Skip to content

GSoC 2019-Installing datasets stored in provenance directory

Posted on:August 3, 2019 at 08:55 AM

Hey everyone,
Just two more weeks to go for the end of GSoC 2019. Here’s a look into what I did in the past two weeks.

Last two weeks I worked on implementing two things.
Created a command to see the log of datasets stored in the provenance directory.

retriever log dataset_name

For eg.

retriver log abalone-age

Commit message: New test
Hash: 02ee77
Date: 07/27/2019, 20:30:00

Commit message: save dataset
Hash: a76e77
Date: 07/19/2019, 18:22:31

In python interface it can be used like this:

from retriver import commit_log

Created a command to install datasets from provenance directory. It’s the same as the command used for installing datasets. The only difference is that if the user provides a hash value with the command for installation of dataset then the installation has to be performed by using already committed dataset in the provenance directory.
Here’s how to use it:

retriever install sqlite dataset --hash hash_value

For eg.

retriever install sqlite abalone-age --hash 02ee77

In python interface it can be used like this:

from retriever import install_csv
install_csv('abalone-age', hash='02ee77')

The pull request for the installlation of dataset from the provenance directory is open and under review.