Homework assessment for Favs

This project will assume you have Node.js, PostgreSQL setup on your computer previously. We will need at least Node 18 for the project to run properly.

1. Clone the repository:

git clone https://github.com/m1771vw/favs-hw.git
cd favs-hw

2. Use npm to install project dependencies listed in package.json:

npm install

3. Create an .env file. Usually we won't keep this here, but for demonstration purposes will give you the .env example. Please replace username and password with you Postgres information

DATABASE_URL="postgresql://<username>:<password>@localhost:5432/favsdb?schema=public"

4. We're going to assume you have Postgres set up with default options. There'll be a script to be run to create a database and a user for that database.

If your postgres user isn't postgres, please update the -U tag

psql -U postgres -d postgres -h localhost -p 5432 -f ./scripts/create_database.sql

5. We need to use Prisma to set up your database.

npx prisma migrate dev --name first_migration

6. Now we can start the project

npm run start

7. You can now navigate to here to see the GraphQL Interface. It's mapped to localhost:3003 so if 3003 is being used please change it to something else.

Example queries and mutations can be found in scripts/graphql_scripts.js.

Here are some examples from that file

mutation {
  createUser(data: {
    first_name: "User1FirstName",
    last_name: "User1LastName",
    phone_number: "1234567890",  
    username: "user1"
  }) {
    id
    first_name
    last_name
    phone_number
    username
  }
}

mutation {
  createUser(data: {
    first_name: "User2FirstName",
    last_name: "User2LastName",
    phone_number: "9876543210",
    username: "user2"
  }) {
    id
    first_name
    last_name
    phone_number
    username
  }
}

query {
  users {
    id
    first_name
    last_name
    phone_number
    username
    created_at
    updated_at
  }
}

query {
  user(id: "<user_id>") {
    id
    first_name
    last_name
    phone_number
    username
    created_at
    updated_at
  }
}

Here are some for Friend Request. This will require you to get the user id from the created users earlier.

mutation {
  createFriendRequest(data: {
    user_requestor_id: "<input user id here>",
    user_requested_id: "<input different id here>",
    status: "pending"
  }) {
    id
    user_requestor_id
    user_requested_id
    status
  }
}

You can accept the request like this. You can change response to REJECTED as well.

mutation {
  responseFriendRequest(
    friendRequestId: "<insert friend request id here>"  
    response: ACCEPTED  
  ) {
    id
    user_requestor_id
    user_requested_id
    status
    created_at
    updated_at
  }
}

You can check friendship of a user like this

query {
  friends(userId: "<input user id here>") {
    id
    first_name
    last_name
    phone_number
    username
    created_at
    updated_at
  }
}

• Design of friend list table
  • Picking single rows over a friend list array
    • Array would be simpler at the beginning, but can run into problems once friend list grows
    • Unable to add more data about it
  • Want to add some extra fields, like when they first became friends
• Keeping friend request table separate
  • Thought about having a status on friend list table but ultimately wanted to seaprate it so I can have a history and also see who sent friend request to whom
• Choosing GraphQL library
  • Picked Apollo because it’s pretty standard
  • Lots of documentation
  • Good for scaling , wide range of features
  • More familiar with the structure as well
• How to store friendships
  • Bidirectional or single row
  • Two entries:
    • Pro simple, easy to query
    • Con: Increased storage
  • Single entry:
    • Pro: Less storage
    • Con: Complex queries, could take more time
  • Going with bidirectional: Storage is cheap, speed is better