Mark-3-Cloud

Mark-3-Cloud is a self-hosted personal cloud app built with Flask. It supports user accounts, file management, profile updates, and lightweight image classification for uploaded images. The current implementation is designed to run offline in Docker and uses local fonts and local file processing only.

Features

• User signup and login with password hashing
• Password policy enforcement:
  • at least 9 characters
  • at least one letter
  • at least one uppercase letter
  • at least one number
  • at least one symbol
• Profile update and profile picture upload
• File upload, batch upload, folder upload, download, preview, rename, and delete
• Image classification for uploaded image files
• Human and non-human classification labels shown on image cards
• Per-user file storage under the app upload folder
• Offline-friendly UI with no external font dependency

Current Database Setup

The app currently uses PostgreSQL in Docker.

• docker-compose.yml starts a postgres:15 service
• The Flask app reads DATABASE_URL from the environment
• Existing databases are repaired at startup so the classification column is added automatically if missing

SQLite is not the default runtime database in the current model. If you want SQLite for local-only development, you would need to set DATABASE_URL manually to a SQLite URI.

Project Structure

Mark-3-Cloud/
├── app/
│   ├── __init__.py
│   ├── models.py
│   ├── routes.py
│   ├── static/
│   │   ├── css/
│   │   │   └── style.css
│   │   ├── images/
│   │   │   └── bg.jpg
│   │   ├── js/
│   │   │   ├── dashboard.js
│   │   │   └── password.js
│   │   └── uploads/
│   │       ├── .gitkeep
│   │       ├── 1/
│   │       ├── 3/
│   │       └── 6.jpg
│   └── templates/
│       ├── dashboard.html
│       ├── login.html
│       └── signup.html
├── instance/
├── static/
│   └── uploads/
├── Dockerfile
├── docker-compose.yml
├── requirements.txt
├── run.py
└── README.md

Notes:

• app/static/uploads/ is the active upload location used by the app.
• static/uploads/ exists as a repository-level runtime folder and may be used for mounted or local storage setups.
• User uploads are created under user-specific subfolders when needed.

Running With Docker

1. Create a .env file in the project root.
2. Set at least:

SECRET_KEY=your-secret-key
DATABASE_URL=postgresql+psycopg2://postgres:password@db:5432/cloud_db

3. Start the stack:

docker compose up --build

4. Open the app at http://localhost:5000.

Local Development

If you want to run the app outside Docker, make sure DATABASE_URL points to a reachable PostgreSQL database. The app still expects that environment variable to exist.

Install dependencies with:

pip install -r requirements.txt

Then run:

python run.py

Offline / USB Transfer Setup

To run the app on another machine without internet (or to avoid downloading large Docker images over a slow connection):

On the Machine with Internet (Build Machine)

1. Build and run the stack locally first:

docker compose build --no-cache web
docker compose up -d

2. Save the Docker images as tar files:

docker save -o mark3-web.tar mark-3-cloud-web:latest
docker save -o postgres-15.tar postgres:15

3. Copy the tar files and the entire project folder to a USB drive.

On the Target Machine (Without Internet)

1. Ensure Docker Desktop is installed and running.

2. Copy the project folder and tar files from USB to a local folder (e.g., C:\mark3-cloud).

3. Load the Docker images:

cd C:\mark3-cloud
docker load -i mark3-web.tar
docker load -i postgres-15.tar

4. Create a .env file (same as above):

SECRET_KEY=your-secret-key
DATABASE_URL=postgresql+psycopg2://postgres:password@db:5432/cloud_db

5. Start the stack without building (using the pre-loaded images):

docker compose up -d --no-build

6. Open the app at http://localhost:5000.

Note: The first startup will take a moment as PostgreSQL initializes the database. After that, subsequent startups are much faster.

Notes

• Uploads are stored under app/static/uploads/<user_id>/.
• Folder uploads preserve relative paths when the browser supports them.
• The app creates missing upload directories automatically.
• Image classification is local and does not require internet access.
• External font loading has been removed so the UI works fully offline.