Fiore is an open-source, event-based social media platform for spontaneous interest-based meetups. It's like a digital common room — students and communities can raise or join last-minute trips, rides, jam sessions, games, movie nights, or any group activity with open spots.

Fiore evolved into a graph-powered engine for discovery and connection, built with FastAPI, PostgreSQL + Apache AGE, and object storage using MinIO. It supports REST, GraphQL, and WebSocket APIs and offers personalized, real-time recommendations.

• FastAPI backend with modular architecture
• REST, GraphQL, and WebSocket APIs
• Real-time event-based communication with WebSockets
• Graph-driven user, event, and community recommendations using Apache AGE
• PostgreSQL for structured data
• MinIO for image and file storage (S3-compatible)
• JWT and API key authentication
• Pytest-based testing suite
• Docker-ready deployment (docker branch)

Imagine you're a student wanting to go for a late-night coffee run, but you're looking for company. Or you're heading for a trip and have 2 empty seats. Or you're starting a pick-up game on campus. Fiore lets you raise events on-the-fly, and others can discover and join them instantly based on shared interests and time windows.

Communities can form around frequent event types. Profiles grow with participation. The social graph evolves. And everything runs in real time.

📜 Swagger UI: http://localhost:8000/docs

Create a .env file in the root directory with the following variables:

DB_USER=fiore
DB_PASSWORD=strong_paa
DB_NAME=fiore
DB_HOST=localhost
DB_PORT=5432

JWT_SECRET=secret

MINIO_ENDPOINT=http://localhost:9000
MINIO_ACCESS_KEY=key
MINIO_SECRET_KEY=secret
MINIO_BUCKET=fiore
MINIO_USE_SSL=False

API_KEY=random

TEST_USER_PASSWORD=user1
TEST_IMAGE_PATH_ABS=/path/to/image

1. Install PostgreSQL (version 13+ recommended):

sudo apt install postgresql postgresql-contrib

2. Install Apache AGE: Follow the official guide: https://github.com/apache/age

3. Enable AGE in your database:

CREATE EXTENSION age;
LOAD 'age';
SET search_path = ag_catalog, "$user", public;

4. Create a graph (you can name it fiore_graph):

SELECT create_graph('fiore_graph');

5. Run the schema:

psql -U fiore_user -d fiore -f schema.sql

Image uploads are handled via MinIO, which is an S3-compatible object store.

Here’s a basic setup snippet using the Go SDK:

package main

import (
  "context"
  "github.com/minio/minio-go/v7"
  "github.com/minio/minio-go/v7/pkg/credentials"
  "log"
)

func main() {
  endpoint := "localhost:9000"
  accessKeyID := "key"
  secretAccessKey := "secret"
  useSSL := false

  minioClient, err := minio.New(endpoint, &minio.Options{
    Creds:  credentials.NewStaticV4(accessKeyID, secretAccessKey, ""),
    Secure: useSSL,
  })
  if err != nil {
    log.Fatalln(err)
  }

  bucketName := "fiore"
  location := "us-east-1"

  err = minioClient.MakeBucket(context.Background(), bucketName, minio.MakeBucketOptions{Region: location})
  if err != nil {
    exists, errBucketExists := minioClient.BucketExists(context.Background(), bucketName)
    if errBucketExists == nil && exists {
      log.Printf("Bucket %s already exists\n", bucketName)
    } else {
      log.Fatalln(err)
    }
  } else {
    log.Printf("Successfully created bucket %s\n", bucketName)
  }
}

This bucket will store all user-uploaded media files like avatars and banners.

To spin up the complete stack including API, PostgreSQL, AGE, and MinIO:

git checkout origin/docker
docker-compose up -d --build

This command launches everything in a self-contained environment using the preconfigured Dockerfiles and Compose file.

pytest tests

Fiore is open-source under the MIT License.