Installation Guide

Prerequisites

Before installing Deeptrain, ensure your environment meets the following requirements:

• Python: Version 3.9 or higher.
• Package Manager: pip (standard with Python) or poetry.
• System Dependencies:
  • FFmpeg: Required for video and audio processing (Transcribe API).
  • Tesseract OCR: Required for processing flowcharts, graphs, and images.

System-Level Installation

Depending on your operating system, you may need to install external binary dependencies to handle multi-modal data.

macOS

Use Homebrew to install the necessary media processing libraries:

brew install ffmpeg tesseract

Ubuntu/Debian

Update your package list and install the required utilities:

sudo apt update
sudo apt install -y ffmpeg tesseract-ocr libtesseract-dev

Windows

1. FFmpeg: Download the builds from ffmpeg.org, extract them, and add the bin folder to your System PATH.
2. Tesseract: Download the installer from the UB Mannheim repository and add the installation path to your System PATH.

Python Package Installation

Install the Deeptrain core library using pip. It is recommended to use a virtual environment (venv or conda) to manage your dependencies.

# Create a virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install Deeptrain
pip install deeptrain

Optional Dependencies

If you plan to use specific local embedding databases or specialized vision models, you can install the extended suites:

pip install "deeptrain[vision,audio]"

Configuration

Deeptrain requires specific environment variables to interface with your chosen LLMs and the Transcribe API. Create a .env file in your project root:

# Core Deeptrain Configuration
DEEPTRAIN_API_KEY=your_deeptrain_key_here

# Model Provider Keys (Example for OpenAI/Anthropic)
OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...

# Storage Configuration (Optional)
LOCAL_DB_PATH=./data/embeddings

Initializing the Client

Once installed, you can initialize the Deeptrain connector in your Python script:

from deeptrain import DeeptrainConnector

# Initialize the connector
# The library automatically loads credentials from environment variables
connector = DeeptrainConnector()

print("Deeptrain successfully initialized.")

Verifying the Installation

To ensure that the multi-modal components (specifically video and audio) are working correctly, run the following verification command:

python -m deeptrain.verify

This internal utility checks for:

1. FFmpeg accessibility: Confirms video frames can be extracted.
2. Tesseract availability: Confirms text can be read from images/graphs.
3. API connectivity: Validates your connection to the Deeptrain processing backend.

Advanced Environment Setup

Docker Installation

For consistent behavior across environments, you can use the official Deeptrain Docker image which comes pre-configured with all system-level dependencies:

docker pull uditakhourii/deeptrain:latest
docker run -it --env-file .env uditakhourii/deeptrain

Private Model Support

If you are using one of the 200+ supported private or open-source models, ensure your local inference server (like Ollama or LocalAI) is running before initializing Deeptrain, and point the BASE_URL in your configuration to your local endpoint.