# PB Name matching
Documenting different name matching options

## `String.Equals()`
Strip down first/last names from PB and BT and see if they match
### Pros
* relatively fast
* simple
### Cons
* could cause a lot of failures due to shortened names
* robert : bob
* william : bill
* steven : steve

## Levenshtein distance
Wiki: https://en.wikipedia.org/wiki/Levenshtein_distance
We can utilize this to gain a configurable confidence intervla of a name match.

### Pros
* Configurable
* Handles fuzzyness well
### Cons
* costly expression
* even performant libraries have a performance overhead
* needs to be configured
* Will need to determine an appropriate "confidence interval"

## FuzzySharp
https://github.com/JakeBayer/FuzzySharp
Generates a "confidence interval" we can use to rate if we think a name matches up.
### Pros
* Configurable
* we can set the confidence interval in the deployment options
* Handles fuzzyness better than pure lavenshtein distance
* May not need to remove special characters?
### Cons
* Somewhat costly
* needs to be configured
### Sample:

// See https://aka.ms/new-console-template for more information

using System.Diagnostics;
using FuzzySharp;



Console.WriteLine("Hello, World!");


var testList = new (string,string)[]{
    ("robert stephenson", "bob stephenson"),
    ("mysmilarstring", "myawfullysimilarstirng"),
    ("Christopher Decker", "Chris Decker"),
    ("Steve McAleer", "Steven Mcaleer"),
    ("Matthew Roebuck", "Matt Roebuck")
};

var stopWatch = new Stopwatch();

foreach (var (name1, name2) in testList)
{
    stopWatch.Restart();
    var score = Fuzz.WeightedRatio(name1, name2).ToString();
    stopWatch.Stop();
    Console.WriteLine($"Comparing '{name1}' : '{name2}' scored {score} and took {stopWatch.ElapsedMilliseconds.ToString()}ms");
}

/*
 * Comparing 'robert stephenson' : 'bob stephenson' scored 84 and took 81ms
 * Comparing 'mysmilarstring' : 'myawfullysimilarstirng' scored 77 and took 10ms
 * Comparing 'Christopher Decker' : 'Chris Decker' scored 80 and took 0ms
 * Comparing 'Steve McAleer' : 'Steven Mcaleer' scored 89 and took 0ms
 * Comparing 'Matthew Roebuck' : 'Matt Roebuck' scored 89 and took 0ms
*/

# Implementation
## Deposits
Implementations would be gateway specific as not all gateways provide account name information.
Cost is relatively cheap, could be done as part of the pre-orchestration validation.

## Withdrawals
Could be completed as part of the pre-assessment validation
Additional field for consideration during assessment (both via admin and WAE)

# Summary
Fuzzy ratio is an additional piece of metadata we can use to help with deposit/withdrawal assessment.