202201102140 Status: #library Tags: [[contents#Misc]]

The SNH's Misc. Tools: snhlib

Install: pip install git+https://github.com/Shidiq/snhlib.git

Latest version: 0.0.1-alpha.14-

Features

Library	Class	Function	Params	Return	Information
finder	FinderScan	scanroot	root, pattern='*.csv', verbose=0	full-path-files, (path, subdirs, files)	get spesific pattern's files in root and sub-root
		openDataGeNose	item, cols=None	Pandas dataframe	Open item full-path file (csv or json) for open data GeNose
stats	Significant	KolmogorovSmirnov	kontrol, test	value and information	Significant hypothesis test using K-S method
		PSI	kontrol, test	value and information	Significant hypothesis test using PSI method
		cohend	kontrol, test	value and information	Calculate effect size using Cohen's method
image	Style	reset	-	-	Reset rcParams (matplotlib)
		paper	loc='best', classic=True, figsize=[10.72, 8.205]	figure, axes	Update rcParams with custom matplotlib theme
dataviz	-	boxplot	data, id_vars, value_vars, hue=None, hue_order=None, options	figure, axes	Boxplot using pandas dataframe as input and spesicif columns for target value
	CalcPCA	init	round_=1, featurename=None, scaler=StandardScaler, colors, markers		Initial configuration
		fit	x, y	pca model	analyze PCA using input x and target y
		getvarpc	-	pc score, variance , eigen value
		getcomponents	-	loading score
		getbestfeature	PC=0, n=3	top n loading score
		plotpc	PC, size, ellipse, ascending, legend, loc	figure	Plot PCA analysis
		screenplot	PC	figure
	CalcLDA	init	round_=1, scaler, colors, markers, cv
		fit	[x, y] or [xtrain, xtest, ytrain, ytest]
		getvarld	-	lda score and variance
		getscore	-	cross-validation scores
		plotlda	ellipse, ascending, legend, loc	figure	Plot LDA analysis

Requirements

• pandas
• scipy
• numpy
• matplotlib
• seaborn
• scikit-learn

---

Release Notes:

• 0.0.1-alpha.1 (2021122800):
  • finder - FinderScan - scanroot
• 0.0.1-alpha.10 (2021122900):
  • stats - Significant - [KolmogorovSmirnov, PSI, cohend]