JCAT

jcat is a self-contained command line tool for viewing jupyter notebook files in terminal. It parses notebook's underlying json content, hence it runs without the dependency of jupyter/ipython core.

Prerequisites

Most systems with g++ and make installed should be fine.

Build and Install

Clone this repo, if you want to be able to easily uninstall jcat use checkinstall.

sudo apt install checkinstall
cd jcat
sudo checkinstall

Otherwise, you can use make to install.

cd jcat
make
sudo make install

Optionally, one could use ./jcat without runing sudo make install;
or use make install PREFIX=/path/to/install for alternative installation directory (by default is /usr/local/bin).

Usage

Usage: jcat FILE [OPTION]

FILE:	A json parsable notebook file (*.ipynb).

OPTION:
  -a:	Align prompt (In/Out) for copy.

Example notebook

• Running with jcat examples/example-cifar10.ipynb:

                =========================================================================
   `markdown`   # CIFAR - 10 
                ## Decode data
                =========================================================================
   `markdown`   Activate virtual environment
                =========================================================================
        In [1]: %%bash
                source ~/kerai/bin/activate
                =========================================================================
   `markdown`   ### Imports
                =========================================================================
        In [2]: %matplotlib inline
                from helper import get_class_names, get_train_data, get_test_data, plot_images
                -------------------------------------------------------------------------
    `stderr`    Using TensorFlow backend.

                =========================================================================
   `markdown`   Get class names
                =========================================================================
        In [3]: class_names = get_class_names()
                class_names
                -------------------------------------------------------------------------
    `stdout`    Decoding file: data/batches.meta

        Out[3]: ['airplane',
                 'automobile',
                 'bird',
                 'cat',
                 'deer',
                 'dog',
                 'frog',
                 'horse',
                 'ship',
                 'truck']
=========================================================================

• Running with jcat examples/example-cifar10.ipynb -a:

Output (click to expand)

=========================================================================
# `markdown`
# CIFAR - 10 
## Decode data
=========================================================================
# `markdown`
Activate virtual environment
=========================================================================
# In [1]: 
%%bash
source ~/kerai/bin/activate
=========================================================================
# `markdown`
### Imports
=========================================================================
# In [2]: 
%matplotlib inline
from helper import get_class_names, get_train_data, get_test_data, plot_images
-------------------------------------------------------------------------
# `stderr`
Using TensorFlow backend.

=========================================================================
# `markdown`
Get class names
=========================================================================
# In [3]: 
class_names = get_class_names()
class_names
-------------------------------------------------------------------------
# `stdout`
Decoding file: data/batches.meta

# Out[3]: 
['airplane',
 'automobile',
 'bird',
 'cat',
 'deer',
 'dog',
 'frog',
 'horse',
 'ship',
 'truck']
=========================================================================

Use jcat with grep to get rid of annoying image datas

• output of cat:

• output of `jcat':

Uninstall

rm /usr/local/bin/jcat, if no extra $PREFIX is supplied during installation.

License

Distributed uder the Boost Software License.

Acknoledgement

jcat parse notebook with jsoncons, a header only c++ json library.
Example notebook from: https://github.com/09rohanchopra/cifar10/blob/master/cifar10-basic.ipynb