inverted-index-api

An Inverted Index Api for implementing efficient search functionality for software applications.

Introduction

Inverted Index API is a NodeJS Powered implementation of the inverted index data structure for finding what word is in what 'book'. A 'book' is an entry in a document structured as a JSON array. Here is a sample document:

Dependencies

Backend Dependencies

This app's functionality depends on multiple NodeJS packages

• NodeJS This is a JavaScript runtime built on Chrome's V8 JavaScript engine. It is used for installing and managing the dependencies.
• Express This is used to create the web routes/endpoints.
• Body-Parser This is used for parsing the content of forms sent to the web app.
• dotenv This handles the management and dynamic assignment of environmental variables
• body-parser This parses the request into a body that can be accessed.
• multer This handles file uploads to the app

Installation guide

To run the app,

• Clone this repository
• Navigate your git terminal to the root of this project
• Run npm install at the root of the project folder to install dependencies
• Run gulp serve to run the app
• Connect to port 8080 on your local host

Usage

With POSTMAN

Creating Index

• Set the address to `localhost:8080/api/create and Method to POST
• Choose form-data in the Body category of the type of data to send
• Set a key called files and set it's type to file
• Ensure there are no headers set, delete any that exists
• Click on Choose Files to select files from your local storage (Files must meet the specs stated earlier)
• Click on send to get and index object for the content of the files uploaded.

Searching Index

• Copy the index created by the api/create route
• Create a new tab on Postman
• Set the address to `localhost:8080/api/search and method to POST
• Choose x-www-form-urlencoded in the body category of the type of data to send
• Set a key of fileName, and for the value, put in the copied index
• Set a key of searchTerms, and set it to the terms to search for
• Set an optional key of fileName, and set it to the name of the file whose indexed content you want to search
• Click send to get a response of each term you searched, and an array of numbers indicating what book they appear in

Limitations

• The Search function for querying multiple files is yet to be implemented. Searching can only be performed on one file.

Sample Index

• Suppose we have a file containing this below:

[
    {
      'title': 'Queen Premier',
      'text': 'Is he to go up? Yes, he is to go up.'
    },
    {
      'title': 'Poem',
      'text': 'Broken crayons still colour.'
    }
]

• With file name validFile.json, the resulting created index will be:

{
  "validFile.json": {
    "go": [
      0
    ],
    "he": [
      0
    ],
    "is": [
      0
    ],
    "premier": [
      0
    ],
    "queen": [
      0
    ],
    "to": [
      0
    ],
    "up": [
      0
    ],
    "yes": [
      0
    ],
    "broken": [
      1
    ],
    "colour": [
      1
    ],
    "crayons": [
      1
    ],
    "poem": [
      1
    ],
    "still": [
      1
    ]
  }

• If the index is queried, for crayons, poem and colour, the result would be

{
  "validFile.json": {
    "crayons": [ 
        1 
      ],
    "poem": [ 
        1 
      ],
    "colour": [ 
        1 
      ]
  }
}

Connecting on Heroku

• Navigate to this link and follow the steps stated above.

Contributing

Contributions are welcome and appreciated. To contribute

• Fork this repository and peruse here
• Submit a pull request if you feel you have substantial updates :)
• I look forward to your Pull Requests!