BUZZ

This is a mini API for a simple dating app that fetches potential user profiles based on location using h3 geospatial index

All that's needed to setup and run this app is Golang and Docker installed locally.

---

Run Tests

go test -v ./...

Run API

docker compose up

Shutdown

docker compose down

---

A few users would be added to the database on startup, you can login with email [email protected] and password password123. It already has a few users located within a 100km distance.

APIs

On start up the app would run on http://localhost:8003

Create user:

This creates a random user with a default password of password123

URL: /user/create

Method: POST

Body: NONE

RESPONSE

{
	"result": {
		"id": 25,
		"email": "[email protected]",
		"password": "77546b6a714655717144705547696e414a71485acbfdac6008f9cab4083784cbd1874f76618d2a97",
		"name": "Asha Streich",
		"gender": "M",
		"age": 19,
		"lat": 0,
		"long": 0
	}
}

===================================

Login:

URL: /login

Method: POST

Body:

{
    "email": "[email protected]",
    "password": "password123"
}

RESPONSE

{
	"token": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJ1c2VySWQiOjh9.pJgdc2DQjtWvHA2rP5_WMzD2Z0Ldd4PPtWWw4ocfGkM"
}

===================================

Update Location:

URL: /user/update

HEADERS: Content-Type: application/json, Authorization: Bearer <Token>

Method: PUT

Body:

{
  "Latitude": 41.089097,
  "Longitude": -81.349288
}

RESPONSE

{
  "result": {
    "id": 8,
    "email": "[email protected]",
    "password": "77546b6a714655717144705547696e414a71485acbfdac6008f9cab4083784cbd1874f76618d2a97",
    "name": "Olga Goodwin",
    "gender": "F",
    "age": 0,
    "longitude": -81.349288,
    "latitude": 41.089097,
    "h3_index": 617744365936443391
  }
}

===================================

Swipe:

Use to respond to a user's profile (either YES or NO)

URL: /swipe

HEADERS: Content-Type: application/json, Authorization: Bearer <Token>

Method: POST

Body:

{
  "userId": 1,
  "action": "YES" // or NO
}

RESPONSE

{
  "results": {
    "matched": true,
    "matchedID": 1
  }
}

===================================

Discover:

Use to fetch potential matches based on age range, gender and radius

URL: /discover

HEADERS: Content-Type: application/json, Authorization: Bearer <Token>

Method: GET

QUERY PARAMS: age_range, gender, distance_from

age_range should follow the `{minimum age}-{maximum age}` e.g 18-30. Defaults to 18-60

gender should be either M, F or 0 for male, female and others respectively.

distance_from should be an integer representing the distance from user in Kilometers. Defaults to 50

RESPONSE

{
  "results": [
    {
      "id": 12,
      "name": "Lottie Ledner",
      "gender": "M",
      "age": 33,
      "distanceFromMe": 60
    },
    {
      "id": 13,
      "name": "Vern King",
      "gender": "M",
      "age": 23,
      "distanceFromMe": 20
    },
    {
      "id": 16,
      "name": "Clifton Hodkiewicz",
      "gender": "M",
      "age": 38,
      "distanceFromMe": 40
    }
  ]
}

---

ASSUMPTIONS

• Minimum age for creating a profile is 18 years.
• Maximum age is 60 years.
• User location is stored as longitude and latitude.
• The distance from users is a distance between 2 point on earth.
• For a dating app we need to find users within a few kilometers radius, hence I used a resolution of 9 for H3 indexing which is a good balance between performance and precision.
• Attractiveness is calculated by number of YES swipes on a user's profile.

GOING FORWARD

• More refactors to be done (probably introduce a dependency injection framework to properly handle dependencies).
• More extensive tests and test coverage.
• Refactor tests setup process across most tests. Currently similar setup logic is repeated in several places. (FIX THAT).
• Handle NO swipes on user profiles, Probably store them to exclude the involved profiles from being shown to the user in future.
• Build separate service that regularly aggregates potential matches for active users based on more profile info, activities and other statistics (might include some ML).
• Add Super Like Feature that ranks profile to the top of the list of potential matches.
• What happens after match????