https://github.com/jay-johnson/redten-python

Python client and docker image for the red10 machine learning api
https://github.com/jay-johnson/redten-python

machine-learning python2 redis rest-api

Last synced: 28 days ago
JSON representation

Python client and docker image for the red10 machine learning api

• Host: GitHub
• URL: https://github.com/jay-johnson/redten-python
• Owner: jay-johnson
• License: apache-2.0
• Created: 2017-05-28T05:08:52.000Z (over 7 years ago)
• Default Branch: master
• Last Pushed: 2017-05-29T19:56:29.000Z (over 7 years ago)
• Last Synced: 2024-10-13T09:41:28.192Z (about 1 month ago)
• Topics: machine-learning, python2, redis, rest-api
• Language: Python
• Homepage: https://redten.io
• Size: 28.3 KB
• Stars: 4
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.rst
  • License: LICENSE

Awesome Lists containing this project

README

        
================

Red10 Python API

================

The standalone python client and docker container for: https://redten.io

.. figure:: http://i.imgur.com/65dsKzX.png

    :alt: Red10 distributed machine learning on Redis

#.  To authenticate you will need valid login credentials

    - Username

    - Password

    - Email

#.  Please export these environment variables

    ::

    

        export ENV_REDTEN_USER=Username

        export ENV_REDTEN_PASS=Password

        export ENV_REDTEN_EMAIL=Email

        export ENV_REDTEN_URL=https://api.redten.io

#.  Optional exports

    Set the csv file path. Note this file must be accessible on all worker nodes.

    ::

        export ENV_REDTEN_CSV_FILE=

    Send forecast results to emails. Comma-separated without any spacing.

    ::

        export [email protected],[email protected]

    Can also set it to an empty string which will skip sending emails:

    

    ::

        export ENV_REDTEN_FORECAST_EMAILS=

#.  Using the ``redten-python`` docker container

    Please note the docker image inflates to over 1 GB as it has the math libraries pandas and numpy installed.

    To run a forecast:

    ::

        ./run-forecast.sh

    

    Run predictions:

    ::

        ./run-predict.sh

#.  Install required pips

    If you do not want to use the docker container, here's how to setup the local environment.

    ::

        pip install -r requirements.txt

    Under the hood these are the current installed pips (pandas takes some time):

    ::

    

        pandas==0.20.0rc1

        requests>=2.13.0

        uuid>=1.30

        unittest2>=1.1.0

        simplejson>=3.10.0

#.  Install the pip from pypi repository

    ::

        pip install redten

#.  Run a forecast job

    ::

        ./bins/forecast.py 

#.  Run a prediction job

    ::

        ./bins/predict.py 

For more information:

#. `Intro`_

#. `Forecast`_

#. `Predictions with the IRIS dataset`_

.. _Intro: https://redten.io:8101/RedTen-Intro.slides.html#/

.. _Forecast: https://redten.io:8103/RT-Price-Forecast.slides.html#/

.. _Predictions with the IRIS dataset: https://redten.io:8102/RT-Run-IRIS.slides.html#/

=======

Version

=======

1.0.1

=========

Debugging

=========

I'm on fedora 24 and on a virtual env I started hitting:

::

    Failed to Run ML Job with exception='unknown error (_ssl.c:2831)'

    Predict job failed with error=Failed to Run ML Job with exception='unknown error (_ssl.c:2831)

After I reinstalled ``certifi`` it started working again.

::

    pip uninstall certifi

    pip install certifi

================

Release Commands

================

#. Upload to pypitest:

::

    python setup.py sdist upload -r pypitest

#. Upload to pypi:

::

    python setup.py sdist upload -r pypi