jupyXplorer

This project aims to generate some util notebooks to get data exploration.

It's based on nbconvert and simply offer you a easy way to analise a feature value (or set of them) of a dataset.

Installation

In order to install the utility just use pip:

pip install jupyXplorer

We provide you a docker image to get started quickly, take a look at hypnosapos/jupyxplorer at dockerhub and select the best one fit your needs.

Usage

Before notebook generation take a look at our config file example (tests/e2e/sample_config.yaml).

jupyxplorer -c config.yaml -o .output

All notebooks will be ready in directory ".output" of the example above.

Once notebooks are generated we can use them in your jupyter to show results of data analysis.

As we said above, you may use a docker container instead:

docker run -it -v </path/my_config.yaml>:/tmp/my_config.yaml </path/output>:/tmp/output hypnosapos/jupyxplorer:latest jupyxplorer -c /tmp/my_config.yaml -o /tmp/output

If you want to use some required files to execute notebooks before generation (-e or --execute argument), provide the value of argument "input-dir" (--input-dir, or the short form -i). Here you have an example:

jupyxplorer -i .input -c config.yaml -o .output -e

The resources under that directory would be particular requirement files for python and local dataset files.

Dev

The development lifecycle is managed by a Makefile and CircleCI, where all steps are executed through docker containers. Type make help to see all available commands.