tedkulp/Readarr
1
0
Fork 0
mirror of https://github.com/Readarr/Readarr.git synced 2026-04-26 22:46:37 -04:00
Code Issues Projects Releases Wiki Activity

Improve the fuzzy matching (#522)

Browse Source 
* Fixed: improve track matching

* Deal with tracks sequentially numbered across discs
This commit is contained in:
ta264
2018-11-16 17:46:46 +00:00
committed by Qstick
parent 8320508688
commit e260a29b57
19 changed files with 767 additions and 164 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/NzbDrone.Common/Extensions/StringExtensions.cs
+30 -20
View File
@@ -143,29 +143,17 @@ namespace NzbDrone.Common.Extensions
public static double FuzzyMatch(this string a, string b)
{
if (a.Contains(" ") && b.Contains(" "))
if (a.IsNullOrWhiteSpace() || b.IsNullOrWhiteSpace())
{
return 0;
}
else if (a.Contains(" ") && b.Contains(" "))
{
var partsA = a.Split(' ');
var partsB = b.Split(' ');
var weightedHighCoefficients = new double[partsA.Length];
var distanceRatios = new double[partsA.Length];
for (int i = 0; i < partsA.Length; i++)
{
double high = 0.0;
int indexDistance = 0;
for (int x = 0; x < partsB.Length; x++)
{
var coef = LevenshteinCoefficient(partsA[i], partsB[x]);
if (coef > high)
{
high = coef;
indexDistance = Math.Abs(i - x);
}
}
double distanceWeight = 1.0 - (double)indexDistance / (double)partsA.Length;
weightedHighCoefficients[i] = high * distanceWeight;
}
return weightedHighCoefficients.Sum() / (double)partsA.Length;
var coef = (FuzzyMatchComponents(partsA, partsB) + FuzzyMatchComponents(partsB, partsA)) / (partsA.Length + partsB.Length);
return Math.Max(coef, LevenshteinCoefficient(a, b));
}
else
{
@@ -173,6 +161,28 @@ namespace NzbDrone.Common.Extensions
}
}
private static double FuzzyMatchComponents(string[] a, string[] b)
{
double weightDenom = Math.Max(a.Length, b.Length);
double sum = 0;
for (int i = 0; i < a.Length; i++)
{
double high = 0.0;
int indexDistance = 0;
for (int x = 0; x < b.Length; x++)
{
var coef = LevenshteinCoefficient(a[i], b[x]);
if (coef > high)
{
high = coef;
indexDistance = Math.Abs(i - x);
}
}
sum += (1.0 - (double)indexDistance / weightDenom) * high;
}
return sum;
}
public static double LevenshteinCoefficient(this string a, string b)
{
return 1.0 - (double)a.LevenshteinDistance(b) / Math.Max(a.Length, b.Length);