featherhelper

Feather Helper is a concise interface to cache numpy arrays and pandas dataframes.


Keywords
cache, numpy, pandas, pyarrow, python
License
MIT
Install
pip install featherhelper==0.0.5

Documentation

Feather Helper

Feather Helper is a concise interface to cache and load numpy arrays and pandas dataframes. I use it with Pandoctools/Knitty.

Contents

Install

Via conda:

conda install -c defaults -c conda-forge featherhelper

Via pip:

pip install featherhelper

Usage example

import pandas as pd
import numpy as np
import featherhelper as fh
fh.setdir("~/feather/mydoc")  # (optional)
# fh.exc(1, 2)  # force raise exceptions for names (optional)

# %%
fh.name(1)  # can also be fh.name('id1'), default is 'default', 1 is the same as '1'
try:
    # raise fh.Err  # (optional)
    df, A, B = fh.pull()  # control length can be set: fh.pull(N)
except fh.Err:
    # calculate:
    print('push')  
    df = pd.DataFrame(np.random.random(16).reshape(4, 4))
    A = df.values
    B = np.random.random(16 * 3).reshape(4, 2, 2, 3)
    #
    fh.push(df, A, B)

print(df, '\n', A, '\n', B)

A shorter example:

import numpy as np
import featherhelper as fh
# fh.exc()

# %%
try:
    A = fh.pull()
except fh.Err:
    A = np.random.random(16).reshape(4, 4)
    fh.push(A)