Skip to content
Related Articles

Related Articles

Improve Article
Auto-complete feature using Trie
  • Difficulty Level : Medium
  • Last Updated : 17 May, 2021

We are given a Trie with a set of strings stored in it. Now the user types in a prefix of his search query, we need to give him all recommendations to auto-complete his query based on the strings stored in the Trie. We assume that the Trie stores past searches by the users.
For example if the Trie store {“abc”, “abcd”, “aa”, “abbbaba”} and the User types in “ab” then he must be shown {“abc”, “abcd”, “abbbaba”}.
Prerequisite Trie Search and Insert.
 

Given a query prefix, we search for all words having this query. 
 

  1. Search for the given query using the standard Trie search algorithm.
  2. If the query prefix itself is not present, return -1 to indicate the same.
  3. If the query is present and is the end of a word in Trie, print query. This can quickly be checked by seeing if the last matching node has isEndWord flag set. We use this flag in Trie to mark the end of word nodes for purpose of searching.
  4. If the last matching node of the query has no children, return.
  5. Else recursively print all nodes under a subtree of last matching node.

Below are few implementations of the above steps.
 

C++




// C++ program to demonstrate auto-complete feature
// using Trie data structure.
#include<bits/stdc++.h>
using namespace std;
 
// Alphabet size (# of symbols)
#define ALPHABET_SIZE (26)
 
// Converts key current character into index
// use only 'a' through 'z' and lower case
#define CHAR_TO_INDEX(c) ((int)c - (int)'a')
 
// trie node
struct TrieNode
{
    struct TrieNode *children[ALPHABET_SIZE];
 
    // isWordEnd is true if the node represents
    // end of a word
    bool isWordEnd;
};
 
// Returns new trie node (initialized to NULLs)
struct TrieNode *getNode(void)
{
    struct TrieNode *pNode = new TrieNode;
    pNode->isWordEnd = false;
 
    for (int i = 0; i < ALPHABET_SIZE; i++)
        pNode->children[i] = NULL;
 
    return pNode;
}
 
// If not present, inserts key into trie.  If the
// key is prefix of trie node, just marks leaf node
void insert(struct TrieNode *root,  const string key)
{
    struct TrieNode *pCrawl = root;
 
    for (int level = 0; level < key.length(); level++)
    {
        int index = CHAR_TO_INDEX(key[level]);
        if (!pCrawl->children[index])
            pCrawl->children[index] = getNode();
 
        pCrawl = pCrawl->children[index];
    }
 
    // mark last node as leaf
    pCrawl->isWordEnd = true;
}
 
// Returns true if key presents in trie, else false
bool search(struct TrieNode *root, const string key)
{
    int length = key.length();
    struct TrieNode *pCrawl = root;
    for (int level = 0; level < length; level++)
    {
        int index = CHAR_TO_INDEX(key[level]);
 
        if (!pCrawl->children[index])
            return false;
 
        pCrawl = pCrawl->children[index];
    }
 
    return (pCrawl != NULL && pCrawl->isWordEnd);
}
 
// Returns 0 if current node has a child
// If all children are NULL, return 1.
bool isLastNode(struct TrieNode* root)
{
    for (int i = 0; i < ALPHABET_SIZE; i++)
        if (root->children[i])
            return 0;
    return 1;
}
 
// Recursive function to print auto-suggestions for given
// node.
void suggestionsRec(struct TrieNode* root, string currPrefix)
{
    // found a string in Trie with the given prefix
    if (root->isWordEnd)
    {
        cout << currPrefix;
        cout << endl;
    }
 
    // All children struct node pointers are NULL
    if (isLastNode(root))
        return;
 
    for (int i = 0; i < ALPHABET_SIZE; i++)
    {
        if (root->children[i])
        {
            // append current character to currPrefix string
            currPrefix.push_back(97 + i);
 
            // recur over the rest
            suggestionsRec(root->children[i], currPrefix);
            // remove last character
            currPrefix.pop_back();
        }
    }
}
 
// print suggestions for given query prefix.
int printAutoSuggestions(TrieNode* root, const string query)
{
    struct TrieNode* pCrawl = root;
 
    // Check if prefix is present and find the
    // the node (of last level) with last character
    // of given string.
    int level;
    int n = query.length();
    for (level = 0; level < n; level++)
    {
        int index = CHAR_TO_INDEX(query[level]);
 
        // no string in the Trie has this prefix
        if (!pCrawl->children[index])
            return 0;
 
        pCrawl = pCrawl->children[index];
    }
 
    // If prefix is present as a word.
    bool isWord = (pCrawl->isWordEnd == true);
 
    // If prefix is last node of tree (has no
    // children)
    bool isLast = isLastNode(pCrawl);
 
    // If prefix is present as a word, but
    // there is no subtree below the last
    // matching node.
    if (isWord && isLast)
    {
        cout << query << endl;
        return -1;
    }
 
    // If there are are nodes below last
    // matching character.
    if (!isLast)
    {
        string prefix = query;
        suggestionsRec(pCrawl, prefix);
        return 1;
    }
}
 
// Driver Code
int main()
{
    struct TrieNode* root = getNode();
    insert(root, "hello");
    insert(root, "dog");
    insert(root, "hell");
    insert(root, "cat");
    insert(root, "a");
    insert(root, "hel");
    insert(root, "help");
    insert(root, "helps");
    insert(root, "helping");
    int comp = printAutoSuggestions(root, "hel");
 
    if (comp == -1)
        cout << "No other strings found with this prefix\n";
 
    else if (comp == 0)
        cout << "No string found with this prefix\n";
 
    return 0;
}

Python3




# Python3 program to demonstrate auto-complete
# feature using Trie data structure.
# Note: This is a basic implementation of Trie
# and not the most optimized one.
class TrieNode():
    def __init__(self):
         
        # Initialising one node for trie
        self.children = {}
        self.last = False
 
class Trie():
    def __init__(self):
         
        # Initialising the trie structure.
        self.root = TrieNode()
        self.word_list = []
 
    def formTrie(self, keys):
         
        # Forms a trie structure with the given set of strings
        # if it does not exists already else it merges the key
        # into it by extending the structure as required
        for key in keys:
            self.insert(key) # inserting one key to the trie.
 
    def insert(self, key):
         
        # Inserts a key into trie if it does not exist already.
        # And if the key is a prefix of the trie node, just
        # marks it as leaf node.
        node = self.root
 
        for a in list(key):
            if not node.children.get(a):
                node.children[a] = TrieNode()
 
            node = node.children[a]
 
        node.last = True
 
    def search(self, key):
         
        # Searches the given key in trie for a full match
        # and returns True on success else returns False.
        node = self.root
        found = True
 
        for a in list(key):
            if not node.children.get(a):
                found = False
                break
 
            node = node.children[a]
 
        return node and node.last and found
 
    def suggestionsRec(self, node, word):
         
        # Method to recursively traverse the trie
        # and return a whole word.
        if node.last:
            self.word_list.append(word)
 
        for a,n in node.children.items():
            self.suggestionsRec(n, word + a)
 
    def printAutoSuggestions(self, key):
         
        # Returns all the words in the trie whose common
        # prefix is the given key thus listing out all
        # the suggestions for autocomplete.
        node = self.root
        not_found = False
        temp_word = ''
 
        for a in list(key):
            if not node.children.get(a):
                not_found = True
                break
 
            temp_word += a
            node = node.children[a]
 
        if not_found:
            return 0
        elif node.last and not node.children:
            return -1
 
        self.suggestionsRec(node, temp_word)
 
        for s in self.word_list:
            print(s)
        return 1
 
# Driver Code
keys = ["hello", "dog", "hell", "cat", "a",
        "hel", "help", "helps", "helping"] # keys to form the trie structure.
key = "hel" # key for autocomplete suggestions.
status = ["Not found", "Found"]
 
# creating trie object
t = Trie()
 
# creating the trie structure with the
# given set of strings.
t.formTrie(keys)
 
# autocompleting the given key using
# our trie structure.
comp = t.printAutoSuggestions(key)
 
if comp == -1:
    print("No other strings found with this prefix\n")
elif comp == 0:
    print("No string found with this prefix\n")
 
# This code is contributed by amurdia

Java




import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
 
public class Trie {
 
    public class TrieNode {
        Map<Character, TrieNode> children;
        char c;
        boolean isWord;
 
        public TrieNode(char c) {
            this.c = c;
            children = new HashMap<>();
        }
 
        public TrieNode() {
            children = new HashMap<>();
        }
 
        public void insert(String word) {
            if (word == null || word.isEmpty())
                return;
            char firstChar = word.charAt(0);
            TrieNode child = children.get(firstChar);
            if (child == null) {
                child = new TrieNode(firstChar);
                children.put(firstChar, child);
            }
 
            if (word.length() > 1)
                child.insert(word.substring(1));
            else
                child.isWord = true;
        }
 
    }
 
    TrieNode root;
 
    public Trie(List<String> words) {
        root = new TrieNode();
        for (String word : words)
            root.insert(word);
 
    }
 
    public boolean find(String prefix, boolean exact) {
        TrieNode lastNode = root;
        for (char c : prefix.toCharArray()) {
            lastNode = lastNode.children.get(c);
            if (lastNode == null)
                return false;
        }
        return !exact || lastNode.isWord;
    }
 
    public boolean find(String prefix) {
        return find(prefix, false);
    }
 
    public void suggestHelper(TrieNode root, List<String> list, StringBuffer curr) {
        if (root.isWord) {
            list.add(curr.toString());
        }
 
        if (root.children == null || root.children.isEmpty())
            return;
 
        for (TrieNode child : root.children.values()) {
            suggestHelper(child, list, curr.append(child.c));
            curr.setLength(curr.length() - 1);
        }
    }
 
    public List<String> suggest(String prefix) {
        List<String> list = new ArrayList<>();
        TrieNode lastNode = root;
        StringBuffer curr = new StringBuffer();
        for (char c : prefix.toCharArray()) {
            lastNode = lastNode.children.get(c);
            if (lastNode == null)
                return list;
            curr.append(c);
        }
        suggestHelper(lastNode, list, curr);
        return list;
    }
   
 
    public static void main(String[] args) {
        List<String> words = List.of("hello", "dog", "hell", "cat", "a", "hel","help","helps","helping");
        Trie trie = new Trie(words);
     
        System.out.println(trie.suggest("hel"));
    }
 
}
Output
hel
hell
hello
help
helping
helps

How can we improve this? 
The number of matches might just be too large so we have to be selective while displaying them. We can restrict ourselves to display only the relevant results. By relevant, we can consider the past search history and show only the most searched matching strings as relevant results. 
Store another value for the each node where isleaf=True which contains the number of hits for that query search. For example if “hat” is searched 10 times, then we store this 10 at the last node for “hat”. Now when we want to show the recommendations, we display the top k matches with the highest hits. Try to implement this on your own.
This article is contributed by Hemang Sarkar. If you like GeeksforGeeks and would like to contribute, you can also write an article using write.geeksforgeeks.org or mail your article to contribute@geeksforgeeks.org. See your article appearing on the GeeksforGeeks main page and help other Geeks.
Please write comments if you find anything incorrect, or you want to share more information about the topic discussed above.
 

Attention reader! Don’t stop learning now. Get hold of all the important DSA concepts with the DSA Self Paced Course at a student-friendly price and become industry ready.  To complete your preparation from learning a language to DS Algo and many more,  please refer Complete Interview Preparation Course.

In case you wish to attend live classes with industry experts, please refer DSA Live Classes




My Personal Notes arrow_drop_up
Recommended Articles
Page :