Run predictions with API on zip folder instead of csv file

[MargauxMMF]

Hi! 

I am trying to use the Predictions API to run predictions on images in a zipped folder. I can upload the zip folder on the UI predict but I want to use the API.

For example, instead of using this:

python predict.py <input-file.csv> <output-file.csv> <deployment_id> --api_key=<API_KEY> --host=https://app2.datarobot.com

I want to do something like this:

python predict.py <dataset.zip> <output-results-file.csv> <deployment_id> --api_key=<API_KEY> --host=https://app2.datarobot.com

Or same question but with this command line:

How can I do this?

Thank you!

[IraWatt · Accepted solution]

Hey @MargauxMMF, 

The code for uploading a zip file is the same for uploading a CSV, you just need to make sure that the zip is structured correctly (post on correct structuring here).

import datarobot as dr
dr.Client(token='', endpoint='')

dr.Project.start("Sub Folders.zip",project_name="test_name")

With that in mind your command should work for both CSV and ZIP.

All the best,

Ira