Going through the docs, I encountered:
...you can call functions directly with an HTTP request or a call from the client.
~ source
there (link in the quote) is a mention about functions.https.onCall.
But in the tutorial here, another function functions.https.onRequest is used, so which one should I use and why? What is the difference/similarity between them?
Documentation for functions.https is here.
The official documentation for these concepts is quite helpful, but from the view of an amateur, the described differences were confusing at first.
Both types, when deployed, are assigned with a unique HTTPS endpoint URL and can be accessed directly using an https client.
However, there is one important difference in the way how they are intended to be called.
onCall: from the client's firebase.functions()onRequest: via standard https client (e.g. fetch() API in JS)Can be invoked directly from the client app, which is its primary purpose.
functions.httpsCallable('getUser')({uid})
.then(r => console.log(r.data.email))
It is implemented using user-provided data and automatic context.
export const getUser = functions.https.onCall((data, context) => {
if (!context.auth) return {status: 'error', code: 401, message: 'Not signed in'}
return new Promise((resolve, reject) => {
// find a user by data.uid and return the result
resolve(user);
})
})
The context automatically contains metadata about the request such as uid and token.
Input data and response objects are automatically (de)serialized.
Serves mostly as an Express API endpoint.
It is implemented with express Request and Response objects.
export const getUser = functions.https.onRequest((req, res) => {
// verify user from req.headers.authorization etc.
res.status(401).send('Authentication required.');
// if authorized
res.setHeader('Content-Type', 'application/json');
res.send(JSON.stringify(user));
})
Depends on user-provided authorization headers.
You are responsible for input and response data.
Read more here Is the new Firebase Cloud Functions https.onCall trigger better?
fetch(), it's just an API endpoint.
onCall is callable and onRequest is not because as you mentioned onCall's context has info about uid but onRequest doesn't.The main difference between onCall and onRequest for the client is the way they are invoked from client side. When you define a function using onCall e.g.
exports.addMessage = functions.https.onCall((data, context) => {
// ...
return ...
});
you can invoke it on the client side using the firebase function client SDK e.g.
// on the client side, you need to import functions client lib
// then you invoke it like this:
const addMessage = firebase.functions().httpsCallable('addMessage');
addMessage({ text: messageText })
.then((result) => {
// Read result of the Cloud Function.
});
more info for onCall: https://firebase.google.com/docs/functions/callable
But if you define your function using onRequest e.g.
exports.addMesssage = functions.https.onRequest((req, res) {
//...
res.send(...);
}
you can call it using normal JS fetch API (no need to import firebase functions client lib on the client code) e.g.
fetch('<your cloud function endpoint>/addMessage').then(...)
this is the big difference that you need to consider when deciding on how to define your functions on the server.
more info for onRequest: https://firebase.google.com/docs/functions/http-events
One caveat I ran into was that the data parameter in functions.https.onCall is actually the raw request. To access the payload, we have to access the data field of the input.
Example
Client sends payload to cloud function:
const callFunction = httpsCallable(functions, 'addMessage');
const result = await callFunction({payload: 'payload'});
Access payload in cloud function:
exports.addMessage = functions.https.onCall((rawRequest, context) => {
const payload = rawRequest.data.payload;
return ...
});
I couldn't find this in the docs, please let me know if I'm missing something.
