noformat 0.1.4

save and load a structured collection of data as folder

Basically, noformat treats a folder structure as a single dict that contains data and attributes. Pandas dataframe and numpy array can be used for dict values.

import numpy as np
import pandas as pd
from noformat import File
data = File('data/temp', 'w-')
data['first_array'] = np.random.randn(10, 10)
data['second_array'] = pd.DataFrame(data=np.random.randn(10, 4), columns=['1', '2', '3', '4'])

Files will be automatically saved upon object destruction. And loaded later

read_data = File('data/temp', 'w+')
assert(read_data['first_array'].shape == (10, 10))

Attributes will be saved in ‘attributes.json’ files

read_data = File('data/temp', 'w+')
read_data.attrs['first_attribute'] = 64

This will create a folder with the following structure

data/temp/
|   first_array.npy
|   second_array.msg
└───attributes.json

Logging files for the data can be included: 1. in json format with .log extension 2. in cell array format with .mat extension It can only be written in the first type