reSearch

reSearch is a simple application that takes in multiple paragraphs(documents) of text, assigns a unique ID to each paragraph and stores the words to paragraph mappings on an inverted index.Given a word to search for, it lists out the top 10 paragraphs in which the word is present.

Note 1 : It is assumed that the user has Postman installed to test the api by sending HTTP requests and MongoDB Cluster to be used as Database.

Note 2 : It is also assumed that the user has Node (LTS version) , npm & nodemon installed globally.

API Specifications at a Glance

• Tokenizes documents to words by splitting at whitespace.
• Converts all words to lowercase.
• Indexes these words against the documents they are from.
• Indexes these words against the documents they are from.
• Generates a unique ID for every document that is indexed.

API Documentation :

# API Endpoint Locally : http://localhost:3000/

Headers
Content-Type : application/json

/research/index

• POST: Index a given document (After having split the input into paragraphs a.k.a document)
• Query :

{
    "documents" : string
}

• Response :

{
    "message": string
}

/research/search

• POST: Given a word, search for it and retrieve the top 10 paragraphs (Documents) that contain it.
• Query :

{
    "query" : string
}

• Response :

{
    "paragraphs": {
        "tags" : array,
        "_id" : string,
        "text" : string,
    }
}

/research/clear

• GET: Clear the index and all indexed documents.
• Response :

{
    "message": string
}