Spell Checking And Auto Completion REST API

In this project. I have implemented a REST API for the spell cheking and for word auto completion. This repository is created by Stephane Miguel. For any suggestion or questions, please feel free to contact me at Skakanakou@gmail.com

Implementation Details

Data

In this project, I have used a text file that contains 479k English words. Here is the link to the github repository that contains the data english-words. I have used the alphabetical version words_alpha.txt

Data Structure

I have implemented 26-Way Trie to store efficiently the words and also to retrieve them efficiently. Here is the link to my implementation 26-Way Trie

REST API

For the REST API implementation, I have used Jersey and Jackson in order to provide JSON format response.

Endpoints of the REST API

The REST API has the following endpoints :

Spelling Checking

The URI is : /restapi/spellingCheck/{word}
This endpoint help to check whether a word is well spelled or not. When the given word is not well spelled, the response suggests 10 words. The response is formated as follow :

• when the query word is "spelling" , the response is :

{
  "queryString": "spelling",
  "wellSpelled": true,
  "suggestion": []
}

• when the query word is "algorithmm", the response is :

{
  "queryString": "algorithmm",
  "wellSpelled": false,
  "suggestion": [
    "algorithm",
    "algorithmic",
    "algorithmically",
    "algorithms"
  ]
}

Word Auto Completion

The URI is : /restapi/autoCompletion/{prefix}
Given a prefix, this endpoint return in alphabetical order at most 20 words that have the given prefix. The response is formated as follow :

• When the given prefix is "sort" , the response is :

{
  "prefix": "sort",
  "propositions": [
    "sort",
    "sortable",
    "sortably",
    "sortal",
    "sortance",
    "sortation",
    "sorted",
    "sorter",
    "sorters",
    "sortes",
    "sorts",
    "sortwith",
    "sorty"
  ]
}

Longest Prefix Word

The URI is : /restapi/longestPrefixWord/{queryString}
Given a queryString, this endpoint return the word that is the longest prefix of the given query sring. The response is formated as follow :

• When the given query string is "computationalblem" , the response is :

{
  "queryString": "computationalblem",
  "longestPrefixWord": "computational"
}