pdpatch

pdpatch adds methods to pandas’ DataFrame and Series for a faster data science pipeline. It also defines drop-in replacements for seaborn and plotly.express that automatically label axes with nicer titles. We use nbdev to build this project.

Install

pip install pdpatch

How to use

from pdpatch.all import *

Interactive Method .less()

Automatically Rename snake_case columns in plotly.express and seaborn

import pandas as pd
from pdpatch.express import *
df = pd.DataFrame({'time__s__': range(10), 'position__m__': [i**1.3 for i in range(10)], 'speed__m/s__': 10*[1]})
#df = pd.DataFrame({'time__s__': range(10), 'position__m__': range(10)})
px.scatter(df, x='time__s__', y='position__m__').show('png')

from pdpatch.seaborn import sns
sns.scatterplot(data=df, x='time__s__', y='position__m__');

Add Altair-like Operation to plotly Figures

fig = px.scatter(df,x='time__s__', y='time__s__') | px.scatter(df,x='time__s__', y=['position__m__', 'speed__m/s__'])
fig.show('png')

fig = px.scatter(df,x='time__s__', y='time__s__') / px.scatter(df,x='time__s__', y=['position__m__', 'speed__m/s__'])
fig.show('png')

fig = px.scatter(df,x='time__s__', y='time__s__') | px.scatter(df,x='time__s__', y=['position__m__', 'speed__m/s__'])
(fig / fig).show('png')

Shorter methods

df.rename(columns={'col_1': 'new_name'})->df.renamec('col_1', 'new_name')

df = dummydf()
df.renamec('col_1', 'new_name').to_html()

Functions as methods

df.len()

5

New methods

df.col_1.minmax

(100, 104)

Utility functions

df = dummydf()
df.to_html()