stephabauva / ML_with_DAGs

Get introduced to Directed Acyclic Graphs (DAGs) through Dagster with a simple ML program

Home Page:https://hackernoon.com/a-quick-introduction-to-machine-learning-with-dagster-gh53336m}{Article}, \href{https://github.com/stephanBV/ML_with_DAGs

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

⚠️ WARNING: This project is not maintained. Its code and dependencies are outdated. To get started with Dagster follow this link: https://docs.dagster.io/tutorial/setup

ML with DAGs ❄️

This application is a simple text classifier using sklearn for newcomers to be introduced to Dagster.

🙀 👉🏼 See a brief guided tour of Dagster and the DAG generated from this program at https://hackernoon.com/a-quick-introduction-to-machine-learning-with-dagster-gh53336m
or from the repository ./pdf/quick_start_with_ML_and_Dagster.pdf

Dagster is a data orchestrator for machine learning, analytics, and ETL. It lets you define pipelines in terms of the data flow between reusable, logical components, then test locally and run anywhere. With a unified view of pipelines and the assets they produce, Dagster can schedule and orchestrate Pandas, Spark, SQL, or anything else that Python can invoke. It makes testing easier and deploying faster 😎.

  • The script creates a single pipeline which:
    • processes the data,
    • searches for optimal parameters between a logistic regression and a random forest,
    • train and test the model

Step 1. Clone this repository 👯‍♂️

git clone https://github.com/stephanBV/ML_with_DAGs.git
cd ML_with_DAGs

Step 2. Create a virtual environment 👾 (Optional)

python3 -m virtualenv venv
source venv/bin/activate

Step 3. Install dependencies 🧞‍♂️

pip install -r requirements.txt

Step 4. Launch Dagster's UI 🐙

python3 -m dagit -f script.py

Step 5. On the main page, at the top, click on Playground.

Then, drag-and-drop the config.yml of the cloned repository to the Playground page.

Step 6. Click on Launch Execution at the bottom right of the Playground page.⚡️