Introduction

In this post, we introduce TDAG, our scalable and distributed engine designed to translate trading strategies directly into directed acyclic graphs (DAGs).

TDAG creates reusable nodes for each step in a pipeline. A node performs calculations or queries data, and the internal logic can be customized to meet specific use cases. For instance, one node could be responsible for querying prices, which another node then uses to calculate returns.

The pipelines are designed for rapid execution and can be quickly adapted to changes in the trading strategy.

Additionally, they are easily reproducible and auditable, ensuring transparency and reliability.

‍

Installation

TDAG is integrated within the Main Sequence Ecosystem. To install TDAG and its components:

• Download VAM, TDAG, and TDAG-GUI from our repository.
• All dependencies are listed in a Pipfile, which can be installed using pipenv install.
• Ray, our chosen framework for executing the pipeline, requires initialization with ray start --head.

‍

Configuration

When you run TDAG for the first time, it automatically sets up a `tdag` folder in your home directory. Within this directory, you will find a `tdag/config.yaml` file containing default configuration settings. You may need to modify certain parameters in this configuration file for TDAG to work with your environment.

‍

Next Steps

With the setup complete, let's proceed by creating a custom portfolio using TDAG in our Getting Started Tutorial. There we demonstrate how to utilize TDAG's capabilities for efficient and flexible trading strategy implementation.

‍