Task
Write an API that will allow quick search of data, as well as recording and editing of this data using Elasticsearch (ES) in conjunction with a database. Adding and editing of the data in database should influence the search and data structure in ES.
Setup
You can use docker container preset for ES, also you will need database for storing actual data (we recomend PostgreSQL, but you can use any database that you are comfortable with)
Given
Database mocks with format
[
{
id: number,
title: string,
content: string,
}
]
Lets call this structure a "document".
This are mocks for your database.
You can store this data in ES data-node in any format you want, try to find the best option.
Scenarios
Save
First of all, you need to be able to save information passed to database to ES. For that we need to have one route that is responsible for saving data.
As a result we should have route
POST /document
request: {
body: {
document: {
title: string, // title of the document to be posted
content: string, // content of the document to be posted
}
}
}
reply: {
document: {
id: number, // id of the posted document
title: string, // title of the posted document
content: string, // content of the posted document
}
}
Search
We need to have route that would search for the given substring across all the documents (full text search), but priority should be given to title field.
We should have this response structure:
{
id: number, // id of the document, where substring was found
fieldName: string, // name of the field where substring was found ('title' or 'content')
fieldContent: string, // content of the field where substring was found
}
e.g.
Given database structure
[
{
id: 1,
title: "Codex",
content: "We are international developers team"
},
{
id: 2,
title: "Our projects",
content: "Codex team developed several big projects like Editor.js, Hawk, Notex"
},
{
id: 3,
title: "Codex Workflow",
content: "Here you can see work progress of the codex team"
}
]
Given search request:
Search by word "Codex"
All three should be found, but priority would be given to document, whose title contain keyword. We should get this response structure:
[
{
id: 1,
fieldName: 'title',
fieldContent: 'Codex',
},
{
id: 3,
fieldName: 'title',
fieldContent: 'Codex Workflow',
},
{
id: 2,
fieldName: 'content',
fieldContent: 'Codex team developed several big projects like Editor.js, Hawk, Notex'
},
]
As a result we should have route
GET /search
request: {
querystring: string // text search substring
}
reply: {
documents: [
{
id: number // id of the document, where substring is found
fieldName: string // field name, where substring is found ('title' or 'content')
fieldContent: string // content of the field, where substring is found
}
]
}
Patch
Last, but not least is that we need to have ability to patch database fields by id. Сhange in the database must be accompanied by a change in data in the ES data-node so that the search occurs only according to current data.
As a result we should have route
PATCH /document
request: {
documentId: number, // id of the document to be patched
document: {
title: string, // new title of the document
content: string, // new content of the document
}
}
reply: {
document: {
id: number, // id of the patched document
title: string, // new document title
content: string, // new document content
}
}
Where to submit?
Publish you code on the GitHub and provide all required instructions for using it
Deadline
October, 16th. It's recommended to show your in-between progress and ask questions in a Telegram Channel.