Skip to content

qri-io/qri-python

Repository files navigation

qri-python

 ██████╗ ██████╗ ██╗
██╔═══██╗██╔══██╗██║
██║   ██║██████╔╝██║
██║▄▄ ██║██╔══██╗██║
╚██████╔╝██║  ██║██║
 ╚══▀▀═╝ ╚═╝  ╚═╝╚═╝

Python client for qri ("query")

Installation

pip install qri

About

Python wrapper to enable usage of qri, the dataset toolchain. Can either use a locally installed qri command-line program to work with your local repository, or can directly get datasets from the Qri Cloud.

Dataset objects returned by this library have the components that exist in the standard qri model. The body is returned as a Pandas DataFrame in order to easily integrate with other data science systems, like Jupyter Notebook.

Usage

The following examples assume you have the latest release of the qri command-line client installed. You can get this from https://github.com/qri-io/qri/releases

import qri
# Pull a dataset from cloud and add it to your repository
$ qri.pull("b5/world_bank_population")

Fetching from registry...

"Added b5/world_bank_population: ..."

# List datasets in your repository
$ qri.list()

[Dataset("b5/world_bank_population")]

# Get that single dataset as a variable
$ d = qri.get("b5/world_bank_population")
# Look at metadata description
$ d.meta.description

( 1 ) United Nations Population Division. World Population Prospects: 2017 Revision...

# Get the dataset body as a pandas DataFrame
$ d.body

. country_name country_code indicator_name ...

0 Afghanistan AFG Population, total ...

...

TODO: Save changes

Development

Clone this repository

git clone https://github.com/qri-io/qri-python

Navigate to the directory where you run jupyter from:

cd /path/where/jupyter/is/run

Symlink the cloned repository's source directory here:

ln -s /path/to/cloned/qri-python/qri .

NOTE: The clone command created the directory "qri-python", and inside is the source directory named "qri". Make sure to symlink the source directory, not just the repository root

This package should now be usable from within Jupyter Notebook