Added embedding API (#654)
Build and Release / Read metadata (push) Waiting to run
Build and Release / Build app (${{ matrix.dotnet_runtime }}) (-aarch64-apple-darwin, osx-arm64, macos-latest, aarch64-apple-darwin, dmg updater) (push) Blocked by required conditions
Build and Release / Build app (${{ matrix.dotnet_runtime }}) (-aarch64-pc-windows-msvc.exe, win-arm64, windows-latest, aarch64-pc-windows-msvc, nsis updater) (push) Blocked by required conditions
Build and Release / Build app (${{ matrix.dotnet_runtime }}) (-aarch64-unknown-linux-gnu, linux-arm64, ubuntu-22.04-arm, aarch64-unknown-linux-gnu, appimage deb updater) (push) Blocked by required conditions
Build and Release / Build app (${{ matrix.dotnet_runtime }}) (-x86_64-apple-darwin, osx-x64, macos-latest, x86_64-apple-darwin, dmg updater) (push) Blocked by required conditions
Build and Release / Build app (${{ matrix.dotnet_runtime }}) (-x86_64-pc-windows-msvc.exe, win-x64, windows-latest, x86_64-pc-windows-msvc, nsis updater) (push) Blocked by required conditions
Build and Release / Build app (${{ matrix.dotnet_runtime }}) (-x86_64-unknown-linux-gnu, linux-x64, ubuntu-22.04, x86_64-unknown-linux-gnu, appimage deb updater) (push) Blocked by required conditions
Build and Release / Prepare & create release (push) Blocked by required conditions
Build and Release / Publish release (push) Blocked by required conditions

Co-authored-by: Thorsten Sommer <SommerEngineering@users.noreply.github.com>
This commit is contained in:
Paul KoudelkaandThorsten Sommer authored and GitHub committed 2026-02-20 15:32:54 +01:00
1 parent af72a45035
commit a8951fe58a
32 files changed
+529 -6

No files matched your search

@@ -86,6 +86,13 @@ public sealed class ProviderAlibabaCloud() : BaseProvider(LLMProviders.ALIBABA_C
{
return Task.FromResult(string.Empty);
}
/// <inhertidoc />
public override async Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
var requestedSecret = await RUST_SERVICE.GetAPIKey(this, SecretStoreType.EMBEDDING_PROVIDER);
return await this.PerformStandardTextEmbeddingRequest(requestedSecret, embeddingModel, token: token, texts: texts);
}
/// <inheritdoc />
public override Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
@@ -113,6 +113,12 @@ public sealed class ProviderAnthropic() : BaseProvider(LLMProviders.ANTHROPIC, "
{
return Task.FromResult(string.Empty);
}
/// <inhertidoc />
public override Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
return Task.FromResult<IReadOnlyList<IReadOnlyList<float>>>([]);
}
/// <inheritdoc />
public override Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
@@ -1,6 +1,7 @@
using System.Net;
using System.Net.Http.Headers;
using System.Runtime.CompilerServices;
using System.Text;
using System.Text.Json;
using System.Text.Json.Serialization;
@@ -98,6 +99,9 @@ public abstract class BaseProvider : IProvider, ISecretId
/// <inheritdoc />
public abstract Task<string> TranscribeAudioAsync(Model transcriptionModel, string audioFilePath, SettingsManager settingsManager, CancellationToken token = default);
/// <inheritdoc />
public abstract Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts);
/// <inheritdoc />
public abstract Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default);
@@ -645,6 +649,81 @@ public abstract class BaseProvider : IProvider, ISecretId
}
}
protected async Task<IReadOnlyList<IReadOnlyList<float>>> PerformStandardTextEmbeddingRequest(RequestedSecret requestedSecret, Model embeddingModel, Host host = Host.NONE, CancellationToken token = default, params List<string> texts)
{
try
{
//
// Add the model name to the form data. Ensure that a model name is always provided.
// Otherwise, the StringContent constructor will throw an exception.
//
var modelName = embeddingModel.Id;
if (string.IsNullOrWhiteSpace(modelName))
modelName = "placeholder";
// Prepare the HTTP embedding request:
var payload = new
{
model = modelName,
input = texts,
encoding_format = "float"
};
var embeddingRequest = JsonSerializer.Serialize(payload, JSON_SERIALIZER_OPTIONS);
using var request = new HttpRequestMessage(HttpMethod.Post, host.EmbeddingURL());
// Handle the authorization header based on the provider:
switch (this.Provider)
{
case LLMProviders.SELF_HOSTED:
if(requestedSecret.Success)
request.Headers.Authorization = new AuthenticationHeaderValue("Bearer", await requestedSecret.Secret.Decrypt(ENCRYPTION));
break;
default:
if(!requestedSecret.Success)
{
this.logger.LogError("No valid API key available for embedding request.");
return [];
}
request.Headers.Authorization = new AuthenticationHeaderValue("Bearer", await requestedSecret.Secret.Decrypt(ENCRYPTION));
break;
}
// Set the content:
request.Content = new StringContent(embeddingRequest, Encoding.UTF8, "application/json");
using var response = await this.httpClient.SendAsync(request, token);
var responseBody = response.Content.ReadAsStringAsync(token).Result;
if (!response.IsSuccessStatusCode)
{
this.logger.LogError("Embedding request failed with status code {ResponseStatusCode} and body: '{ResponseBody}'.", response.StatusCode, responseBody);
return [];
}
var embeddingResponse = JsonSerializer.Deserialize<EmbeddingResponse>(responseBody, JSON_SERIALIZER_OPTIONS);
if (embeddingResponse is { Data: not null })
{
return embeddingResponse.Data
.Select(d => d.Embedding?.ToArray() ?? [])
.Cast<IReadOnlyList<float>>()
.ToArray();
}
else
{
this.logger.LogError("Was not able to deserialize the embedding response.");
return [];
}
}
catch (Exception e)
{
this.logger.LogError("Failed to perform embedding request: '{Message}'.", e.Message);
return [];
}
}
/// <summary>
/// Parse and convert API parameters from a provided JSON string into a dictionary,
/// optionally merging additional parameters and removing specific keys.
@@ -86,6 +86,12 @@ public sealed class ProviderDeepSeek() : BaseProvider(LLMProviders.DEEP_SEEK, "h
{
return Task.FromResult(string.Empty);
}
/// <inhertidoc />
public override Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
return Task.FromResult<IReadOnlyList<IReadOnlyList<float>>>([]);
}
/// <inheritdoc />
public override Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
@@ -0,0 +1,12 @@
// ReSharper disable CollectionNeverUpdated.Global
namespace AIStudio.Provider;
// ReSharper disable once ClassNeverInstantiated.Global
public sealed record EmbeddingData
{
public string? Object { get; set; }
public List<float>? Embedding { get; set; }
public int? Index { get; set; }
}
@@ -0,0 +1,14 @@
namespace AIStudio.Provider;
public sealed record EmbeddingResponse
{
public string? Id { get; init; }
public string? Object { get; init; }
public List<EmbeddingData>? Data { get; init; }
public string? Model { get; init; }
public EmbeddingUsage? Usage { get; init; }
}
@@ -0,0 +1,11 @@
// ReSharper disable ClassNeverInstantiated.Global
namespace AIStudio.Provider;
public sealed record EmbeddingUsage
{
public int? PromptTokens { get; set; }
public int? TotalTokens { get; set; }
public int? CompletionTokens { get; set; }
}
@@ -88,6 +88,12 @@ public class ProviderFireworks() : BaseProvider(LLMProviders.FIREWORKS, "https:/
var requestedSecret = await RUST_SERVICE.GetAPIKey(this, SecretStoreType.TRANSCRIPTION_PROVIDER);
return await this.PerformStandardTranscriptionRequest(requestedSecret, transcriptionModel, audioFilePath, token: token);
}
/// <inhertidoc />
public override Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
return Task.FromResult<IReadOnlyList<IReadOnlyList<float>>>([]);
}
/// <inheritdoc />
public override Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
@@ -87,6 +87,12 @@ public sealed class ProviderGWDG() : BaseProvider(LLMProviders.GWDG, "https://ch
var requestedSecret = await RUST_SERVICE.GetAPIKey(this, SecretStoreType.TRANSCRIPTION_PROVIDER);
return await this.PerformStandardTranscriptionRequest(requestedSecret, transcriptionModel, audioFilePath, token: token);
}
/// <inhertidoc />
public override Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
return Task.FromResult<IReadOnlyList<IReadOnlyList<float>>>([]);
}
/// <inheritdoc />
public override async Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
@@ -0,0 +1,6 @@
namespace AIStudio.Provider.Google;
public sealed record GoogleEmbedding
{
public List<float>? Values { get; init; }
}
@@ -0,0 +1,30 @@
using System.Text.Json;
using System.Text.Json.Serialization;
namespace AIStudio.Provider.Google;
public sealed record GoogleEmbeddingResponse
{
[JsonConverter(typeof(GoogleEmbeddingListConverter))]
public List<GoogleEmbedding>? Embedding { get; init; }
private sealed class GoogleEmbeddingListConverter : JsonConverter<List<GoogleEmbedding>>
{
public override List<GoogleEmbedding> Read(ref Utf8JsonReader reader, Type typeToConvert, JsonSerializerOptions options)
{
if (reader.TokenType == JsonTokenType.StartObject)
{
var single = JsonSerializer.Deserialize<GoogleEmbedding>(ref reader, options);
return single is null ? new() : new() { single };
}
if (reader.TokenType == JsonTokenType.StartArray)
return JsonSerializer.Deserialize<List<GoogleEmbedding>>(ref reader, options) ?? new();
throw new JsonException("Expected object or array for embedding.");
}
public override void Write(Utf8JsonWriter writer, List<GoogleEmbedding> value, JsonSerializerOptions options) =>
JsonSerializer.Serialize(writer, value, options);
}
}
@@ -87,6 +87,78 @@ public class ProviderGoogle() : BaseProvider(LLMProviders.GOOGLE, "https://gener
{
return Task.FromResult(string.Empty);
}
/// <inhertidoc />
public override async Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
var requestedSecret = await RUST_SERVICE.GetAPIKey(this, SecretStoreType.EMBEDDING_PROVIDER);
try
{
var modelName = embeddingModel.Id;
if (string.IsNullOrWhiteSpace(modelName))
{
LOGGER.LogError("No model name provided for embedding request.");
return [];
}
if (modelName.StartsWith("models/", StringComparison.OrdinalIgnoreCase))
modelName = modelName.Substring("models/".Length);
if (!requestedSecret.Success)
{
LOGGER.LogError("No valid API key available for embedding request.");
return [];
}
// Prepare the Google Gemini embedding request:
var payload = new
{
content = new
{
parts = texts.Select(text => new { text }).ToArray()
},
taskType = "SEMANTIC_SIMILARITY"
};
var embeddingRequest = JsonSerializer.Serialize(payload, JSON_SERIALIZER_OPTIONS);
var embedUrl = $"https://generativelanguage.googleapis.com/v1beta/models/{modelName}:embedContent";
using var request = new HttpRequestMessage(HttpMethod.Post, embedUrl);
request.Headers.Add("x-goog-api-key", await requestedSecret.Secret.Decrypt(ENCRYPTION));
// Set the content:
request.Content = new StringContent(embeddingRequest, Encoding.UTF8, "application/json");
using var response = await this.httpClient.SendAsync(request, token);
var responseBody = await response.Content.ReadAsStringAsync(token);
if (!response.IsSuccessStatusCode)
{
LOGGER.LogError("Embedding request failed with status code {ResponseStatusCode} and body: '{ResponseBody}'.", response.StatusCode, responseBody);
return [];
}
var embeddingResponse = JsonSerializer.Deserialize<GoogleEmbeddingResponse>(responseBody, JSON_SERIALIZER_OPTIONS);
if (embeddingResponse is { Embedding: not null })
{
return embeddingResponse.Embedding
.Select(d => d.Values?.ToArray() ?? [])
.Cast<IReadOnlyList<float>>()
.ToArray();
}
else
{
LOGGER.LogError("Was not able to deserialize the embedding response.");
return [];
}
}
catch (Exception e)
{
LOGGER.LogError("Failed to perform embedding request: '{Message}'.", e.Message);
return [];
}
}
/// <inheritdoc />
public override async Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
@@ -87,6 +87,12 @@ public class ProviderGroq() : BaseProvider(LLMProviders.GROQ, "https://api.groq.
{
return Task.FromResult(string.Empty);
}
/// <inhertidoc />
public override Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
return Task.FromResult<IReadOnlyList<IReadOnlyList<float>>>([]);
}
/// <inheritdoc />
public override Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
@@ -86,6 +86,13 @@ public sealed class ProviderHelmholtz() : BaseProvider(LLMProviders.HELMHOLTZ, "
{
return Task.FromResult(string.Empty);
}
/// <inhertidoc />
public override async Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
var requestedSecret = await RUST_SERVICE.GetAPIKey(this, SecretStoreType.EMBEDDING_PROVIDER);
return await this.PerformStandardTextEmbeddingRequest(requestedSecret, embeddingModel, token: token, texts: texts);
}
/// <inheritdoc />
public override async Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
@@ -91,6 +91,12 @@ public sealed class ProviderHuggingFace : BaseProvider
{
return Task.FromResult(string.Empty);
}
/// <inhertidoc />
public override Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
return Task.FromResult<IReadOnlyList<IReadOnlyList<float>>>([]);
}
/// <inheritdoc />
public override Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
@@ -59,6 +59,16 @@ public interface IProvider
/// <param name="token">The cancellation token.</param>
/// <returns>>The transcription result.</returns>
public Task<string> TranscribeAudioAsync(Model transcriptionModel, string audioFilePath, SettingsManager settingsManager, CancellationToken token = default);
/// <summary>
/// Embed a text file.
/// </summary>
/// <param name="embeddingModel">The model to use for embedding.</param>
/// <param name="settingsManager">The settings manager instance to use.</param>
/// <param name="token">The cancellation token.</param>
/// /// <param name="texts">A single string or a list of strings to embed.</param>
/// <returns>>The embedded text as a single vector or as a list of vectors.</returns>
public Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts);
/// <summary>
/// Load all possible text models that can be used with this provider.
@@ -89,6 +89,13 @@ public sealed class ProviderMistral() : BaseProvider(LLMProviders.MISTRAL, "http
return await this.PerformStandardTranscriptionRequest(requestedSecret, transcriptionModel, audioFilePath, token: token);
}
/// <inhertidoc />
public override async Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Provider.Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
var requestedSecret = await RUST_SERVICE.GetAPIKey(this, SecretStoreType.EMBEDDING_PROVIDER);
return await this.PerformStandardTextEmbeddingRequest(requestedSecret, embeddingModel, token: token, texts: texts);
}
/// <inheritdoc />
public override async Task<IEnumerable<Provider.Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
{
@@ -40,6 +40,8 @@ public class NoProvider : IProvider
public Task<string> TranscribeAudioAsync(Model transcriptionModel, string audioFilePath, SettingsManager settingsManager, CancellationToken token = default) => Task.FromResult(string.Empty);
public Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts) => Task.FromResult<IReadOnlyList<IReadOnlyList<float>>>([]);
public IReadOnlyCollection<Capability> GetModelCapabilities(Model model) => [ Capability.NONE ];
#endregion
@@ -224,6 +224,13 @@ public sealed class ProviderOpenAI() : BaseProvider(LLMProviders.OPEN_AI, "https
var requestedSecret = await RUST_SERVICE.GetAPIKey(this, SecretStoreType.TRANSCRIPTION_PROVIDER);
return await this.PerformStandardTranscriptionRequest(requestedSecret, transcriptionModel, audioFilePath, token: token);
}
/// <inhertidoc />
public override async Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
var requestedSecret = await RUST_SERVICE.GetAPIKey(this, SecretStoreType.EMBEDDING_PROVIDER);
return await this.PerformStandardTextEmbeddingRequest(requestedSecret, embeddingModel, token: token, texts: texts);
}
/// <inheritdoc />
public override async Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
@@ -94,6 +94,13 @@ public sealed class ProviderOpenRouter() : BaseProvider(LLMProviders.OPEN_ROUTER
{
return Task.FromResult(string.Empty);
}
/// <inhertidoc />
public override async Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
var requestedSecret = await RUST_SERVICE.GetAPIKey(this, SecretStoreType.EMBEDDING_PROVIDER);
return await this.PerformStandardTextEmbeddingRequest(requestedSecret, embeddingModel, token: token, texts: texts);
}
/// <inheritdoc />
public override Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
@@ -94,6 +94,12 @@ public sealed class ProviderPerplexity() : BaseProvider(LLMProviders.PERPLEXITY,
{
return Task.FromResult(string.Empty);
}
/// <inhertidoc />
public override Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
return Task.FromResult<IReadOnlyList<IReadOnlyList<float>>>([]);
}
/// <inheritdoc />
public override Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
@@ -30,6 +30,11 @@ public static class HostExtensions
_ => "audio/transcriptions",
};
public static string EmbeddingURL(this Host host) => host switch
{
_ => "embeddings",
};
public static bool IsChatSupported(this Host host)
{
switch (host)
@@ -95,6 +95,13 @@ public sealed class ProviderSelfHosted(Host host, string hostname) : BaseProvide
return await this.PerformStandardTranscriptionRequest(requestedSecret, transcriptionModel, audioFilePath, host, token);
}
/// <inhertidoc />
public override async Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Provider.Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
var requestedSecret = await RUST_SERVICE.GetAPIKey(this, SecretStoreType.EMBEDDING_PROVIDER);
return await this.PerformStandardTextEmbeddingRequest(requestedSecret, embeddingModel, token: token, texts: texts);
}
public override async Task<IEnumerable<Provider.Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
{
try
@@ -88,6 +88,12 @@ public sealed class ProviderX() : BaseProvider(LLMProviders.X, "https://api.x.ai
return Task.FromResult(string.Empty);
}
/// <inhertidoc />
public override Task<IReadOnlyList<IReadOnlyList<float>>> EmbedTextAsync(Model embeddingModel, SettingsManager settingsManager, CancellationToken token = default, params List<string> texts)
{
return Task.FromResult<IReadOnlyList<IReadOnlyList<float>>>([]);
}
/// <inheritdoc />
public override async Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
{