Merge pull request #30 from Encamina/@lmarcos/issue-27#-token-countin…

…g-functions-are-slow

@lmarcos/issue 27# token counting functions are slow

CHANGELOG.md

@@ -12,10 +12,56 @@ Each version and revision is followed by a date of change, specially for under d
 - **Major Changes**: big improvements in the code, like adding or enabling features, or bug fixes.
 - **Minor Changes**: small changes that have little impact, like spell checks in an API's documentation, adding or removing comments, etc.
 
-Also, any bug fix must start with the prefix «Bug fix:» followed by the description of the changes _per se_.
+Also, any bug fix must start with the prefix �Bug fix:� followed by the description of the changes _per se_.
 
 Previous classification is not required if changes are simple or all belong to the same category.
 
+## [8.0.1]
+
+### **Major Changes**
+ - In `Encamina.Enmarcha.SemanticKernel.Abstractions.ILengthFunctions`, `GptEncoding` is now cached and reused to improve performance. [(#30)](https://github.com/Encamina/enmarcha/pull/30)
+
+### Minor Changes
+  - Changes from version 8.0.0 have been added to the `CHANGELOG.md` file.
+
+## [8.0.0]
+
+### Major Changes
+
+ - Changed .NET version from 6 to 8, therefore closes issue `Everything ready for ENMARCHA 8.0.0 #7`.
+ - Updated the following .NET libraries to their newest version (8.0.0):
+   - Microsoft.AspNetCore.Authentication.JwtBearer
+   - Microsoft.AspNetCore.Authentication.OpenIdConnect
+   - Microsoft.EntityFrameworkCore
+   - Microsoft.EntityFrameworkCore.SqlServer
+   - Microsoft.Extensions.Caching.Abstractions
+   - Microsoft.Extensions.Configuration.Abstractions
+   - Microsoft.Extensions.DependencyInjection.Abstractions
+   - Microsoft.Extensions.Hosting
+   - Microsoft.Extensions.Http
+   - Microsoft.Extensions.Logging.Abstractions
+   - Microsoft.Extensions.Options
+   - Microsoft.Extensions.Options.ConfigurationExtensions
+   - Microsoft.Extensions.Options.DataAnnotations
+   - System.Net.Http.Json
+   - System.Text.Json
+ - Updated library Azure.Data.Tables from 12.8.1 to 12.8.2.
+ - Updated library Microsoft.Azure.Cosmos from 3.36.0 to 3.37.0.
+ - Updated Bot Framework related libraries from version 4.21.1 to 4.21.2. These libraries are:
+   - Microsoft.Bot.Builder.Azure
+   - Microsoft.Bot.Builder.Azure.Blobs
+   - Microsoft.Bot.Builder.Dialogs
+   - Microsoft.Bot.Builder.Integration.ApplicationInsights.Core
+   - Microsoft.Bot.Builder.Integration.AspNet.Core
+ - Updated library Moq from 4.20.69 to 4.20.70.
+ - Updated library xunit from 2.6.1 to 2.6.2.
+ - Updated library xunit.analyzers from 1.5.0 to 1.6.0.
+ - Updated library xunit.extensibility.core from 2.6.1 to 2.6.2.
+ - Updated library xunit.runner.visualstudio from 2.5.3 to 2.5.4.
+
+## Minor Changes
+ - Some minor tweaks.
+
 ## [6.0.4]
 
 ### Important

Directory.Build.props

@@ -16,7 +16,7 @@
   </PropertyGroup>
 
   <PropertyGroup>
-    <VersionPrefix>8.0.0</VersionPrefix>
+    <VersionPrefix>8.0.1</VersionPrefix>
     <VersionSuffix></VersionSuffix>
   </PropertyGroup>
 

src/Encamina.Enmarcha.SemanticKernel.Abstractions/ILengthFunctions.cs

@@ -5,18 +5,45 @@ namespace Encamina.Enmarcha.SemanticKernel.Abstractions;
 /// <inheritdoc/>
 public interface ILengthFunctions : AI.Abstractions.ILengthFunctions
 {
+    /// <summary>
+    /// Gets the default <see cref="GptEncoding">encoding</see> for models like `GPT-3.5-Turbo` and `GPT-4` from OpenAI.
+    /// </summary>
+    public static readonly GptEncoding DefaultGptEncoding = GptEncoding.GetEncoding("cl100k_base");
+
+    /// <summary>
+    /// Dictionary to cache GptEncoding instances based on encoding names.
+    /// </summary>
+    private static readonly Dictionary<string, GptEncoding> EncodingCache = [];
+
     /// <summary>
     /// Gets the number of tokens using encodings for models like `GPT-3.5-Turbo` and `GPT-4` from OpenAI on the specified text.
     /// If the text is <see langword="null"/> or empty (i.e., <see cref="string.Empty"/>), returns zero (<c>0</c>).
     /// </summary>
     /// <seealso href="https://platform.openai.com/tokenizer"/>
     /// <seealso href="https://github.com/openai/openai-cookbook/blob/main/examples/How_to_count_tokens_with_tiktoken.ipynb"/>
-    public static Func<string, int> LengthByTokenCount => (text) => string.IsNullOrEmpty(text) ? 0 : GptEncoding.GetEncoding("cl100k_base").Encode(text).Count;
+    public static Func<string, int> LengthByTokenCount => (text) => string.IsNullOrEmpty(text) ? 0 : DefaultGptEncoding.Encode(text).Count;
 
     /// <summary>
     /// Gets the number of tokens using a given encoding on the specified text.
     /// If the text is <see langword="null"/> or empty (i.e., <see cref="string.Empty"/>), returns zero (<c>0</c>).
     /// </summary>
     /// <seealso href="https://github.com/openai/openai-cookbook/blob/main/examples/How_to_count_tokens_with_tiktoken.ipynb"/>
-    public static Func<string, string, int> LengthByTokenCountUsingEncoding => (encoding, text) => string.IsNullOrEmpty(text) ? 0 : GptEncoding.GetEncoding(encoding).Encode(text).Count;
+    public static Func<string, string, int> LengthByTokenCountUsingEncoding => (encoding, text) => string.IsNullOrEmpty(text) ? 0 : GetCachedEncoding(encoding).Encode(text).Count;
+
+    /// <summary>
+    /// Gets the GptEncoding instance based on the specified encoding name, caching it for future use.
+    /// </summary>
+    /// <param name="encoding">The name of the GptEncoding.</param>
+    /// <returns>The GptEncoding instance.</returns>
+    private static GptEncoding GetCachedEncoding(string encoding)
+    {
+        if (EncodingCache.TryGetValue(encoding, out var gptEncoding))
+        {
+            return gptEncoding;
+        }
+
+        gptEncoding = GptEncoding.GetEncoding(encoding);
+        EncodingCache[encoding] = gptEncoding;
+        return gptEncoding;
+    }
 }