MinHashSharp 1.1.0
See the version list below for details.
dotnet add package MinHashSharp --version 1.1.0
NuGet\Install-Package MinHashSharp -Version 1.1.0
<PackageReference Include="MinHashSharp" Version="1.1.0" />
paket add MinHashSharp --version 1.1.0
#r "nuget: MinHashSharp, 1.1.0"
// Install MinHashSharp as a Cake Addin #addin nuget:?package=MinHashSharp&version=1.1.0 // Install MinHashSharp as a Cake Tool #tool nuget:?package=MinHashSharp&version=1.1.0
MinHashSharp - A Robust Library for Similarity Estimation
MinHashSharp
offers a simple lightweight data structure designed to index and estimate Jaccard similarity between sets. Leveraging its robust structure, it has been successfully tested on datasets as large as 60GB, encompassing tens of millions of documents, while ensuring smooth and efficient operations.
Installation
To incorporate MinHashSharp into your project, choose one of the following methods:
.NET CLI
dotnet add package MinHashSharp
NuGet Package Manager
Install-Package MinHashSharp
For detailed package information, visit MinHashSharp on NuGet.
Key Features
The library currently offers two classes:
MinHash
: A probabilistic data structure for computing Jaccard similarity between sets.
MinHashLSH
: A class for supporting big-data fast querying using an approximate Jaccard similarity
threshold.
Sample usage
string s1 = "The quick brown fox jumps over the lazy dog and proceeded to run towards the other room";
string s2 = "The slow purple elephant runs towards the happy fox and proceeded to run towards the other room";
string s3 = "The quick brown fox jumps over the angry dog and proceeded to run towards the other room";
var m1 = new MinHash(numPerm: 128).Update(s1.Split());
var m2 = new MinHash(numPerm: 128).Update(s2.Split());
var m3 = new MinHash(numPerm: 128).Update(s3.Split());
Console.WriteLine(m1.Jaccard(m2));// 0.51
var lsh = new MinHashLSH(threshold: 0.8, numPerm: 128);
lsh.Insert("s1", m1);
lsh.Insert("s2", m2);
Console.WriteLine(string.Join(", ", lsh.Query(m3))); // s1
Multi-threading
The library is entirely thread-safe except for the MinHashLSH.Insert
function (and the custom injected hash function, if relevant). Therefore, you can create MinHash
objects on multiple threads and query the same MinHashLSH
object freely. If you are indexing sets on multiple threads, then just make sure to gain exclusive access to the LSH around every Insert
call:
lock (lsh) {
lsh.Insert("s3", m3);
}
Custom hash function
By default, the library uses the Farmhash function introduced by Google for efficiency. For more accurate hashes, one can inject a custom hash function into the MinHash
object.
For example, if you want to use the C# default string hash function:
static uint StringHash(string s) => (uint)s.GetHashCode();
var m = new MinHash(numPerm: 128, hashFunc: StringHash).Update(s1.Split());
Product | Versions Compatible and additional computed target framework versions. |
---|---|
.NET | net6.0 is compatible. net6.0-android was computed. net6.0-ios was computed. net6.0-maccatalyst was computed. net6.0-macos was computed. net6.0-tvos was computed. net6.0-windows was computed. net7.0 was computed. net7.0-android was computed. net7.0-ios was computed. net7.0-maccatalyst was computed. net7.0-macos was computed. net7.0-tvos was computed. net7.0-windows was computed. net8.0 was computed. net8.0-android was computed. net8.0-browser was computed. net8.0-ios was computed. net8.0-maccatalyst was computed. net8.0-macos was computed. net8.0-tvos was computed. net8.0-windows was computed. |
-
net6.0
- Farmhash.Sharp (>= 0.9.0)
- MathNet.Numerics (>= 5.0.0)
- TqdmSharp (>= 1.3.1)
NuGet packages
This package is not used by any NuGet packages.
GitHub repositories
This package is not used by any popular GitHub repositories.
1.1.0: Added serialize/deserialize option for an LSH index; Added option to "freeze" index.