This document is mostly aimed towards people who are interested in contributing but feel intimidated by the project and the technologies used in it or want to find out how things work.
There are 3 main technologies Hifumi uses that you might want to become more familiar with before starting out.
If you are already familiar with these, skip over to Contributing
Docker is a technology that wraps programs into standalone packages called containers. Docker is used in this project for running Hasura and PostgreSQL separately to make sure the programmer doesn't have to bother with setting them up individually which would be very tedious.
You can download docker and docker-compose here.
Once you have them downloaded you can setup both the dependencies by typing docker-compose up at the root of the project. That's really all you need to worry about when it comes to docker.
If you want to stop the containers without losing any of the data, just do docker-compose stop
Read up on docker more here
We use it for almost every part of Hifumi in production
Graphql is a query language that aims to be a statically typed alternative to REST. Unlike REST, it's made up of a single endpoint where you send the structure of the request as the body of a POST request.
GraphQL is used in hifumi to interface with the database.
An example query that you might use with Hifumi's API might look something like this
query {
images(limit: 2) {
user {
user_id
name
}
url
image_tags {
name
}
}
}What this does is simply retrieve the first 2 images from the database, but unlike REST where you don't get to specify what kind of data you want to get back from the server, GraphQL lets you configure the response to exactly the way you want it.
A response you might get from this query could be something like this
{
"data": {
"images": [{
"user": {
"user_id": "136663446337880064",
"name": "Nemu Nemu the Sleep Queen"
},
"url": "https://cdn.discordapp.com/attachments/231790205676290049/565545316950999050/image0.png",
"image_tags": [{
"tagger_id": "136663446337880064",
"name": "relatable"
}]
},
{
"user": {
"user_id": "185800739090726913",
"name": "Jordan"
},
"url": "https://cdn.discordapp.com/attachments/231790205676290049/565554323468058624/image0.jpg",
"image_tags": [{
"tagger_id": "185800739090726913",
"name": "cute"
}]
}
]
}
}As you can see, the response fields match the requests made which is very beneficial since there won't be any unexpected responses.
In Hifumi's codebase, however, you might come across something along the lines of
const { data } = await _client.query({
image_tags_aggregate: [{
where: {
name: {
_eq: query
}
}
}, {
aggregate: {
count: 1,
}
}]
});Which might seem odd at first. It's actually this repo which provides a typesafe way of creating graphql queries. It's the exact same thing as this query
query {
image_tags_aggregate(
where: {
name: {
_eq: "whatever query is"
}
}
) {
aggregate {
count
}
}
}To use you have to have the graphql server set up and run npm run codegen which will pull all the type information from Hasura and generate typescript types from it.
Learn more about graphql here
Hasura is one of the best open source projects out there! It allows us to interface with the PostgreSQL database using graphql and prevents us from having to write a separate API to be able to do that.
It's not necessarily useful for the bot itself, but this is the secret ingredient that allows Hifumi's website to receive live updates from the database with no extra work.
In order to have the bot working properly, you're going to need to have Hasura and PostgreSQL running
Once you have the bot running, you can go check out its console at http://localhost:8080 which will let you query the database as you wish.
The production URL for the hasura endpoint is https://db.hifumi.io/v1alpha1/graphql
The development URL is http://localhost:8080/v1alpha1/graphql
Before you push a commit or make a pull request, make sure that:
- Your code compiles (duh)
- You have no linter errors from tslint
You can check for linter errors by using npm run lint