Skip to content
/ gordo Public

An API-first distributed deployment system of deep learning models using timeseries data to predict the behaviour of systems

License

Notifications You must be signed in to change notification settings

equinor/gordo

Repository files navigation

Gordo


Building thousands of models with timeseries data to monitor systems.


Table of Contents


About

Gordo fulfills the role of inhaling config files and supplying components to the pipeline of:

  1. Fetching data
  2. Training model
  3. Serving model

Components


Documentation is available on Read the Docs


Install

gordo-helm - you can use gordo helm chart from this repository to deploy gordo infrastructure to your Kubernetes cluster.

Python package

pip install --upgrade gordo

With additional extras: pip install gordo[postgres,mlflow]

Bleeding edge:
pip install git+https://github.com/equinor/gordo.git

Developer manual

This section will explain how to start development of Gordo.

Setup

Create and activate a virtual environment first. As a default option, it can be venv module.

Install pip-tools

pip install --upgrade pip
pip install --upgrade pip-tools

Install requirements

pip install -r requirements/full_requirements.txt
pip install -r requirements/test_requirements.txt

Install package:

python3 setup.py install

How to update packages

Note: you have to install pip-tools version higher then 6 for requirements to have same multi-line output format.

To update some package in full_requirements.txt:

  • Change its version in requirements.in file;
  • Compile and upgrade requirements:
pip-compile --upgrade --output-file=full_requirements.txt mlflow_requirements.in postgres_requirements.in requirements.in  

Examples

See our example notebooks for how to develop with gordo locally.

How to run tests locally

List of commands to run tests can be found here. Running of tests takes some time, so it's faster to run tests in parallel:

pytest -n auto -m 'not dockertest' --ignore benchmarks

Run docker-related tests:

pytest -m 'dockertest'

NOTE: To run tests it's required for your system to has (note: commands might differ from your OS):

  • Running docker daemon.
  • Available 5432 port for postgres container.

NOTE: this example is for Pycharm IDE to use breakpoints in the code of the tests.
On the configuration setup for test running add to Additional arguments: in pytest section following string: --ignore benchmarks --cov-report= --no-cov

Build the documentation

This command will run the local documentation server:

> cd docs/
> make watch

Contributing

We welcome contributions to this project! To get started, please follow these steps:

  1. Fork this repository to your own GitHub account and then clone it to your local device.
git clone https://github.com/your-account/your-project.git
  1. Create a new branch for your feature or bug fix.
git checkout -b your-feature-or-bugfix-branch
  1. Make your changes and commit them with a descriptive message.
git commit -m "Add a new feature" -a
  1. Push your changes to your forked repository.
git push origin your-feature-or-bugfix-branch
  1. Open a pull request in this repository and describe the changes you made.

We'll review your changes and work with you to get them merged into the main branch of the project.

About

An API-first distributed deployment system of deep learning models using timeseries data to predict the behaviour of systems

Topics

Resources

License

Security policy

Stars

Watchers

Forks

Packages

 
 
 

Languages