ruhuka
v0.0.2
Published
A small REST API Client
Downloads
1
Readme
ruhuka
A REST API Client.
In these days, I couldn't find a viable REST API Client for Javascript so I created one. This is in it's initial days and may change very quickly. Inspired by another-rest-client
Installation
Using npm
npm install ruhuka
# or using yarn
yarn add ruhuka
Then import it
// commonjs
const { RESTResource, Resource } = require("ruhuka");
// esm
import RESTResource, { Resource } from "ruhuka";
Using deno
import RESTResource, { Resource } from "https://deno.land/x/[email protected]/mod.ts";
Usage
The library exports 2 main classes: RESTResource
and Resource
.
RESTResource
is also the default export. It has some helper methods to help query
REST APIs.
REST Resource
Importing REST Resource.
// import the rest resource class
import RESTResource from "ruhuka";
import { RESTResource } from "ruhuka";
Initialization
// initialise a new resource
const Posts = new RESTResource(
"todos",
"https://jsonplaceholder.typicode.com/posts",
);
NOTE: all returned values are
Response
s by default, so you will have to parse the response yourself for example by callingposts.json()
CRUD
The CRUD (Create Read Update Delete) methods are only available on a
RESTResource
(not the Resource
).
- Get all items in collection (READ)
const posts = await Posts.get();
// GET https://jsonplaceholder.typicode.com/posts
- Get a particular resource
const post = await Posts.get_item("1");
// GET https://jsonplaceholder.typicode.com/posts/1
- Creating a post
await Posts.post({
userId: 1,
id: 101,
title: "Test post",
body: "easy peasy",
}, {
encode: "json",
});
// POST https://jsonplaceholder.typicode.com/posts
// Headers content-type=json
// Body=JSON.stringify(body)
- Updating a resource
// using PUT
await Posts.put_item("1", {
title: "This is changed",
}, {
encode: "json",
});
// PUT https://jsonplaceholder.typicode.com/posts/1
// Headers content-type=json
// Body=JSON.stringify(body)
// using PATCH
await Posts.put_item("1", {
title: "This is changed",
}, {
encode: "json",
});
// PATCH https://jsonplaceholder.typicode.com/posts/1
// Headers content-type=json
// Body=JSON.stringify(body)
- Deleting a resource
await Posts.delete_item("1");
Resource
Importing Resource.
// import the resource class
import { Resource } from "ruhuka";
Initialization
// initialise a new resource
const Posts = new Resource(
"todos",
"https://jsonplaceholder.typicode.com/posts",
);
Because a Resource does not provide helper methods such as get
, get_item
etc. You have to initialise them yourself. Here is an example roughly based on
Ruby on Rails.
Posts.create = Posts.new_method("create", "POST");
Posts.all = Posts.new_method("all", "GET");
// the third parameter means it's acting on a document rather than a collection in whole
Posts.get = Posts.new_method("get", "GET", true);
Posts.update = Posts.new_method("update", "PUT", true);
Posts.delete = Posts.new_method("delete", "DELETE", true);
// Now we can use our methods
Posts.delete("1");
TIP: You can also use
new_method
on plainResource
. In fact, RESTResource is an extended version of Resource that callsnew_method
several times in the constructor to make it able to interact to a REST API.
Configuration
While initialising the resource, you can also set some options as outlined below:
- route: options for the child routes (
get
,get_item
) etc.
const ConfigedResource = new Resource("name", "uri://url", {
route: { ...options },
});
Route Configuration
You can set route options per resource or per API call as shown below:
// per resource
// the third paramter
const JSONAPI = new Resource("name", "uri://url", {
route: {
encode: "json",
decode: "json",
headers: {
"content-type": "application/json",
"authorization": "JWT somekey",
},
},
});
// per call
// the first parameter on collection methods
const result = JSONAPI.get({
decode: "text",
headers: {
"etag": "W\\some-etag",
},
});
// the second parameter on document methods
const result = JSONAPI.get_item("1", {
decode: "text",
headers: {
"etag": "W\\some-etag",
},
});
- encode:
Function | "json" | null
Encode the request body before sending it. Can be a function that receives a body and encode it to one of the following types:string | Blob | BufferSource | FormData | URLSearchParams | ReadableStream<Uint8Array>
. If it is the string"json"
.JSON.stringify
will be called on the body before sending it. By default the value is set to null and the body isn't encoded at all. - decode:
Function | "response" | "json" | "text" | "array-buffer" | null
. Converts the got Response to the given type. By default it is set to"response"
and therefore returns the plainResponse
. - fetch:
Function
thefetch
function to use. By default usesglobalThis.fetch
. - additionalData:
object
additional data to send in the 2nd argument tofetch
. - params:
string | string[][] | Record<string, string> | URLSearchParams | undefined
Query search params to send (first argument toURLSearchParams
). - mustBeOk:
boolean
the response must be in the 200-399 range (successful).
Usage examples
Sample JSON API.
import { Resource, RESTResource } from "ruhuka";
const Posts = new RESTResource(
"todos",
"https://jsonplaceholder.typicode.com/posts",
{
route: {
encode: "json",
decode: "json",
},
},
);
Experimental nested documents
Gets resources attached to another resource.
const Comments = Posts.document("1", "comments");
// get all comments
await Comments.get();
// get a specific comment
await Comments.get_item("1");
NOTE: The nested resources are cached in case of multiple calls. The following code hence creates a single resource.
const allComments = await Posts.document("1", "comments").get();
const specificComment = await Posts.document("1", "comments").get_item("1");
await Posts.document("1", "comments").delete_item(specificComment.id);
TypeScript
You can add 2 parameters to Resource
to denote the type of the (decoded) data
returned from the document and collection APIs respectively
interface IPost {
title: string;
body: string;
}
interface IPaginated {
total: number;
limit: number;
skip: number;
data: IPost[];
}
const Posts = new Resource<IPost, IPaginated>("posts", "uri://url", {
route: {
encode: "json",
decode: "json",
},
});
const posts: IPaginated = await Posts.get();
const post: IPost = await Posts.get("1");
Error handling
The methods return promises which can catch when an issue occur (request was
aborted, network lost) or when the result doesn't have a status code (200-399 or
response.ok === false
) and the route configuration option mustBeOk
is set.
The package throws the original response.
const Posts = new Resource("posts", "uri://url", {
route: {
encode: "json",
decode: "json",
mustBeOk: true,
},
});
const post = await Posts.get("400042342")
.catch(response => console.log(response.status));
// logs 404 (there is no post with id 400042342)
Caching
The package doesn't do any caching at all. But I will work on a caching fetching function. Here is an idea of how you might implement cache.
// stores a map of urls and response
const cache = new Map();
// this implementation is very incomplete. It caches everything. In practice you would want to honor the cache headers `Cache-Control`, `ETag` etc. and would also compare the request options (changed headers may mean a different response.)
const cachingFetch = (url, options) => {
// if the url is cached, return the cached response
if (cache.has(url)) {
return cache.get(url).clone();
};
// else fetch the url then cache it
else {
return fetch(url, options)
.then(response => {
// don't cache non-ok responses
if (!response.ok) return response;
cache.set(url, response);
return response.clone();
});
}
}
const Posts = new Resource("posts", "uri://url", {
route: {
encode: "json",
decode: "json",
mustBeOk: true,
fetch: cachingFetch
},
});
Authentication
You may be required to send some headers everytime to access protected resources. It's super easy with ruhuka:
// could also be a plain object
const headers = new Headers();
headers.set("Authorization", `JWT YOUR_KEY_HERE`);
const Posts = new Resource("posts", "uri://url", {
route: {
encode: "json",
decode: "json",
headers
},
});
Contributing
Note that this is an early draft and some features are missing. However all types of pull requests are welcome.