@aims-api/aims-tagging-node
v0.0.8
Published
A client package for AIMS Tagging API
Downloads
685
Readme
Getting Started
npm install @aims-api/aims-tagging-node
Authentication
In order to use the lirbary you need to obtain credentials by contacting us at [email protected]. Credentials consist of CLIENT_ID
, CLIENT_SECRET
, LOGIN
and PASSWORD
.
To access protected routes you need to authenticate in order to get a time-limited token. You need to use this token while creating a client instance.
Example with Next.js
It is common to make a proxy request from client app to the server "Next.js API route" in order to hide foreign URL.
You may store the token in a cookie or local storage (browser), but you need to retrieve it from the request object. Storing the token in cookies is recommended for security reasons. In case you decide to store it in cookies, you can set Axios interceptor to check if cookie is still valid and if not, terminate user session in order to obtain a new token.
// pages/api/authenticate.ts
import { NextApiRequest, NextApiResponse } from 'next'
import { Client as TaggingApiClient } from '@aims-api/aims-tagging-node'
export const createClient = async (req: NextApiRequest) => {
const { clientId, clientSecret } = await getSiteConfigForRequest(req)
const apiUserToken = getTokenFromCookie(req)
// You can retrieve auth_token from cookies or local storage
// Authentication endpoints (e.g. "authenticate", "register", ...)
// do not require "apiUserToken" field
return new TaggingApiClient({
apiHost: 'HOST_URL', // optional
clientId: 'YOUR_CLIENT_ID', // required
clientSecret: 'YOUR_CLIENT_SECRET', // required
apiUserToken: JSON.parse(req.cookies.auth_token),
})
}
const handler = async (req: NextApiRequest, res: NextApiResponse) => {
if (req.method === 'POST') {
try {
const { userEmail, userPassword } = req.body // credentials: LOGIN, PASSWORD
// passed from the client-side code
const aimsClient = createClient(req)
const response = await aimsClient.endpoints.authentication.authenticate({
userEmail,
userPassword,
})
const { token, ...rest } = response // removes token from the response to hide
// from the client-side code (optional)
return
res
.setHeader(
'Set-Cookie',
`auth_token=${JSON.stringify(response.token)}; HttpOnly; Secure; path=/`,
) // sets token to cookie (optional)
.status(200)
.json(rest)
} catch (error) {
console.warn(error)
}
}
return res.status(400).json('Method not allowed')
}
export default handler
Usage
When you create a client instance in your codebase, you can then easily access all the existing endpoints via IDE autocomplete, as well as the required and optional parameters.
It's recommended to create a new file for each request handler.
For calling an endpoint batch.startTagging
credits are required, for each track in a batch user needs to have exactly one credit. To get information about credits availability, please call apiUser.remainingMonthlyRequests
endpoint, property is named remainingMonthlyRequests
.
TypeScript
You can import input and response types that are provided in every endpoint file. In order to validate response structure and prevent application from crashing, you may use Zod library.
Example
// src/types/index.ts
...
type Track = Partial<{
id: string
title: string
tags: Tags | null
filesize: number | null
duration: number | null
processedAt: DateTime | null
hookUrl: string | null
modelVersion: string
queuedAt: DateTime | null
batchId: string | null
}>
Routes
The library provides a set of endpoints that can be found in src/client/index.ts file by the endpoints
property on line #175.
Response Structure
Responses are not validated by this library, therefore you need to parse the structure on your own, referring to provided response types. If you find any incosistency, please open an issue or contact us via email.
All endpoints excerpt userData
object that contains remainingRequests
property, the value is an integer (of type String) that represents the number of credits available for tagging. To access exact response data in the client-side code you need to read response.data.data
.
type ResponseType = {
data: any
userData: {
remainingRequests: string
}
}
This library uses Axios under the hood to make requests, so any network error will be of type AxiosError
. Please, check Axios documentation for more information.
Example of parsing an error with Zod on the client side:
// src/helpers/response.ts
import { z } from 'zod'
import { isAxiosError } from 'axios'
interface CustomError {
status: number | null
message: string
field?: string
}
const errorSchema = z.object({
status: z.number(),
message: z.string(),
field: z.optional(z.string()),
})
const notParsed: CustomError = {
status: 500,
message: 'Not able to parse error',
}
export const parseError = (error: unknown) => {
if (isAxiosError(error) && error.response !== undefined) {
const errorParserResponse = errorSchema.safeParse(error.response.data)
if (!errorParserResponse.success) {
return notParsed
}
return errorParserResponse.data
}
return notParsed
}
License
See LICENSE for more information.