Add support for Dimension parameter for embeddings (#144)

README.md

@@ -136,7 +136,7 @@ We can also set ChatGPT parameters for chat completion at startup. Check the [of
 
 The configuration can be automatically read from [IConfiguration](https://learn.microsoft.com/en-us/dotnet/api/microsoft.extensions.configuration.iconfiguration), using for example a _ChatGPT_ section in the _appsettings.json_ file:
 
-```yaml
+```
 "ChatGPT": {
  "Provider": "OpenAI", // Optional. Allowed values: OpenAI (default) or Azure
  "ApiKey": "", // Required
@@ -159,6 +159,9 @@ The configuration can be automatically read from [IConfiguration](https://learn.
  // "FrequencyPenalty": 0,
  // "ResponseFormat": { "Type": "text" }, // Allowed values for Type: text (default) or json_object
  // "Seed": 42 // Optional (any integer value)
+ //},
+ //"DefaultEmbeddingParameters": {
+ // "Dimensions": 1536
  //}
 }
 ```
@@ -550,7 +553,52 @@ var response = await chatGptClient.GenerateEmbeddingAsync(message);
 var embeddings = response.GetEmbedding();
 ```
 
-This code will give you a float array containing all the embeddings for the specified message. The length of the array depends on the model used. For example, if we use the _text-embedding-ada-002_ model, the array will contain 1536 elements.
+This code will give you a float array containing all the embeddings for the specified message. The length of the array depends on the model used:
+
+| Model| Output dimension |
+| - | - |
+| text-embedding-ada-002 | 1536 |
+| text-embedding-3-small | 1536 |
+| text-embedding-3-large | 3072 |
+
+Newer models like _text-embedding-3-small_ and _text-embedding-3-large_ allows developers to trade-off performance and cost of using embeddings. Specifically, developers can shorten embeddings without the embedding losing its concept-representing properties.
+
+As for ChatGPT, this settings can be done in various ways:
+
+- Via code:
+
+```csharp
+builder.Services.AddChatGpt(options =>
+{
+ // ...
+
+ options.DefaultEmbeddingParameters = new EmbeddingParameters
+ {
+ Dimensions = 256
+ };
+});
+```
+
+- Using the _appsettings.json_ file:
+
+```
+"ChatGPT": { 
+ "DefaultEmbeddingParameters": {
+ "Dimensions": 256
+ }
+}
+```
+
+Then, if you want to change the dimension for a particular request, you can specify the *EmbeddingParameters* argument in the **GetEmbeddingAsync** invocation:
+
+```csharp
+var response = await chatGptClient.GenerateEmbeddingAsync(request.Message, new EmbeddingParameters
+{
+ Dimensions = 512
+});
+
+var embeddings = response.GetEmbedding(); // The length of the array is 512
+```
 
 If you need to calculate the [cosine similarity](https://en.wikipedia.org/wiki/Cosine_similarity) between two embeddings, you can use the **EmbeddingUtility.CosineSimilarity** method.
 

docs/ChatGptNet.Models.Embeddings/EmbeddingParameters.md

@@ -0,0 +1,25 @@
+# EmbeddingParameters class
+
+Represents embeddings parameters.
+
+```csharp
+public class EmbeddingParameters
+```
+
+## Public Members
+
+| name | description |
+| --- | --- |
+| [EmbeddingParameters](EmbeddingParameters/EmbeddingParameters.md)() | The default constructor. |
+| [Dimensions](EmbeddingParameters/Dimensions.md) { get; set; } | The number of dimensions the resulting output embeddings should have. Only supported in `text-embedding-3` and later models. |
+
+## Remarks
+
+See [Create embeddings](https://platform.openai.com/docs/api-reference/embeddings/create) for more information.
+
+## See Also
+
+* namespace [ChatGptNet.Models.Embeddings](../ChatGptNet.md)
+* [EmbeddingParameters.cs](https://github.com/marcominerva/ChatGptNet/tree/master/src/ChatGptNet/Models/Embeddings/EmbeddingParameters.cs)
+
+<!-- DO NOT EDIT: generated by xmldocmd for ChatGptNet.dll -->

docs/ChatGptNet.Models.Embeddings/EmbeddingParameters/Dimensions.md

@@ -0,0 +1,14 @@
+# EmbeddingParameters.Dimensions property
+
+The number of dimensions the resulting output embeddings should have. Only supported in `text-embedding-3` and later models.
+
+```csharp
+public int? Dimensions { get; set; }
+```
+
+## See Also
+
+* class [EmbeddingParameters](../EmbeddingParameters.md)
+* namespace [ChatGptNet.Models.Embeddings](../../ChatGptNet.md)
+
+<!-- DO NOT EDIT: generated by xmldocmd for ChatGptNet.dll -->

docs/ChatGptNet.Models.Embeddings/EmbeddingParameters/EmbeddingParameters.md

@@ -0,0 +1,14 @@
+# EmbeddingParameters constructor
+
+The default constructor.
+
+```csharp
+public EmbeddingParameters()
+```
+
+## See Also
+
+* class [EmbeddingParameters](../EmbeddingParameters.md)
+* namespace [ChatGptNet.Models.Embeddings](../../ChatGptNet.md)
+
+<!-- DO NOT EDIT: generated by xmldocmd for ChatGptNet.dll -->

docs/ChatGptNet.Models.Embeddings/OpenAIEmbeddingModels.md

@@ -10,7 +10,9 @@ public static class OpenAIEmbeddingModels
 
 | name | description |
 | --- | --- |
-| const [TextEmbeddingAda002](OpenAIEmbeddingModels/TextEmbeddingAda002.md) | The second generation embedding model provided by OpenAI. |
+| const [TextEmbedding3Large](OpenAIEmbeddingModels/TextEmbedding3Large.md) | Most capable embedding model for both english and non-english tasks. It uses a 3072 output dimension. |
+| const [TextEmbedding3Small](OpenAIEmbeddingModels/TextEmbedding3Small.md) | Increased performance over 2nd generation ada embedding model. It uses a 1536 output dimension. |
+| const [TextEmbeddingAda002](OpenAIEmbeddingModels/TextEmbeddingAda002.md) | The second generation embedding model provided by OpenAI. It uses a 1536 output dimension. |
 
 ## Remarks
 

docs/ChatGptNet.Models.Embeddings/OpenAIEmbeddingModels/TextEmbedding3Large.md

@@ -0,0 +1,14 @@
+# OpenAIEmbeddingModels.TextEmbedding3Large field
+
+Most capable embedding model for both english and non-english tasks. It uses a 3072 output dimension.
+
+```csharp
+public const string TextEmbedding3Large;
+```
+
+## See Also
+
+* class [OpenAIEmbeddingModels](../OpenAIEmbeddingModels.md)
+* namespace [ChatGptNet.Models.Embeddings](../../ChatGptNet.md)
+
+<!-- DO NOT EDIT: generated by xmldocmd for ChatGptNet.dll -->

docs/ChatGptNet.Models.Embeddings/OpenAIEmbeddingModels/TextEmbedding3Small.md

@@ -0,0 +1,14 @@
+# OpenAIEmbeddingModels.TextEmbedding3Small field
+
+Increased performance over 2nd generation ada embedding model. It uses a 1536 output dimension.
+
+```csharp
+public const string TextEmbedding3Small;
+```
+
+## See Also
+
+* class [OpenAIEmbeddingModels](../OpenAIEmbeddingModels.md)
+* namespace [ChatGptNet.Models.Embeddings](../../ChatGptNet.md)
+
+<!-- DO NOT EDIT: generated by xmldocmd for ChatGptNet.dll -->

docs/ChatGptNet.Models.Embeddings/OpenAIEmbeddingModels/TextEmbeddingAda002.md

@@ -1,6 +1,6 @@
 # OpenAIEmbeddingModels.TextEmbeddingAda002 field
 
-The second generation embedding model provided by OpenAI.
+The second generation embedding model provided by OpenAI. It uses a 1536 output dimension.
 
 ```csharp
 public const string TextEmbeddingAda002;

docs/ChatGptNet/ChatGptOptions.md

@@ -12,6 +12,7 @@ public class ChatGptOptions
 | --- | --- |
 | [ChatGptOptions](ChatGptOptions/ChatGptOptions.md)() | The default constructor. |
 | [DefaultEmbeddingModel](ChatGptOptions/DefaultEmbeddingModel.md) { get; set; } | Gets or sets the default model for embedding. (default: [`TextEmbeddingAda002`](../ChatGptNet.Models.Embeddings/OpenAIEmbeddingModels/TextEmbeddingAda002.md) when the provider is OpenAI). |
+| [DefaultEmbeddingParameters](ChatGptOptions/DefaultEmbeddingParameters.md) { get; } | Gets or sets the default parameters for embeddings. |
 | [DefaultModel](ChatGptOptions/DefaultModel.md) { get; set; } | Gets or sets the default model for chat completion. (default: [`Gpt35Turbo`](../ChatGptNet.Models/OpenAIChatGptModels/Gpt35Turbo.md) when the provider is OpenAI). |
 | [DefaultParameters](ChatGptOptions/DefaultParameters.md) { get; } | Gets or sets the default parameters for chat completion. |
 | [MessageExpiration](ChatGptOptions/MessageExpiration.md) { get; set; } | Gets or sets the expiration for cached conversation messages (default: 1 hour). |

docs/ChatGptNet/ChatGptOptions/DefaultEmbeddingParameters.md

@@ -0,0 +1,15 @@
+# ChatGptOptions.DefaultEmbeddingParameters property
+
+Gets or sets the default parameters for embeddings.
+
+```csharp
+public EmbeddingParameters DefaultEmbeddingParameters { get; }
+```
+
+## See Also
+
+* class [EmbeddingParameters](../../ChatGptNet.Models.Embeddings/EmbeddingParameters.md)
+* class [ChatGptOptions](../ChatGptOptions.md)
+* namespace [ChatGptNet](../../ChatGptNet.md)
+
+<!-- DO NOT EDIT: generated by xmldocmd for ChatGptNet.dll -->

docs/ChatGptNet/ChatGptOptionsBuilder.md

@@ -11,7 +11,8 @@ public class ChatGptOptionsBuilder
 | name | description |
 | --- | --- |
 | [ChatGptOptionsBuilder](ChatGptOptionsBuilder/ChatGptOptionsBuilder.md)() | The default constructor. |
-| [DefaultEmbeddingModel](ChatGptOptionsBuilder/DefaultEmbeddingModel.md) { get; set; } | Gets or sets the default model for embedding. (default: [`TextEmbeddingAda002`](../ChatGptNet.Models.Embeddings/OpenAIEmbeddingModels/TextEmbeddingAda002.md) when the provider is OpenAI). |
+| [DefaultEmbeddingModel](ChatGptOptionsBuilder/DefaultEmbeddingModel.md) { get; set; } | Gets or sets the default model for embeddings. (default: [`TextEmbeddingAda002`](../ChatGptNet.Models.Embeddings/OpenAIEmbeddingModels/TextEmbeddingAda002.md) when the provider is OpenAI). |
+| [DefaultEmbeddingParameters](ChatGptOptionsBuilder/DefaultEmbeddingParameters.md) { get; } | Gets or sets the default parameters for embeddings. |
 | [DefaultModel](ChatGptOptionsBuilder/DefaultModel.md) { get; set; } | Gets or sets the default model for chat completion. (default: [`Gpt35Turbo`](../ChatGptNet.Models/OpenAIChatGptModels/Gpt35Turbo.md) when the provider is OpenAI). |
 | [DefaultParameters](ChatGptOptionsBuilder/DefaultParameters.md) { get; set; } | Gets or sets the default parameters for chat completion. |
 | [MessageExpiration](ChatGptOptionsBuilder/MessageExpiration.md) { get; set; } | Gets or sets the expiration for cached conversation messages (default: 1 hour). |

docs/ChatGptNet/ChatGptOptionsBuilder/DefaultEmbeddingModel.md

@@ -1,6 +1,6 @@
 # ChatGptOptionsBuilder.DefaultEmbeddingModel property
 
-Gets or sets the default model for embedding. (default: [`TextEmbeddingAda002`](../../ChatGptNet.Models.Embeddings/OpenAIEmbeddingModels/TextEmbeddingAda002.md) when the provider is OpenAI).
+Gets or sets the default model for embeddings. (default: [`TextEmbeddingAda002`](../../ChatGptNet.Models.Embeddings/OpenAIEmbeddingModels/TextEmbeddingAda002.md) when the provider is OpenAI).
 
 ```csharp
 public string? DefaultEmbeddingModel { get; set; }

docs/ChatGptNet/ChatGptOptionsBuilder/DefaultEmbeddingParameters.md

@@ -0,0 +1,15 @@
+# ChatGptOptionsBuilder.DefaultEmbeddingParameters property
+
+Gets or sets the default parameters for embeddings.
+
+```csharp
+public EmbeddingParameters DefaultEmbeddingParameters { get; }
+```
+
+## See Also
+
+* class [EmbeddingParameters](../../ChatGptNet.Models.Embeddings/EmbeddingParameters.md)
+* class [ChatGptOptionsBuilder](../ChatGptOptionsBuilder.md)
+* namespace [ChatGptNet](../../ChatGptNet.md)
+
+<!-- DO NOT EDIT: generated by xmldocmd for ChatGptNet.dll -->

docs/ChatGptNet/IChatGptClient.md

@@ -16,7 +16,7 @@ public interface IChatGptClient
 | [AskStreamAsync](IChatGptClient/AskStreamAsync.md)(…) | Requests a new chat interaction with streaming response, like in ChatGPT. (2 methods) |
 | [ConversationExistsAsync](IChatGptClient/ConversationExistsAsync.md)(…) | Checks if a chat conversation exists. |
 | [DeleteConversationAsync](IChatGptClient/DeleteConversationAsync.md)(…) | Deletes a chat conversation, clearing all the history. |
-| [GenerateEmbeddingAsync](IChatGptClient/GenerateEmbeddingAsync.md)(…) | Generates embeddings for a message. (2 methods) |
+| [GenerateEmbeddingAsync](IChatGptClient/GenerateEmbeddingAsync.md)(…) | Generates embeddings for a text. (2 methods) |
 | [GetConversationAsync](IChatGptClient/GetConversationAsync.md)(…) | Retrieves a chat conversation from the cache. |
 | [LoadConversationAsync](IChatGptClient/LoadConversationAsync.md)(…) | Loads messages into a new conversation. (2 methods) |
 | [SetupAsync](IChatGptClient/SetupAsync.md)(…) | Setups a new conversation with a system message, that is used to influence assistant behavior. (2 methods) |

docs/ChatGptNet/IChatGptClient/GenerateEmbeddingAsync.md

@@ -1,15 +1,17 @@
 # IChatGptClient.GenerateEmbeddingAsync method (1 of 2)
 
-Generates embeddings for a list of messages.
+Generates embeddings for a list of texts.
 
 ```csharp
-public Task<EmbeddingResponse> GenerateEmbeddingAsync(IEnumerable<string> messages, 
- string? model = null, CancellationToken cancellationToken = default)
+public Task<EmbeddingResponse> GenerateEmbeddingAsync(IEnumerable<string> texts, 
+ EmbeddingParameters? parameters = null, string? model = null, 
+ CancellationToken cancellationToken = default)
 ```
 
 | parameter | description |
 | --- | --- |
-| messages | The messages to use for generating embeddings. |
+| texts | The texts to use for generating embeddings. |
+| parameters | An [`EmbeddingParameters`](../../ChatGptNet.Models.Embeddings/EmbeddingParameters.md) object used to override the default embedding parameters in the [`DefaultEmbeddingParameters`](../ChatGptOptions/DefaultEmbeddingParameters.md) property. |
 | model | The name of the embedding model. If *model* is `null`, then the one specified in the [`DefaultEmbeddingModel`](../ChatGptOptions/DefaultEmbeddingModel.md) property will be used. |
 | cancellationToken | The token to monitor for cancellation requests. |
 
@@ -26,23 +28,26 @@ The embeddings for the provided messages.
 ## See Also
 
 * class [EmbeddingResponse](../../ChatGptNet.Models.Embeddings/EmbeddingResponse.md)
+* class [EmbeddingParameters](../../ChatGptNet.Models.Embeddings/EmbeddingParameters.md)
 * interface [IChatGptClient](../IChatGptClient.md)
 * namespace [ChatGptNet](../../ChatGptNet.md)
 
 ---
 
 # IChatGptClient.GenerateEmbeddingAsync method (2 of 2)
 
-Generates embeddings for a message.
+Generates embeddings for a text.
 
 ```csharp
-public Task<EmbeddingResponse> GenerateEmbeddingAsync(string message, string? model = null, 
+public Task<EmbeddingResponse> GenerateEmbeddingAsync(string text, 
+ EmbeddingParameters? parameters = null, string? model = null, 
  CancellationToken cancellationToken = default)
 ```
 
 | parameter | description |
 | --- | --- |
-| message | The message to use for generating embeddings. |
+| text | The text to use for generating embeddings. |
+| parameters | An [`EmbeddingParameters`](../../ChatGptNet.Models.Embeddings/EmbeddingParameters.md) object used to override the default embedding parameters in the [`DefaultEmbeddingParameters`](../ChatGptOptions/DefaultEmbeddingParameters.md) property. |
 | model | The name of the embedding model. If *model* is `null`, then the one specified in the [`DefaultEmbeddingModel`](../ChatGptOptions/DefaultEmbeddingModel.md) property will be used. |
 | cancellationToken | The token to monitor for cancellation requests. |
 
@@ -59,6 +64,7 @@ The embeddings for the provided message.
 ## See Also
 
 * class [EmbeddingResponse](../../ChatGptNet.Models.Embeddings/EmbeddingResponse.md)
+* class [EmbeddingParameters](../../ChatGptNet.Models.Embeddings/EmbeddingParameters.md)
 * interface [IChatGptClient](../IChatGptClient.md)
 * namespace [ChatGptNet](../../ChatGptNet.md)
 

docs/README.md

@@ -71,6 +71,7 @@
 | public type | description |
 | --- | --- |
 | class [EmbeddingData](./ChatGptNet.Models.Embeddings/EmbeddingData.md) | Represents an embedding. |
+| class [EmbeddingParameters](./ChatGptNet.Models.Embeddings/EmbeddingParameters.md) | Represents embeddings parameters. |
 | class [EmbeddingResponse](./ChatGptNet.Models.Embeddings/EmbeddingResponse.md) | Represents an embedding response. |
 | static class [OpenAIEmbeddingModels](./ChatGptNet.Models.Embeddings/OpenAIEmbeddingModels.md) | Contains all the embedding models that are currently supported by OpenAI. |
 

samples/ChatGptApi/Program.cs

@@ -117,7 +117,7 @@ await foreach (var delta in responseStream.AsDeltas())
 })
 .WithOpenApi();
 
-app.MapPost("/api/embeddings/CosineSimilarity", async (CosineSimilarityRequest request, IChatGptClient chatGptClient) =>
+app.MapPost("/api/embeddings/cosine-similarity", async (CosineSimilarityRequest request, IChatGptClient chatGptClient) =>
 {
  var firstEmbeddingResponse = await chatGptClient.GenerateEmbeddingAsync(request.FirstMessage);
  var secondEmbeddingResponse = await chatGptClient.GenerateEmbeddingAsync(request.SecondMessage);

samples/ChatGptApi/appsettings.json

@@ -1,17 +1,17 @@
 {
  "ChatGPT": {
- "Provider": "OpenAI", // Optional. Allowed values: OpenAI (default) or Azure
+ "Provider": "OpenaI", // Optional. Allowed values: OpenAI (default) or Azure
  "ApiKey": "", // Required
- //"Organization": "",  // Optional, used only by OpenAI
+ //"Organization": "", // Optional, used only by OpenAI
  "ResourceName": "", // Required when using Azure OpenAI Service
- "ApiVersion": "2023-12-01-preview", // Optional, used only by Azure OpenAI Service (default: 2023-12-01-preview)
+ "ApiVersion": "2023-12-01-preview", // Optional, used only by Azure OpenAI Service (default: 2023-08-01-preview)
  "AuthenticationType": "ApiKey", // Optional, used only by Azure OpenAI Service. Allowed values: ApiKey (default) or ActiveDirectory
 
  "DefaultModel": "my-model",
  "DefaultEmbeddingModel": "text-embedding-ada-002", // Optional, set it if you want to use embeddings
  "MessageLimit": 20,
  "MessageExpiration": "00:30:00",
- "ThrowExceptionOnError": true, // Optional, default: true
+ "ThrowExceptionOnError": true // Optional, default: true
  //"User": "UserName",
  //"DefaultParameters": {
  // "Temperature": 0.8,
@@ -21,6 +21,9 @@
  // "FrequencyPenalty": 0,
  // "ResponseFormat": { "Type": "text" }, // Allowed values for Type: text (default) or json_object
  // "Seed": 42 // Optional (any integer value)
+ //},
+ //"DefaultEmbeddingParameters": {
+ // "Dimensions": 1536
  //}
  },
  "Logging": {

samples/ChatGptConsole/appsettings.json

@@ -11,7 +11,7 @@
  "DefaultEmbeddingModel": "text-embedding-ada-002", // Optional, it set if you want to use embeddings
  "MessageLimit": 20,
  "MessageExpiration": "00:30:00",
- "ThrowExceptionOnError": true,
+ "ThrowExceptionOnError": true
  //"User": "UserName",
  //"DefaultParameters": {
  // "Temperature": 0.8,
@@ -21,6 +21,9 @@
  // "FrequencyPenalty": 0,
  // "ResponseFormat": { "Type": "text" }, // Allowed values for Type: text (default) or json_object
  // "Seed": 42 // Optional (any integer value)
+ //},
+ //"DefaultEmbeddingParameters": {
+ // "Dimensions": 1536
  //}
  },
  "Logging": {

samples/ChatGptFunctionCallingConsole/appsettings.json

@@ -11,7 +11,7 @@
  "DefaultEmbeddingModel": "text-embedding-ada-002", // Optional, it set if you want to use embeddings
  "MessageLimit": 20,
  "MessageExpiration": "00:30:00",
- "ThrowExceptionOnError": true, // Optional, default: true
+ "ThrowExceptionOnError": true // Optional, default: true
  //"User": "UserName",
  //"DefaultParameters": {
  // "Temperature": 0.8,
@@ -21,6 +21,9 @@
  // "FrequencyPenalty": 0,
  // "ResponseFormat": { "Type": "text" }, // Allowed values for Type: text (default) or json_object
  // "Seed": 42 // Optional (any integer value)
+ //},
+ //"DefaultEmbeddingParameters": {
+ // "Dimensions": 1536
  //}
  },
  "Logging": {