_init() {
return (TernaryTrieNode *)calloc(1, sizeof(TernaryTrieNode));
}
TrieNode *ttnode_init() { return calloc(1, sizeof(TrieNode)); }
/**
* Free a TernaryTrieInnerNode and its underlying tree structure. This should
* Free a TrieInnerNode and its underlying tree structure. This should
* usually only be called on the root of a binary tree to free the entire
* structure.
*
* @param node node whose tree to free
*/
void ttinode_free_cascade(TernaryTrieInnerNode *node) {
void ttinode_free_cascade(TrieInnerNode *node) {
if (node->left != NULL) {
ttinode_free_cascade(node->left);
}
@ -94,11 +91,11 @@ void ttinode_free_cascade(TernaryTrieInnerNode *node) {
}
/**
* Free a TernaryTrieNode and its underlying tree structure.
* Free a TrieNode and its underlying tree structure.
*
* @param node node to free
*/
void ttnode_free(TernaryTrieNode *node) {
void ttnode_free(TrieNode *node) {
if (node->type == 2) {
free(node->ptr.string);
} else if (node->size != 0) {
@ -119,7 +116,7 @@ void ttnode_free(TernaryTrieNode *node) {
* @param node node to add string to
* @param string string to add
*/
void ttnode_set_string(TernaryTrieNode *node, const char *string) {
void ttnode_set_string(TrieNode *node, const char *string) {
node->type = 2;
node->size = strlen(string);
node->ptr.string = strdup(string);
@ -127,8 +124,8 @@ void ttnode_set_string(TernaryTrieNode *node, const char *string) {
/**
* This function performs a lookup in the underlying binary tree of the given
* TernaryTrieNode. If found, the return value is a pointer to the memory
* location where the TernaryTrieInnerNode representing the given character
* TrieNode. If found, the return value is a pointer to the memory
* location where the TrieInnerNode representing the given character
* stores its `next` field. If not found, the return value is NULL, unless
* `create` is true.
*
@ -138,12 +135,12 @@ void ttnode_set_string(TernaryTrieNode *node, const char *string) {
*
* @param node node to perform lookup in. If node is a full leaf, the return
* value will always be NULL, regardless of the value of create.
* @param create whether to create the TernaryTrieInnerNode if it isn't present
* @param create whether to create the TrieInnerNode if it isn't present
* yet. If this is set to true, the function will never return NULL unless the
* node represents a leaf with a string, because the struct and therefore the
* address is created if it doesn't exist yet.
*/
TernaryTrieNode **ttnode_search(TernaryTrieNode *node, const char c,
TrieNode **ttnode_search(TrieNode *node, const char c,
bool create) {
// Full leafs will always return NULL
if (node->type == 2) {
@ -162,8 +159,8 @@ TernaryTrieNode **ttnode_search(TernaryTrieNode *node, const char c,
return NULL;
}
TernaryTrieInnerNode *parent = node->ptr.root;
TernaryTrieInnerNode *child;
TrieInnerNode *parent = node->ptr.root;
TrieInnerNode *child;
// Iterate through the tree until we either find the character or realize it's
// not present in the tree
@ -189,7 +186,7 @@ TernaryTrieNode **ttnode_search(TernaryTrieNode *node, const char c,
// If create is true, we create the new node so that we can still return a
// non-NULL pointer.
if (create) {
TernaryTrieInnerNode *new_node = ttinode_init(c);
TrieInnerNode *new_node = ttinode_init(c);
if (c < parent->key) {
parent->left = new_node;
@ -211,8 +208,8 @@ TernaryTrieNode **ttnode_search(TernaryTrieNode *node, const char c,
*
* @param node node to split
*/
void ttnode_split(TernaryTrieNode *node) {
TernaryTrieNode *new_node = ttnode_init();
void ttnode_split(TrieNode *node) {
TrieNode *new_node = ttnode_init();
char key = node->ptr.string[0];
// There's a chance the remaining string was only 1 character, meaning the new
@ -233,16 +230,16 @@ void ttnode_split(TernaryTrieNode *node) {
node->ptr.string = NULL;
// Initialize node's binary tree with the correct character
TernaryTrieNode **node_ptr = ttnode_search(node, key, true);
TrieNode **node_ptr = ttnode_search(node, key, true);
*node_ptr = new_node;
}
/*
* Remove the given character from a TernaryTrieInnerNode's subtree. The
* Remove the given character from a TrieInnerNode's subtree. The
* function assumes the character is indeed in the subtree.
*/
void ttinode_remove(TernaryTrieInnerNode *node, const char c) {
TernaryTrieInnerNode **to_remove_ptr = &node;
void ttinode_remove(TrieInnerNode *node, const char c) {
TrieInnerNode **to_remove_ptr = &node;
// We use pointers to pointers here so we can later free the removed node
// without having to know what its parent is
@ -253,11 +250,11 @@ void ttinode_remove(TernaryTrieInnerNode *node, const char c) {
// If the node isn't a leaf, we have to replace it with another
if ((*to_remove_ptr)->left != NULL || (*to_remove_ptr)->right != NULL) {
TernaryTrieInnerNode *to_replace = *to_remove_ptr;
TrieInnerNode *to_replace = *to_remove_ptr;
// Replace with its only right child
if (to_replace->left == NULL) {
TernaryTrieInnerNode *to_remove = to_replace->right;
TrieInnerNode *to_remove = to_replace->right;
to_replace->key = to_remove->key;
to_replace->next = to_remove->next;
@ -268,7 +265,7 @@ void ttinode_remove(TernaryTrieInnerNode *node, const char c) {
}
// Replace with its only left child
else if (to_replace->right == NULL) {
TernaryTrieInnerNode *to_remove = to_replace->left;
TrieInnerNode *to_remove = to_replace->left;
to_replace->key = to_remove->key;
to_replace->next = to_remove->next;
@ -279,8 +276,8 @@ void ttinode_remove(TernaryTrieInnerNode *node, const char c) {
}
// Node has two children, so replace with successor
else {
TernaryTrieInnerNode *to_remove_parent = to_replace;
TernaryTrieInnerNode *to_remove = to_replace->right;
TrieInnerNode *to_remove_parent = to_replace;
TrieInnerNode *to_remove = to_replace->right;
while (to_remove->left != NULL) {
to_remove_parent = to_remove;
@ -307,14 +304,14 @@ void ttinode_remove(TernaryTrieInnerNode *node, const char c) {
}
/**
* Remove the given character from a TernaryTrieNode, respecting the rules
* Remove the given character from a TrieNode, respecting the rules
* of a binary search tree. This function assumes the character is in the search
* tree.
*
* @param node node to remove character from
* @param c character to remove
*/
void ttnode_remove(TernaryTrieNode *node, const char c) {
void ttnode_remove(TrieNode *node, const char c) {
ttinode_remove(node->ptr.root, c);
node->size--;