Google Firestore storage adapter for grammY
Storage adapter that can be used to store your session data on Google Cloud Firestore when using sessions.
Installation
npm install @grammyjs/storage-firestore --save
You should also have @google-cloud/firestore
installed.
Introduction
Set up your Firestore database by following https://cloud.google.com/firestore/docs/quickstart-servers#node.js.
You should now have:
- A project ID for you google cloud project
- A keyfile containing the firestore authentication credentials
- A Telegram bot token
Put those values into the following example code:
import { Bot, Context, session, SessionFlavor } from "grammy";
import { adapter } from "@grammyjs/storage-firestore";
import { Firestore } from "@google-cloud/firestore";
interface SessionData {
counter: number;
}
type MyContext = Context & SessionFlavor<SessionData>;
const db = new Firestore({
projectId: "cloudconvert-bot-257814",
keyFilename: "firestore-keyfile.json",
});
const bot = new Bot<MyContext>("797115886:AAEsOSRAqzpWBpQocm-_VcfT6S_HgRslf_g");
bot.use(
session({
initial: () => ({ counter: 0 }),
storage: adapter(db.collection("test-sessions")),
})
);
bot.command("stats", (ctx) =>
ctx.reply(`Already got ${ctx.session.counter} photos!`)
);
bot.on(":photo", (ctx) => ctx.session.counter++);
bot.start();
On firestore, you are billed per operation.
If your bot does not need the session data for most of the messages it processes, this would cause a lot of superfluous reads and writes.
Imagine your bot is in a group chat where it only counts photos but otherwise ignores all messages.
Then it would not makes sense to retrieve the session data for every request and write it back.
The solution is to use the built-in lazy mode of grammY, so just replace bot.use(session ...)
by bot.use(lazySession ...)
, and remember to await ctx.session
whenever you use it.
You must also flavor your context with LazySessionFlavor
instead of just SessionFlavor
.