Red10 Python API

Python client for using: https://redten.io

1. To authenticate you will need valid login credentials

   • Username
   • Password
   • Email

2. 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
   

3. Optional exports

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

   export ENV_REDTEN_CSV_FILE=<path to csv>
   

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

   export ENV_REDTEN_FORECAST_EMAILS=email1@email.com,email2@email.com
   

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

   export ENV_REDTEN_FORECAST_EMAILS=
   

# Install required pips

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

1. Install the pip from pypi repository

   pip install redten
   

2. Run a prediction job

   ./bins/forecast.py <Dataset-Name>
   

3. Run a forecast job

   ./bins/predict.py
   

For more information:

1. Intro
2. Forecast
3. Predictions with the IRIS dataset

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

1. Upload to pypitest:

python setup.py sdist upload -r pypitest

1. Upload to pypi:

python setup.py sdist upload -r pypi