Computer Vision Bot

CV Bot is a lightweight assistant for computer vision projects.
It provides basic tools to manage cameras, run inference, handle model weights, and prepare datasets for training.

---

🚀 Features (v1)

• Cameras

  • Check camera availability
  • Capture snapshots

• Weights (YOLO only)

  • Add YOLO weights
  • Use weights for inference and testing

• Recording

  • Record video streams from cameras

• Inference Recording

  • Run YOLO inference on video streams
  • Save inference results

• Training

  • Collect datasets for future model training

---

📦 Installation

1. Clone the repository

git clone https://github.com/Zzior/cv_bot.git
cd cv_bot

2. Configure environment variables

Copy the example environment file and edit it:

cp .env.example .env
nano .env

Example .env configuration:

DEFAULT_LANGUAGE="en"  # currently only English is supported
TIME_ZONE="UTC"

# If you are using Docker, do not edit Postgres configs
POSTGRES_DB="cv_bot"
POSTGRES_USER="postgres"
POSTGRES_PASSWORD="password"
POSTGRES_PORT="5432"
POSTGRES_HOST="cv_bot_db"

BOT_TOKEN="your-telegram-bot-token"
USER_IDS="1111111111 2222222222"
LOG_CHAT_ID="100"

---

⚙️ Optional: Override Docker Compose

If you need custom volume mappings or ports, create or edit compose.override.yaml:

nano compose.override.yaml

Example override configuration:

services:
  cv_bot:
    volumes: !override
      - ./:/app
      - /mnt/hdd/cv_bot:/app/storage

  cv_bot_fb:
    volumes: !override
      - /mnt/hdd/cv_bot:/srv
      - ./docker_storage/filebrowser:/database
    ports: !override
      - "7000:8080"

---

🐳 Build and Run with Docker

sudo docker compose build
sudo docker compose up -d

The bot should now be up and running in detached mode.

---

📌 Notes

• Only YOLO-based models are currently supported.
• Make sure your cameras are accessible from the Docker container.