Finding most similar element

Steps to calculate similarity of html element based on attributes values:

• Calculate weight coefficient of each attribute of a target element based on its lentgh:
  attr1.value.length / ( attr1.value.length + attr2.value.length ... + attrN.value.length)
• For each attribute with same key, calculate string similarity coefficient for values of two elements
• For each shared attribute multiply weight and similarity coefficients
• Summarize all result from previous step

Calculation example

Attribute key	Element 1 attr	Element 1 attr	Length	Weight	Similarity	Weighted Similarity
class	btn btn-success	btn btn-default	15	0.32	0.88	0.28
href	#success	#done	8	0.16	0.34	0.05
rel	next	last	4	0.08	0.50	0.04
title	It's all ok		11	0.22	0.00	0.00
onclick	function1()	function2()	11	0.22	0.98	0.21
TOTAL:			49	1.00		0.58

In this example elemets similarity coefficient is 0.58

So, all compared elements are stored in a sorted map in this implementation, with reverse order sorting, where key is similarity coefficient and value is element itself. And top element in this sorted map is the most similar to a target element.

Run

java -jar app.jar <required: abs path file1> <required: abs path file2> <optoinal: element id from file1>

output is just an xPath to the most similar element from file2

Run on examples

$ java -jar app.jar sample-0-origin.html sample-1-evil-gemini.html
html > body > div > div > div[3] > div[1] > div > div[2] > a[2]

$ java -jar app.jar sample-0-origin.html sample-2-container-and-clone.html 
html > body > div > div > div[3] > div[1] > div > div[2] > div > a

$ java -jar app.jar sample-0-origin.html sample-3-the-escape.html 
html > body > div > div > div[3] > div[1] > div > div[3] > a

$ java -jar app.jar sample-0-origin.html sample-4-the-mash.html 
html > body > div > div > div[3] > div[1] > div > div[3] > a