MindWorkAI/AI-Studio
MindWorkAI/AI-Studio
1
0
Fork 0
mirror of https://github.com/MindWorkAI/AI-Studio.git synced 2026-08-11 23:12:10 +00:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

current main is now successfully merged into branch

Browse Source
This commit is contained in:
Paul Koudelka 2026-05-27 20:02:43 +02:00
parent 00b6c98f3b
commit ea87d79ca9
32 changed files with 607 additions and 478 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

52
app/MindWork AI Studio/Assistants/I18N/allTexts.lua
View File

@ -4905,6 +4905,7 @@ UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T146957
-- Refresh all
UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T1503082343"] = "Refresh all"
-- Kerberos/SSO ERI data sources cannot be exported yet. Please configure them manually in the configuration plugin.
UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T1577531115"] = "Kerberos/SSO ERI data sources cannot be exported yet. Please configure them manually in the configuration plugin."
@ -4968,11 +4969,12 @@ UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T352566
-- No data sources configured yet.
UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T3549650120"] = "No data sources configured yet."
-- Local data sources refresh when files change.
UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T3687976654"] = "Local data sources refresh when files change."
-- Export Access Token?
UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T3595669127"] = "Export Access Token?"
-- Local data sources refresh when files change.
UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T3687976654"] = "Local data sources refresh when files change."
-- Export ERI Data Source
UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T3831281036"] = "Export ERI Data Source"
@ -4994,11 +4996,12 @@ UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T750361
-- External Data (ERI-Server v1)
UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T774473996"] = "External Data (ERI-Server v1)"
-- Local data sources refresh only when triggered manually.
UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T854231603"] = "Local data sources refresh only when triggered manually."
-- Cannot export this ERI data source because no authentication secret is configured. The issue was: {0}
UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T782820095"] = "Cannot export this ERI data source because no authentication secret is configured. The issue was: {0}"
-- Local data sources refresh only when triggered manually.
UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T854231603"] = "Local data sources refresh only when triggered manually."
-- Local Directory
UI_TEXT_CONTENT["AISTUDIO::DIALOGS::SETTINGS::SETTINGSDIALOGDATASOURCES::T926703547"] = "Local Directory"
@ -6130,6 +6133,9 @@ UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T1019424746"] = "Startup log file
-- Browse AI Studio's source code on GitHub — we welcome your contributions.
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T1107156991"] = "Browse AI Studio's source code on GitHub — we welcome your contributions."
-- Vector store version
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T1124039623"] = "Vector store version"
-- The Tokenizer library serves as the base framework for integrating the DeepSeek tokenizer.
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T1132433749"] = "The Tokenizer library serves as the base framework for integrating the DeepSeek tokenizer."
@ -6145,9 +6151,6 @@ UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T1290340974"] = "Unknown configur
-- This library is used to read PDF files. This is necessary, e.g., for using PDFs as a data source for a chat.
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T1388816916"] = "This library is used to read PDF files. This is necessary, e.g., for using PDFs as a data source for a chat."
-- Database version
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T1420062548"] = "Database version"
-- This library is used to extend the MudBlazor library. It provides additional components that are not part of the MudBlazor library.
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T1421513382"] = "This library is used to extend the MudBlazor library. It provides additional components that are not part of the MudBlazor library."
@ -6295,6 +6298,9 @@ UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T2924964415"] = "AI Studio runs w
-- Changelog
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T3017574265"] = "Changelog"
-- Vector store
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T3046399223"] = "Vector store"
-- Enterprise configuration ID:
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T3092349641"] = "Enterprise configuration ID:"
@ -6370,9 +6376,6 @@ UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T3986423270"] = "Check Pandoc Ins
-- Versions
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T4010195468"] = "Versions"
-- Database
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T4036243672"] = "Database"
-- This library is used by the Rust runtime to read the current user's username, e.g. when an organization-managed ERI server uses the OS username for authentication.
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T4060906280"] = "This library is used by the Rust runtime to read the current user's username, e.g. when an organization-managed ERI server uses the OS username for authentication."
@ -6396,6 +6399,7 @@ UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T585329785"] = "Used .NET SDK"
-- We use the DeepSeek Tokenizer to estimate the number of tokens an input will generate.
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T591393704"] = "We use the DeepSeek Tokenizer to estimate the number of tokens an input will generate."
-- starting
UI_TEXT_CONTENT["AISTUDIO::PAGES::INFORMATION::T594602073"] = "starting"
@ -7087,31 +7091,37 @@ UI_TEXT_CONTENT["AISTUDIO::TOOLS::CONFIDENCESCHEMESEXTENSIONS::T3893997203"] = "
UI_TEXT_CONTENT["AISTUDIO::TOOLS::CONFIDENCESCHEMESEXTENSIONS::T4107860491"] = "Trust all LLM providers"
-- Reason
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::NOEMBEDDINGSTORE::T1093747001"] = "Reason"
-- Starting
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::NODATABASECLIENT::T1233211769"] = "Starting"
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::NODATABASECLIENT::T1093747001"] = "Reason"
-- Unavailable
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::NOEMBEDDINGSTORE::T3662391977"] = "Unavailable"
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::NODATABASECLIENT::T3662391977"] = "Unavailable"
-- Status
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::NOEMBEDDINGSTORE::T6222351"] = "Status"
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::NODATABASECLIENT::T6222351"] = "Status"
-- Reason
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::VECTORSTORE::NOVECTORSTORECLIENT::T1093747001"] = "Reason"
-- Unavailable
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::VECTORSTORE::NOVECTORSTORECLIENT::T3662391977"] = "Unavailable"
-- Status
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::VECTORSTORE::NOVECTORSTORECLIENT::T6222351"] = "Status"
-- Storage size
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::QDRANT::QDRANTCLIENTIMPLEMENTATION::T1230141403"] = "Storage size"
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::VECTORSTORE::QDRANTCLIENTIMPLEMENTATION::T1230141403"] = "Storage size"
-- HTTP port
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::QDRANT::QDRANTCLIENTIMPLEMENTATION::T1717573768"] = "HTTP port"
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::VECTORSTORE::QDRANTCLIENTIMPLEMENTATION::T1717573768"] = "HTTP port"
-- Reported version
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::QDRANT::QDRANTCLIENTIMPLEMENTATION::T3556099842"] = "Reported version"
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::VECTORSTORE::QDRANTCLIENTIMPLEMENTATION::T3556099842"] = "Reported version"
-- gRPC port
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::QDRANT::QDRANTCLIENTIMPLEMENTATION::T757840040"] = "gRPC port"
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::VECTORSTORE::QDRANTCLIENTIMPLEMENTATION::T757840040"] = "gRPC port"
-- Number of collections
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::QDRANT::QDRANTCLIENTIMPLEMENTATION::T842647336"] = "Number of collections"
UI_TEXT_CONTENT["AISTUDIO::TOOLS::DATABASES::VECTORSTORE::QDRANTCLIENTIMPLEMENTATION::T842647336"] = "Number of collections"
-- The related data is not allowed to be sent to any LLM provider. This means that this data source cannot be used at the moment.
UI_TEXT_CONTENT["AISTUDIO::TOOLS::ERICLIENT::DATAMODEL::PROVIDERTYPEEXTENSIONS::T1555790630"] = "The related data is not allowed to be sent to any LLM provider. This means that this data source cannot be used at the moment."

6
app/MindWork AI Studio/Components/ChatComponent.razor.cs
View File

@ -4,7 +4,7 @@ using AIStudio.Provider;
using AIStudio.Settings;
using AIStudio.Settings.DataModel;
using AIStudio.Tools.AIJobs;
using AIStudio.Tools.Services;
using Microsoft.AspNetCore.Components;
using Microsoft.AspNetCore.Components.Web;
@ -1051,7 +1051,7 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
return;
if (!tokenizerResponse.Value.Success)
{
this.Logger.LogWarning($"Failed to initialize the tokenizer for the provider: {tokenizerResponse.Value.Message}");
this.Logger.LogWarning($"Failed to initialize the tokenizer for the provider: status='{tokenizerResponse.Value.Status}', reason='{tokenizerResponse.Value.Message}'");
return;
}
@ -1060,7 +1060,7 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
return;
if (!response.Value.Success)
{
this.Logger.LogWarning($"Failed to calculate token count: {response.Value.Message}");
this.Logger.LogWarning($"Failed to calculate token count: status='{response.Value.Status}', reason='{response.Value.Message}'");
return;
}
this.tokenCount = response.Value.TokenCount.ToString();

2
app/MindWork AI Studio/Dialogs/EmbeddingProviderDialog.razor.cs
View File

@ -249,7 +249,7 @@ public partial class EmbeddingProviderDialog : MSGComponentBase, ISecretId
await this.form.Validate();
return;
}
this.dataFilePath = response.Message;
this.dataFilePath = response.StoredPath;
// Use the data model to store the provider.
// We just return this data to the parent component:

2
app/MindWork AI Studio/Dialogs/ProviderDialog.razor.cs
View File

@ -276,7 +276,7 @@ public partial class ProviderDialog : MSGComponentBase, ISecretId
await this.form.Validate();
return;
}
this.dataFilePath = tokenizerResponse.Message;
this.dataFilePath = tokenizerResponse.StoredPath;
// Use the data model to store the provider.
// We just return this data to the parent component:

2
app/MindWork AI Studio/Dialogs/Settings/SettingsDialogDataSources.razor.cs
View File

@ -2,6 +2,8 @@ using AIStudio.Settings;
using AIStudio.Settings.DataModel;
using AIStudio.Tools.ERIClient.DataModel;
using AIStudio.Tools.PluginSystem;
using AIStudio.Tools.Services;
using Microsoft.AspNetCore.Components;
namespace AIStudio.Dialogs.Settings;

6
app/MindWork AI Studio/MindWork AI Studio.csproj
View File

@ -88,7 +88,7 @@
<MetaAppCommitHash>$([System.String]::Copy( $(Metadata) ).Split( ';' )[ 8 ])</MetaAppCommitHash>
<MetaArchitecture>$([System.String]::Copy( $(Metadata) ).Split( ';' )[ 9 ])</MetaArchitecture>
<MetaPdfiumVersion>$([System.String]::Copy( $(Metadata) ).Split( ';' )[ 10 ])</MetaPdfiumVersion>
<MetaQdrantVersion>$([System.String]::Copy( $(Metadata) ).Split( ';' )[ 11 ])</MetaQdrantVersion>
<MetaVectorStoreVersion>$([System.String]::Copy( $(Metadata) ).Split( ';' )[ 11 ])</MetaVectorStoreVersion>
<GenerateAssemblyInfo>true</GenerateAssemblyInfo>
@ -116,8 +116,8 @@
<AssemblyAttribute Include="AIStudio.Tools.Metadata.MetaDataLibraries">
<_Parameter1>$(MetaPdfiumVersion)</_Parameter1>
</AssemblyAttribute>
<AssemblyAttribute Include="AIStudio.Tools.Metadata.MetaDataDatabases">
<_Parameter1>$(MetaQdrantVersion)</_Parameter1>
<AssemblyAttribute Include="AIStudio.Tools.Metadata.MetaDataVectorStore">
<_Parameter1>$(MetaVectorStoreVersion)</_Parameter1>
</AssemblyAttribute>
</ItemGroup>

12
app/MindWork AI Studio/Pages/Information.razor
View File

@ -21,11 +21,11 @@
<MudListItem T="string" Icon="@Icons.Material.Outlined.Build" Text="@this.VersionRust"/>
<MudListItem T="string" Icon="@Icons.Material.Outlined.Storage">
<MudText Typo="Typo.body1">
@this.VersionDatabase
@this.VersionVectorStore
</MudText>
<MudCollapse Expanded="@this.showDatabaseDetails">
<MudCollapse Expanded="@this.showVectorStoreDetails">
<MudText Typo="Typo.body1" Class="mt-2 mb-2">
@foreach (var item in this.databaseDisplayInfo)
@foreach (var item in this.vectorStoreDisplayInfo)
{
<div style="display: flex; align-items: center; gap: 8px;">
<MudIcon Icon="@Icons.Material.Filled.ArrowRightAlt"/>
@ -35,11 +35,11 @@
}
</MudText>
</MudCollapse>
<MudButton StartIcon="@(this.showDatabaseDetails ? Icons.Material.Filled.ExpandLess : Icons.Material.Filled.ExpandMore)"
<MudButton StartIcon="@(this.showVectorStoreDetails ? Icons.Material.Filled.ExpandLess : Icons.Material.Filled.ExpandMore)"
Size="Size.Small"
Variant="Variant.Text"
OnClick="@this.ToggleDatabaseDetails">
@(this.showDatabaseDetails ? T("Hide Details") : T("Show Details"))
OnClick="@this.ToggleVectorStoreDetails">
@(this.showVectorStoreDetails ? T("Hide Details") : T("Show Details"))
</MudButton>
</MudListItem>
<MudListItem T="string" Icon="@Icons.Material.Outlined.DocumentScanner" Text="@this.VersionPdfium"/>

73
app/MindWork AI Studio/Pages/Information.razor.cs
View File

@ -29,13 +29,13 @@ public partial class Information : MSGComponentBase
private ISnackbar Snackbar { get; init; } = null!;
[Inject]
private DatabaseClient DatabaseClient { get; init; } = null!;
private DatabaseClientProvider DatabaseClientProvider { get; init; } = null!;
private static readonly Assembly ASSEMBLY = Assembly.GetExecutingAssembly();
private static readonly MetaDataAttribute META_DATA = ASSEMBLY.GetCustomAttribute<MetaDataAttribute>()!;
private static readonly MetaDataArchitectureAttribute META_DATA_ARCH = ASSEMBLY.GetCustomAttribute<MetaDataArchitectureAttribute>()!;
private static readonly MetaDataLibrariesAttribute META_DATA_LIBRARIES = ASSEMBLY.GetCustomAttribute<MetaDataLibrariesAttribute>()!;
private static readonly MetaDataEmbeddingStoreAttribute META_DATA_EMBEDDING_STORE = ASSEMBLY.GetCustomAttribute<MetaDataEmbeddingStoreAttribute>()!;
private static readonly MetaDataVectorStoreAttribute META_DATA_VECTOR_STORE = ASSEMBLY.GetCustomAttribute<MetaDataVectorStoreAttribute>()!;
private static string TB(string fallbackEN) => I18N.I.T(fallbackEN, typeof(Information).Namespace, nameof(Information));
@ -62,18 +62,18 @@ public partial class Information : MSGComponentBase
private string VersionPdfium => $"{T("Used PDFium version")}: v{META_DATA_LIBRARIES.PdfiumVersion}";
private string VersionEmbeddingStore
private string VersionVectorStore
{
get
{
if (this.embeddingStore is null)
return $"{T("Embedding store")}: {T("checking availability")}";
if (this.vectorStore is null)
return $"{T("Vector store")}: {T("checking availability")}";
return this.embeddingStore.Status switch
return this.vectorStore.Status switch
{
EmbeddingStoreStatus.AVAILABLE => $"{T("Embedding store version")}: {this.embeddingStore.Name} v{META_DATA_EMBEDDING_STORE.DatabaseVersion}",
EmbeddingStoreStatus.STARTING => $"{T("Embedding store")}: {this.embeddingStore.Name} - {T("starting")}",
_ => $"{T("Embedding store")}: {this.embeddingStore.Name} - {T("not available")}"
DatabaseClientStatus.AVAILABLE => $"{T("Vector store version")}: {this.vectorStore.Name} v{META_DATA_VECTOR_STORE.VectorStoreVersion}",
DatabaseClientStatus.STARTING => $"{T("Vector store")}: {this.vectorStore.Name} - {T("starting")}",
_ => $"{T("Vector store")}: {this.vectorStore.Name} - {T("not available")}"
};
}
}
@ -85,7 +85,7 @@ public partial class Information : MSGComponentBase
private bool showEnterpriseConfigDetails;
private bool showEmbeddingStoreDetails;
private bool showVectorStoreDetails;
private List<IAvailablePlugin> configPlugins = PluginFactory.AvailablePlugins
.Where(x => x.Type is PluginType.CONFIGURATION)
@ -95,14 +95,13 @@ public partial class Information : MSGComponentBase
private List<EnterpriseEnvironment> enterpriseEnvironments = EnterpriseEnvironmentService.CURRENT_ENVIRONMENTS.ToList();
private List<MandatoryInfoPanelData> mandatoryInfoPanels = [];
private sealed record EmbeddingStoreDisplayInfo(string Label, string Value);
private sealed record MandatoryInfoPanelData(string HeaderText, string PluginName, DataMandatoryInfo Info, DataMandatoryInfoAcceptance? Acceptance);
private readonly List<EmbeddingStoreDisplayInfo> embeddingStoreDisplayInfo = new();
private DatabaseClient? embeddingStore;
private CancellationTokenSource? databaseRefreshCancellationTokenSource;
private sealed record VectorStoreDisplayInfo(string Label, string Value);
private readonly List<VectorStoreDisplayInfo> vectorStoreDisplayInfo = new();
private DatabaseClient? vectorStore;
private CancellationTokenSource? vectorStoreRefreshCancellationTokenSource;
private bool HasAnyActiveEnvironment => this.enterpriseEnvironments.Any(e => e.IsActive);
@ -148,9 +147,9 @@ public partial class Information : MSGComponentBase
this.osUserName = await this.RustService.ReadUserName();
this.logPaths = await this.RustService.GetLogPaths();
await this.RefreshDatabaseInfo(CancellationToken.None);
if (this.databaseClient?.Status is DatabaseClientStatus.STARTING)
this.StartShortDatabaseRefreshLoop();
await this.RefreshVectorStoreInfo(CancellationToken.None);
if (this.vectorStore?.Status is DatabaseClientStatus.STARTING)
this.StartShortVectorStoreRefreshLoop();
// Determine the Pandoc version may take some time, so we start it here
// without waiting for the result:
@ -249,22 +248,22 @@ public partial class Information : MSGComponentBase
this.showEnterpriseConfigDetails = !this.showEnterpriseConfigDetails;
}
private void ToggleDatabaseDetails()
private void ToggleVectorStoreDetails()
{
this.showEmbeddingStoreDetails = !this.showEmbeddingStoreDetails;
this.showVectorStoreDetails = !this.showVectorStoreDetails;
}
private async Task RefreshDatabaseInfo(CancellationToken cancellationToken)
private async Task RefreshVectorStoreInfo(CancellationToken cancellationToken)
{
var refreshedClient = await this.DatabaseClientProvider.RefreshClientAsync(DatabaseRole.VECTOR_STORE, cancellationToken);
this.databaseClient = refreshedClient;
this.embeddingStoreDisplayInfo.Clear();
this.vectorStore = refreshedClient;
this.vectorStoreDisplayInfo.Clear();
try
{
await foreach (var (label, value) in refreshedClient.GetDisplayInfo().WithCancellation(cancellationToken))
{
this.embeddingStoreDisplayInfo.Add(new EmbeddingStoreDisplayInfo(label, value));
this.vectorStoreDisplayInfo.Add(new VectorStoreDisplayInfo(label, value));
}
}
catch (OperationCanceledException)
@ -273,20 +272,20 @@ public partial class Information : MSGComponentBase
}
catch (Exception e)
{
this.databaseClient = new NoDatabaseClient(refreshedClient.Name, e.Message, DatabaseClientStatus.STARTING);
await foreach (var (label, value) in this.databaseClient.GetDisplayInfo().WithCancellation(cancellationToken))
this.vectorStore = new NoDatabaseClient(refreshedClient.Name, e.Message, DatabaseClientStatus.STARTING);
await foreach (var (label, value) in this.vectorStore.GetDisplayInfo().WithCancellation(cancellationToken))
{
this.embeddingStoreDisplayInfo.Add(new EmbeddingStoreDisplayInfo(label, value));
this.vectorStoreDisplayInfo.Add(new VectorStoreDisplayInfo(label, value));
}
}
}
private void StartShortDatabaseRefreshLoop()
private void StartShortVectorStoreRefreshLoop()
{
this.databaseRefreshCancellationTokenSource?.Cancel();
this.databaseRefreshCancellationTokenSource?.Dispose();
this.databaseRefreshCancellationTokenSource = new CancellationTokenSource();
var cancellationToken = this.databaseRefreshCancellationTokenSource.Token;
this.vectorStoreRefreshCancellationTokenSource?.Cancel();
this.vectorStoreRefreshCancellationTokenSource?.Dispose();
this.vectorStoreRefreshCancellationTokenSource = new CancellationTokenSource();
var cancellationToken = this.vectorStoreRefreshCancellationTokenSource.Token;
_ = Task.Run(async () =>
{
@ -298,11 +297,11 @@ public partial class Information : MSGComponentBase
await Task.Delay(TimeSpan.FromSeconds(1), cancellationToken);
await this.InvokeAsync(async () =>
{
await this.RefreshDatabaseInfo(cancellationToken);
await this.RefreshVectorStoreInfo(cancellationToken);
this.StateHasChanged();
});
if (this.databaseClient?.Status is not DatabaseClientStatus.STARTING)
if (this.vectorStore?.Status is not DatabaseClientStatus.STARTING)
return;
}
catch (OperationCanceledException)
@ -331,8 +330,8 @@ public partial class Information : MSGComponentBase
protected override void DisposeResources()
{
this.databaseRefreshCancellationTokenSource?.Cancel();
this.databaseRefreshCancellationTokenSource?.Dispose();
this.vectorStoreRefreshCancellationTokenSource?.Cancel();
this.vectorStoreRefreshCancellationTokenSource?.Dispose();
base.DisposeResources();
}

14
app/MindWork AI Studio/Program.cs
View File

@ -28,7 +28,7 @@ internal sealed class Program
public static string API_TOKEN = null!;
public static IServiceProvider SERVICE_PROVIDER = null!;
public static ILoggerFactory LOGGER_FACTORY = null!;
public static EmbeddingStore EMBEDDING_STORE = null!;
public static DatabaseClientProvider DATABASE_CLIENT_PROVIDER = null!;
public static async Task Main()
{
@ -87,10 +87,6 @@ internal sealed class Program
return;
}
var embeddingStoreConfig = await rust.GetEmbeddingStoreConfiguration(EmbeddingStoreKind.QDRANT_REMOTE);
var embeddingStore = EmbeddingStoreFactory.Create(embeddingStoreConfig);
var builder = WebApplication.CreateBuilder();
builder.WebHost.ConfigureKestrel(kestrelServerOptions =>
{
@ -148,7 +144,7 @@ internal sealed class Program
builder.Services.AddHostedService<TemporaryChatService>();
builder.Services.AddHostedService<EnterpriseEnvironmentService>();
builder.Services.AddHostedService(sp => sp.GetRequiredService<DataSourceEmbeddingService>());
builder.Services.AddSingleton(embeddingStore);
builder.Services.AddSingleton<DatabaseClientProvider>();
builder.Services.AddHostedService<GlobalShortcutService>();
builder.Services.AddHostedService<RustAvailabilityMonitorService>();
@ -208,9 +204,7 @@ internal sealed class Program
RUST_SERVICE = rust;
ENCRYPTION = encryption;
var databaseLogger = app.Services.GetRequiredService<ILogger<EmbeddingStore>>();
embeddingStore.SetLogger(databaseLogger);
EMBEDDING_STORE = embeddingStore;
DATABASE_CLIENT_PROVIDER = app.Services.GetRequiredService<DatabaseClientProvider>();
programLogger.LogInformation("Initialize internal file system.");
app.Use(Redirect.HandlerContentAsync);
@ -248,7 +242,7 @@ internal sealed class Program
await serverTask;
RUST_SERVICE.Dispose();
EMBEDDING_STORE.Dispose();
DATABASE_CLIENT_PROVIDER.Dispose();
PluginFactory.Dispose();
programLogger.LogInformation("The AI Studio server was stopped.");
}

6
app/MindWork AI Studio/Tools/Databases/DatabaseClient.cs
View File

@ -1,6 +1,6 @@
namespace AIStudio.Tools.Databases;
public abstract class EmbeddingStore(string name, string path)
public abstract class DatabaseClient(string name, string path)
{
public string Name => name;
@ -12,7 +12,7 @@ public abstract class EmbeddingStore(string name, string path)
private string Path => path;
private ILogger<EmbeddingStore>? logger;
private ILogger<DatabaseClient>? logger;
public abstract IAsyncEnumerable<(string Label, string Value)> GetDisplayInfo();
@ -50,7 +50,7 @@ public abstract class EmbeddingStore(string name, string path)
return $"{convertedSize:0.##} {suffixes[suffixIndex]}";
}
public void SetLogger(ILogger<EmbeddingStore> logService)
public void SetLogger(ILogger<DatabaseClient> logService)
{
this.logger = logService;
}

5
app/MindWork AI Studio/Tools/Databases/DatabaseClientProvider.Qdrant.cs
View File

@ -1,4 +1,5 @@
using AIStudio.Tools.Rust;
using AIStudio.Tools.Databases.VectorStore;
using AIStudio.Tools.Rust;
namespace AIStudio.Tools.Databases;
@ -6,7 +7,7 @@ public sealed partial class DatabaseClientProvider
{
private async Task<DatabaseClient> CreateQdrantClientAsync(CancellationToken cancellationToken)
{
var qdrantInfo = await this.rustService.GetQdrantInfo(cancellationToken);
var qdrantInfo = await rustService.GetQdrantInfo(cancellationToken);
if (qdrantInfo.Status is QdrantStatus.STARTING)
{
return this.CreateNoDatabaseClient(

94
app/MindWork AI Studio/Tools/Databases/DatabaseClientProvider.cs
View File

@ -1,15 +1,14 @@
using AIStudio.Tools.Databases.Qdrant;
using AIStudio.Tools.Rust;
using AIStudio.Tools.Services;
using AIStudio.Tools.Databases.VectorStore;
namespace AIStudio.Tools.Databases;
public sealed class EmbeddingStoreProvider(RustService rustService, ILoggerFactory loggerFactory) : IDisposable
public sealed partial class DatabaseClientProvider(RustService rustService, ILoggerFactory loggerFactory) : IDisposable
{
private readonly Dictionary<DatabaseRole, EmbeddingStore> clients = new();
private readonly Dictionary<DatabaseRole, DatabaseClient> clients = new();
private readonly Dictionary<DatabaseRole, SemaphoreSlim> locks = new();
private readonly Lock locksLock = new();
private readonly ILogger<EmbeddingStoreProvider> logger = loggerFactory.CreateLogger<EmbeddingStoreProvider>();
private readonly ILogger<DatabaseClientProvider> logger = loggerFactory.CreateLogger<DatabaseClientProvider>();
private readonly ILogger<DatabaseClient> databaseClientLogger = loggerFactory.CreateLogger<DatabaseClient>();
public async Task<DatabaseClient> GetClientAsync(DatabaseRole databaseRole, CancellationToken cancellationToken = default)
@ -45,6 +44,18 @@ public sealed class EmbeddingStoreProvider(RustService rustService, ILoggerFacto
}
}
public async Task<IVectorStoreClient> GetVectorStoreAsync(CancellationToken cancellationToken = default)
{
var client = await this.GetClientAsync(DatabaseRole.VECTOR_STORE, cancellationToken);
if (client is IVectorStoreClient vectorStore)
return vectorStore;
return new NoVectorStoreClient(
client.Name,
"The configured database client does not support vector store operations.",
client.Status);
}
private DatabaseClient CacheIfAvailable(DatabaseRole databaseRole, DatabaseClient client)
{
if (!client.IsAvailable)
@ -84,79 +95,6 @@ public sealed class EmbeddingStoreProvider(RustService rustService, ILoggerFacto
_ => new NoDatabaseClient(databaseRole.ToString(), "The requested database role is not supported.")
};
private async Task<DatabaseClient> CreateQdrantClientAsync(CancellationToken cancellationToken)
{
var qdrantInfo = await rustService.GetQdrantInfo(cancellationToken);
if (qdrantInfo.Status is QdrantStatus.STARTING)
{
return this.CreateNoDatabaseClient(
"Qdrant",
"Qdrant is starting. Details will appear shortly.",
DatabaseClientStatus.STARTING);
}
if (!qdrantInfo.IsAvailable || qdrantInfo.Status is QdrantStatus.UNAVAILABLE)
{
var reason = qdrantInfo.UnavailableReason ?? "unknown";
this.logger.LogWarning("Qdrant is not available. Starting without vector database. Reason: '{Reason}'.", reason);
return this.CreateNoDatabaseClient("Qdrant", qdrantInfo.UnavailableReason, DatabaseClientStatus.UNAVAILABLE);
}
if (!HasValidQdrantConnectionInfo(qdrantInfo, out var invalidReason))
return this.CreateNoDatabaseClient("Qdrant", invalidReason, DatabaseClientStatus.UNAVAILABLE);
var client = new QdrantClientImplementation("Qdrant", qdrantInfo.Path, qdrantInfo.PortHttp, qdrantInfo.PortGrpc, qdrantInfo.Fingerprint, qdrantInfo.ApiToken);
client.SetLogger(this.databaseClientLogger);
try
{
await client.CheckAvailabilityAsync();
return client;
}
catch (Exception e)
{
client.Dispose();
this.logger.LogWarning(e, "Qdrant reported as available by Rust, but the health check failed.");
return this.CreateNoDatabaseClient("Qdrant", e.Message, DatabaseClientStatus.STARTING);
}
}
private static bool HasValidQdrantConnectionInfo(QdrantInfo qdrantInfo, out string invalidReason)
{
if (qdrantInfo.Path == string.Empty)
{
invalidReason = "Failed to get the Qdrant path from Rust.";
return false;
}
if (qdrantInfo.PortHttp == 0)
{
invalidReason = "Failed to get the Qdrant HTTP port from Rust.";
return false;
}
if (qdrantInfo.PortGrpc == 0)
{
invalidReason = "Failed to get the Qdrant gRPC port from Rust.";
return false;
}
if (qdrantInfo.Fingerprint == string.Empty)
{
invalidReason = "Failed to get the Qdrant fingerprint from Rust.";
return false;
}
if (qdrantInfo.ApiToken == string.Empty)
{
invalidReason = "Failed to get the Qdrant API token from Rust.";
return false;
}
invalidReason = string.Empty;
return true;
}
private NoDatabaseClient CreateNoDatabaseClient(string name, string? unavailableReason, DatabaseClientStatus status)
{
var client = new NoDatabaseClient(name, unavailableReason, status);

33
app/MindWork AI Studio/Tools/Databases/EmbeddingStoreFactory.cs
View File

@ -1,33 +0,0 @@
using AIStudio.Tools.Databases.Qdrant;
namespace AIStudio.Tools.Databases;
public class EmbeddingStoreFactory
{
public static EmbeddingStore Create(EmbeddingStoreConfiguration configuration) => configuration.Kind switch
{
EmbeddingStoreKind.NONE => new NoEmbeddingStore(configuration.Name, configuration.UnavailableReason ?? "unknown"),
_ when configuration.Location is null => new NoEmbeddingStore(configuration.Name, $"No location specified for {configuration.Name}"),
EmbeddingStoreKind.QDRANT_REMOTE when configuration.Location is RemoteLocation location=> new QdrantClientImplementation(configuration.Name, location.Path, location.HttpPort, location.GrpcPort, location.Fingerprint, location.ApiToken),
_ => throw new ArgumentException("Invalid configuration for " + configuration.Name, nameof(configuration)),
};
}
public enum EmbeddingStoreKind
{
NONE,
QDRANT_EMBED,
QDRANT_REMOTE,
}
public abstract record EmbeddingStoreLocation;
public sealed record EmbeddedLocation(string Path) : EmbeddingStoreLocation;
public sealed record RemoteLocation(string Path, int? HttpPort, int? GrpcPort, string? Fingerprint, string? ApiToken) : EmbeddingStoreLocation;
public sealed record EmbeddingStoreConfiguration(
EmbeddingStoreKind Kind,
string Name,
EmbeddingStoreLocation? Location,
string? UnavailableReason);

4
app/MindWork AI Studio/Tools/Databases/NoDatabaseClient.cs
View File

@ -2,9 +2,9 @@ using AIStudio.Tools.PluginSystem;
namespace AIStudio.Tools.Databases;
public sealed class NoEmbeddingStore(string name, string? unavailableReason, DatabaseClientStatus status = DatabaseClientStatus.UNAVAILABLE) : EmbeddingStore(name, string.Empty)
public sealed class NoDatabaseClient(string name, string? unavailableReason, DatabaseClientStatus status = DatabaseClientStatus.UNAVAILABLE) : DatabaseClient(name, string.Empty)
{
private static string TB(string fallbackEN) => I18N.I.T(fallbackEN, typeof(NoEmbeddingStore).Namespace, nameof(NoEmbeddingStore));
private static string TB(string fallbackEN) => I18N.I.T(fallbackEN, typeof(NoDatabaseClient).Namespace, nameof(NoDatabaseClient));
public override DatabaseClientStatus Status => status;

2
app/MindWork AI Studio/Tools/Databases/EmbeddingStoragePoint.cs → app/MindWork AI Studio/Tools/Databases/VectorStoragePoint.cs
View File

@ -1,6 +1,6 @@
namespace AIStudio.Tools.Databases;
public sealed record EmbeddingStoragePoint(
public sealed record VectorStoragePoint(
string PointId,
IReadOnlyList<float> Vector,
string DataSourceId,

20
app/MindWork AI Studio/Tools/Databases/VectorStore/IVectorStoreClient.cs Normal file
View File

@ -0,0 +1,20 @@
namespace AIStudio.Tools.Databases.VectorStore;
public interface IVectorStoreClient
{
string Name { get; }
DatabaseClientStatus Status { get; }
bool IsAvailable { get; }
IAsyncEnumerable<(string Label, string Value)> GetDisplayInfo();
Task EnsureVectorStoreExists(string storeName, int vectorSize, CancellationToken token);
Task InsertEmbedding(string storeName, IReadOnlyList<VectorStoragePoint> points, CancellationToken token);
Task DeleteEmbeddingByFile(string storeName, string filePath, CancellationToken token);
Task DeleteVectorStore(string storeName, CancellationToken token);
}

39
app/MindWork AI Studio/Tools/Databases/VectorStore/NoVectorStoreClient.cs Normal file
View File

@ -0,0 +1,39 @@
using AIStudio.Tools.PluginSystem;
namespace AIStudio.Tools.Databases.VectorStore;
public sealed class NoVectorStoreClient(string name, string? unavailableReason, DatabaseClientStatus status = DatabaseClientStatus.UNAVAILABLE) : IVectorStoreClient
{
private static string TB(string fallbackEN) => I18N.I.T(fallbackEN, typeof(NoVectorStoreClient).Namespace, nameof(NoVectorStoreClient));
public string Name => name;
public DatabaseClientStatus Status => status;
public bool IsAvailable => false;
public async IAsyncEnumerable<(string Label, string Value)> GetDisplayInfo()
{
yield return (TB("Status"), TB("Unavailable"));
if (!string.IsNullOrWhiteSpace(unavailableReason))
yield return (TB("Reason"), unavailableReason);
await Task.CompletedTask;
}
public Task EnsureVectorStoreExists(string storeName, int vectorSize, CancellationToken token) =>
Task.FromException(this.CreateUnavailableException());
public Task InsertEmbedding(string storeName, IReadOnlyList<VectorStoragePoint> points, CancellationToken token) =>
Task.FromException(this.CreateUnavailableException());
public Task DeleteEmbeddingByFile(string storeName, string filePath, CancellationToken token) =>
Task.FromException(this.CreateUnavailableException());
public Task DeleteVectorStore(string storeName, CancellationToken token) =>
Task.FromException(this.CreateUnavailableException());
private InvalidOperationException CreateUnavailableException() =>
new(unavailableReason ?? "The vector store is not available.");
}

14
app/MindWork AI Studio/Tools/Databases/Qdrant/QdrantClientImplementation.cs → app/MindWork AI Studio/Tools/Databases/VectorStore/QdrantClientImplementation.cs
View File

@ -4,9 +4,9 @@ using Grpc.Core;
using AIStudio.Tools.PluginSystem;
using static Qdrant.Client.Grpc.Conditions;
namespace AIStudio.Tools.Databases.Qdrant;
namespace AIStudio.Tools.Databases.VectorStore;
public class QdrantClientImplementation : EmbeddingStore
public class QdrantClientImplementation : DatabaseClient, IVectorStoreClient
{
private static string TB(string fallbackEN) => I18N.I.T(fallbackEN, typeof(QdrantClientImplementation).Namespace, nameof(QdrantClientImplementation));
@ -71,7 +71,7 @@ public class QdrantClientImplementation : EmbeddingStore
yield return (TB("Number of collections"), await this.GetCollectionsAmount());
}
public override async Task EnsureEmbeddingStoreExists(string collectionName, int vectorSize, CancellationToken token)
public async Task EnsureVectorStoreExists(string collectionName, int vectorSize, CancellationToken token)
{
var exists = await this.GrpcClient.CollectionExistsAsync(collectionName, token);
if (exists)
@ -87,7 +87,7 @@ public class QdrantClientImplementation : EmbeddingStore
cancellationToken: token);
}
public override Task InsertEmbedding(string collectionName, IReadOnlyList<EmbeddingStoragePoint> points, CancellationToken token)
public Task InsertEmbedding(string collectionName, IReadOnlyList<VectorStoragePoint> points, CancellationToken token)
{
var qdrantPoints = points.Select(point => new PointStruct
{
@ -112,7 +112,7 @@ public class QdrantClientImplementation : EmbeddingStore
return this.GrpcClient.UpsertAsync(collectionName, qdrantPoints, true, null, null, token);
}
public override async Task DeleteEmbeddingByFile(string collectionName, string filePath, CancellationToken token)
public async Task DeleteEmbeddingByFile(string collectionName, string filePath, CancellationToken token)
{
try
{
@ -120,11 +120,10 @@ public class QdrantClientImplementation : EmbeddingStore
}
catch (RpcException exception) when (exception.StatusCode is StatusCode.NotFound)
{
return;
}
}
public override async Task DeleteEmbeddingStore(string collectionName, CancellationToken token)
public async Task DeleteVectorStore(string collectionName, CancellationToken token)
{
var exists = await this.GrpcClient.CollectionExistsAsync(collectionName, token);
if (!exists)
@ -136,7 +135,6 @@ public class QdrantClientImplementation : EmbeddingStore
}
catch (RpcException exception) when (exception.StatusCode is StatusCode.NotFound)
{
return;
}
}

6
app/MindWork AI Studio/Tools/Metadata/MetaDataEmbeddingStoreAttribute.cs
View File

@ -1,6 +0,0 @@
namespace AIStudio.Tools.Metadata;
public class MetaDataDatabasesAttribute(string databaseVersion) : Attribute
{
public string DatabaseVersion => databaseVersion;
}

6
app/MindWork AI Studio/Tools/Metadata/MetaDataVectorStoreAttribute.cs Normal file
View File

@ -0,0 +1,6 @@
namespace AIStudio.Tools.Metadata;
public class MetaDataVectorStoreAttribute(string vectorStoreVersion) : Attribute
{
public string VectorStoreVersion => vectorStoreVersion;
}

2
app/MindWork AI Studio/Tools/PluginSystem/PluginConfigurationObject.cs
View File

@ -438,7 +438,7 @@ public sealed record PluginConfigurationObject
return string.Empty;
}
return storeResult.Message;
return storeResult.StoredPath;
}
private static string? ResolvePluginTokenizerPath(string configuredTokenizerPath, string pluginPath)

7
app/MindWork AI Studio/Tools/Rust/TokenizerResponse.cs
View File

@ -1,3 +1,8 @@
namespace AIStudio.Tools.Rust;
public readonly record struct TokenizerResponse(bool Success, int TokenCount, string Message);
public readonly record struct TokenizerResponse(
bool Success,
int TokenCount,
string Message,
TokenizerStatus Status = TokenizerStatus.UNAVAILABLE,
string StoredPath = "");

8
app/MindWork AI Studio/Tools/Rust/TokenizerStatus.cs Normal file
View File

@ -0,0 +1,8 @@
namespace AIStudio.Tools.Rust;
public enum TokenizerStatus
{
UNAVAILABLE,
RUNNING,
AVAILABLE,
}

6
app/MindWork AI Studio/Tools/Services/DataSourceEmbeddingService.Files.cs
View File

@ -23,7 +23,7 @@ public sealed partial class DataSourceEmbeddingService
var currentChunk = new StringBuilder();
await foreach (var segment in this.rustService.StreamArbitraryFileData(filePath, token: token))
await foreach (var segment in rustService.StreamArbitraryFileData(filePath, token: token))
{
var normalized = NormalizeChunkSegment(segment);
if (string.IsNullOrWhiteSpace(normalized))
@ -118,7 +118,7 @@ public sealed partial class DataSourceEmbeddingService
}
catch (Exception exception)
{
this.logger.LogWarning(exception, "Cannot access directory '{DirectoryPath}' while indexing.", currentPath);
logger.LogWarning(exception, "Cannot access directory '{DirectoryPath}' while indexing.", currentPath);
result.FailedFiles++;
result.LastError = $"The directory '{currentPath}' could not be accessed.";
continue;
@ -135,7 +135,7 @@ public sealed partial class DataSourceEmbeddingService
}
catch (Exception exception)
{
this.logger.LogWarning(exception, "Cannot inspect file '{FilePath}' while indexing.", filePath);
logger.LogWarning(exception, "Cannot inspect file '{FilePath}' while indexing.", filePath);
result.FailedFiles++;
result.LastError = $"The file '{filePath}' could not be inspected.";
continue;

11
app/MindWork AI Studio/Tools/Services/DataSourceEmbeddingService.State.cs
View File

@ -1,6 +1,7 @@
using System.Text.Json;
using AIStudio.Settings;
using AIStudio.Tools.Databases.VectorStore;
namespace AIStudio.Tools.Services;
@ -70,13 +71,13 @@ public sealed partial class DataSourceEmbeddingService
await File.WriteAllTextAsync(statePath, json, token);
}
private async Task ResetPersistedStateAsync(string dataSourceId)
private async Task ResetPersistedStateAsync(string dataSourceId, IVectorStoreClient? vectorStore, CancellationToken token)
{
await this.EnsureStateLoadedAsync(CancellationToken.None);
await this.EnsureStateLoadedAsync(token);
this.manifests.Remove(dataSourceId);
await this.DeleteCollectionAsync(this.GetCollectionName(dataSourceId));
await this.SaveStateAsync(CancellationToken.None);
this.logger.LogInformation("Reset persisted embedding state for data source '{DataSourceId}'.", dataSourceId);
await this.DeleteCollectionAsync(this.GetCollectionName(dataSourceId), vectorStore, token);
await this.SaveStateAsync(token);
logger.LogInformation("Reset persisted embedding state for data source '{DataSourceId}'.", dataSourceId);
}
private string GetStatePath()

22
app/MindWork AI Studio/Tools/Services/DataSourceEmbeddingService.Watchers.cs
View File

@ -14,13 +14,13 @@ public sealed partial class DataSourceEmbeddingService
private void RefreshWatchers()
{
if (!this.settingsManager.ConfigurationData.DataSourceIndexing.AutomaticRefresh)
if (!settingsManager.ConfigurationData.DataSourceIndexing.AutomaticRefresh)
{
this.RemoveAllWatchers();
return;
}
var supportedSources = this.settingsManager.ConfigurationData.DataSources
var supportedSources = settingsManager.ConfigurationData.DataSources
.Where(this.IsSupportedInternalDataSource)
.ToDictionary(source => source.Id, StringComparer.OrdinalIgnoreCase);
@ -33,7 +33,7 @@ public sealed partial class DataSourceEmbeddingService
private void EnsureWatcher(IDataSource dataSource)
{
if (!this.settingsManager.ConfigurationData.DataSourceIndexing.AutomaticRefresh)
if (!settingsManager.ConfigurationData.DataSourceIndexing.AutomaticRefresh)
return;
var configuration = GetWatchConfiguration(dataSource);
@ -73,7 +73,7 @@ public sealed partial class DataSourceEmbeddingService
watcher.Renamed += (_, _) => this.OnWatchedDataSourceChanged(dataSourceId);
watcher.Error += (_, args) =>
{
this.logger.LogWarning(args.GetException(), "The file watcher for data source '{DataSourceId}' failed. Recreating it.", dataSourceId);
logger.LogWarning(args.GetException(), "The file watcher for data source '{DataSourceId}' failed. Recreating it.", dataSourceId);
this.RemoveWatcher(dataSourceId);
this.EnsureWatcher(dataSourceId);
this.OnWatchedDataSourceChanged(dataSourceId);
@ -83,7 +83,7 @@ public sealed partial class DataSourceEmbeddingService
}
catch (Exception exception)
{
this.logger.LogWarning(exception, "Failed to create file watcher for data source '{DataSourceId}' at '{RootPath}'.", dataSourceId, configuration.RootPath);
logger.LogWarning(exception, "Failed to create file watcher for data source '{DataSourceId}' at '{RootPath}'.", dataSourceId, configuration.RootPath);
return null;
}
}
@ -114,10 +114,10 @@ public sealed partial class DataSourceEmbeddingService
private void OnWatchedDataSourceChanged(string dataSourceId)
{
if (!this.settingsManager.ConfigurationData.DataSourceIndexing.AutomaticRefresh)
if (!settingsManager.ConfigurationData.DataSourceIndexing.AutomaticRefresh)
return;
this.logger.LogDebug("Detected file system change for data source '{DataSourceId}'. Scheduling a debounced embedding run.", dataSourceId);
logger.LogDebug("Detected file system change for data source '{DataSourceId}'. Scheduling a debounced embedding run.", dataSourceId);
var debounceToken = new CancellationTokenSource();
lock (this.watcherDebounceLock)
@ -136,12 +136,12 @@ public sealed partial class DataSourceEmbeddingService
if (!this.TryCompletePendingWatcherRefresh(dataSourceId, debounceToken))
return;
var dataSource = this.settingsManager.ConfigurationData.DataSources
var dataSource = settingsManager.ConfigurationData.DataSources
.FirstOrDefault(source => source.Id.Equals(dataSourceId, StringComparison.OrdinalIgnoreCase));
if (dataSource is not null)
{
this.logger.LogInformation("Queueing data source '{DataSourceName}' ({DataSourceId}) after file system changes settled.", dataSource.Name, dataSource.Id);
logger.LogInformation("Queueing data source '{DataSourceName}' ({DataSourceId}) after file system changes settled.", dataSource.Name, dataSource.Id);
await this.QueueDataSourceAsync(dataSource);
}
}
@ -150,7 +150,7 @@ public sealed partial class DataSourceEmbeddingService
}
catch (Exception exception)
{
this.logger.LogWarning(exception, "Failed to queue watched data source '{DataSourceId}' after a file system change.", dataSourceId);
logger.LogWarning(exception, "Failed to queue watched data source '{DataSourceId}' after a file system change.", dataSourceId);
}
finally
{
@ -161,7 +161,7 @@ public sealed partial class DataSourceEmbeddingService
private void EnsureWatcher(string dataSourceId)
{
var dataSource = this.settingsManager.ConfigurationData.DataSources
var dataSource = settingsManager.ConfigurationData.DataSources
.FirstOrDefault(source => source.Id.Equals(dataSourceId, StringComparison.OrdinalIgnoreCase));
if (dataSource is not null)

131
app/MindWork AI Studio/Tools/Services/DataSourceEmbeddingService.cs
View File

@ -6,22 +6,19 @@ using AIStudio.Provider;
using AIStudio.Settings;
using AIStudio.Settings.DataModel;
using AIStudio.Tools.Databases;
using AIStudio.Tools.Databases.VectorStore;
using AIStudio.Tools.PluginSystem;
using AIStudio.Tools.Rust;
namespace AIStudio.Tools.Services;
public sealed partial class DataSourceEmbeddingService : BackgroundService
public sealed partial class DataSourceEmbeddingService(SettingsManager settingsManager, RustService rustService, DatabaseClientProvider databaseClientProvider, ILogger<DataSourceEmbeddingService> logger)
: BackgroundService
{
private const int MAX_CHUNK_LENGTH = 3_200;
private const int MIN_CHUNK_LENGTH = 800;
private const int CHUNK_OVERLAP_LENGTH = 320;
private const int EMBEDDING_BATCH_SIZE = 16;
private readonly SettingsManager settingsManager;
private readonly RustService rustService;
private readonly EmbeddingStore embeddingStore;
private readonly ILogger<DataSourceEmbeddingService> logger;
private readonly Channel<string> queue = Channel.CreateUnbounded<string>();
private readonly ConcurrentDictionary<string, byte> queuedIds = new(StringComparer.OrdinalIgnoreCase);
private readonly ConcurrentDictionary<string, DataSourceEmbeddingStatus> statuses = new(StringComparer.OrdinalIgnoreCase);
@ -32,14 +29,6 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
private Dictionary<string, DataSourceEmbeddingManifest> manifests = new(StringComparer.OrdinalIgnoreCase);
private bool stateLoaded;
public DataSourceEmbeddingService(SettingsManager settingsManager, RustService rustService, EmbeddingStore embeddingStore, ILogger<DataSourceEmbeddingService> logger)
{
this.settingsManager = settingsManager;
this.rustService = rustService;
this.embeddingStore = embeddingStore;
this.logger = logger;
}
public IReadOnlyList<DataSourceEmbeddingStatus> GetStatuses()
{
return this.statuses.Values
@ -78,7 +67,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
{
this.RefreshWatchers();
var tasks = this.settingsManager.ConfigurationData.DataSources
var tasks = settingsManager.ConfigurationData.DataSources
.Where(this.IsSupportedInternalDataSource)
.Select(this.QueueDataSourceAsync);
@ -87,7 +76,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
public Task QueueAllInternalDataSourcesIfAutomaticRefreshAsync()
{
if (!this.settingsManager.ConfigurationData.DataSourceIndexing.AutomaticRefresh)
if (!settingsManager.ConfigurationData.DataSourceIndexing.AutomaticRefresh)
{
this.RefreshWatchers();
return Task.CompletedTask;
@ -106,16 +95,16 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
if (!this.IsSupportedInternalDataSource(dataSource))
return;
this.logger.LogInformation("Queueing data source '{DataSourceName}' ({DataSourceId}) for background embeddings.", dataSource.Name, dataSource.Id);
logger.LogInformation("Queueing data source '{DataSourceName}' ({DataSourceId}) for background embeddings.", dataSource.Name, dataSource.Id);
this.RefreshWatchers();
this.logger.LogDebug("Adding watcher for data source '{DataSourceName}' ({DataSourceId}).", dataSource.Name, dataSource.Id);
logger.LogDebug("Adding watcher for data source '{DataSourceName}' ({DataSourceId}).", dataSource.Name, dataSource.Id);
if (!this.statuses.TryGetValue(dataSource.Id, out var currentStatus) || currentStatus.State is not DataSourceEmbeddingState.RUNNING)
this.UpsertStatus(this.CreateStatus(dataSource, DataSourceEmbeddingState.QUEUED, currentStatus?.TotalFiles ?? 0, currentStatus?.IndexedFiles ?? 0, currentStatus?.FailedFiles ?? 0));
this.logger.LogDebug("Upserting status for data source '{DataSourceName}' ({DataSourceId}).", dataSource.Name, dataSource.Id);
logger.LogDebug("Upserting status for data source '{DataSourceName}' ({DataSourceId}).", dataSource.Name, dataSource.Id);
if (this.queuedIds.TryAdd(dataSource.Id, 0))
await this.queue.Writer.WriteAsync(dataSource.Id);
this.logger.LogDebug("Queued data source '{DataSourceName}' ({DataSourceId}).", dataSource.Name, dataSource.Id);
logger.LogDebug("Queued data source '{DataSourceName}' ({DataSourceId}).", dataSource.Name, dataSource.Id);
}
public async Task RemoveDataSourceAsync(IDataSource dataSource)
@ -125,7 +114,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
this.RemoveWatcher(dataSource.Id);
this.statuses.TryRemove(dataSource.Id, out _);
await this.ResetPersistedStateAsync(dataSource.Id);
await this.ResetPersistedStateAsync(dataSource.Id, null, CancellationToken.None);
this.PublishStatusChanged();
}
@ -138,7 +127,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
var dataSourceId = await this.queue.Reader.ReadAsync(stoppingToken);
this.queuedIds.TryRemove(dataSourceId, out _);
var dataSource = this.settingsManager.ConfigurationData.DataSources
var dataSource = settingsManager.ConfigurationData.DataSources
.FirstOrDefault(source => source.Id.Equals(dataSourceId, StringComparison.OrdinalIgnoreCase));
if (dataSource is null || !this.IsSupportedInternalDataSource(dataSource))
@ -154,7 +143,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
}
catch (Exception exception)
{
this.logger.LogError(exception, "Background embedding failed for data source '{DataSourceName}' ({DataSourceId}).", dataSource.Name, dataSource.Id);
logger.LogError(exception, "Background embedding failed for data source '{DataSourceName}' ({DataSourceId}).", dataSource.Name, dataSource.Id);
this.UpsertStatus(this.GetFallbackStatus(dataSource, exception.Message));
}
}
@ -170,15 +159,17 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
private async Task ProcessDataSourceAsync(IDataSource dataSource, CancellationToken token)
{
await this.EnsureStateLoadedAsync(token);
this.logger.LogInformation("Starting background embeddings for data source '{DataSourceName}' ({DataSourceId}).", dataSource.Name, dataSource.Id);
logger.LogInformation("Starting background embeddings for data source '{DataSourceName}' ({DataSourceId}).", dataSource.Name, dataSource.Id);
if (!this.embeddingStore.IsAvailable)
var vectorStore = await databaseClientProvider.GetVectorStoreAsync(token);
if (!vectorStore.IsAvailable)
{
this.logger.LogWarning(
logger.LogWarning(
"Skipping background embeddings for data source '{DataSourceName}' ({DataSourceId}) because the database client '{DatabaseName}' is unavailable.",
dataSource.Name,
dataSource.Id,
this.embeddingStore.Name);
vectorStore.Name);
this.UpsertStatus(this.GetFallbackStatus(dataSource, "The vector database is not available."));
return;
}
@ -189,7 +180,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
return;
}
this.logger.LogInformation(
logger.LogInformation(
"Using embedding provider '{EmbeddingProviderId}' with model '{EmbeddingModelId}' for data source '{DataSourceName}' ({DataSourceId}).",
embeddingProvider.Id,
embeddingProvider.Model.Id,
@ -197,12 +188,12 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
dataSource.Id);
var collectionName = this.GetCollectionName(dataSource.Id);
var manifest = await this.EnsureCompatibleManifestAsync(dataSource, embeddingProvider, collectionName, token);
var manifest = await this.EnsureCompatibleManifestAsync(dataSource, embeddingProvider, collectionName, vectorStore, token);
var inputFiles = this.GetInputFiles(dataSource);
var indexedFiles = inputFiles.Files;
var totalFiles = indexedFiles.Count + inputFiles.FailedFiles;
this.logger.LogInformation(
logger.LogInformation(
"Prepared data source '{DataSourceName}' ({DataSourceId}) for embedding. AccessibleFiles={AccessibleFiles}, FailedFiles={FailedFiles}, Collection='{CollectionName}'.",
dataSource.Name,
dataSource.Id,
@ -210,7 +201,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
inputFiles.FailedFiles,
collectionName);
await this.RemoveMissingFileEmbeddingsAsync(dataSource, collectionName, manifest, indexedFiles, token);
await this.RemoveMissingFileEmbeddingsAsync(vectorStore, dataSource, collectionName, manifest, indexedFiles, token);
await this.SaveStateAsync(token);
this.UpsertStatus(this.CreateStatus(
@ -235,7 +226,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
if (manifest.Files.TryGetValue(file.FullName, out var existingRecord) &&
string.Equals(existingRecord.Fingerprint, fingerprint, StringComparison.Ordinal))
{
this.logger.LogDebug(
logger.LogDebug(
"Skipping unchanged file '{FilePath}' for data source '{DataSourceName}' ({DataSourceId}).",
file.FullName,
dataSource.Name,
@ -249,7 +240,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
try
{
this.logger.LogInformation(
logger.LogInformation(
"Embedding file '{FilePath}' for data source '{DataSourceName}' ({DataSourceId}). Progress={CompletedFiles}/{TotalFiles}.",
file.FullName,
dataSource.Name,
@ -257,7 +248,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
skippedFiles + completedFiles + 1,
totalFiles);
var startedAtUtc = DateTime.UtcNow;
var chunkCount = await this.IndexOneFileAsync(dataSource, file, fingerprint, embeddingProvider, provider, manifest, token);
var chunkCount = await this.IndexOneFileAsync(vectorStore, dataSource, file, fingerprint, embeddingProvider, provider, manifest, token);
manifest.Files[file.FullName] = new EmbeddedFileRecord(
fingerprint,
file.Length,
@ -266,7 +257,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
chunkCount);
await this.SaveStateAsync(token);
completedFiles++;
this.logger.LogInformation(
logger.LogInformation(
"Embedded file '{FilePath}' for data source '{DataSourceName}' ({DataSourceId}) successfully. Chunks={ChunkCount}, DurationMs={DurationMs}.",
file.FullName,
dataSource.Name,
@ -279,16 +270,16 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
failedFiles++;
lastError = exception.Message;
manifest.Files.Remove(file.FullName);
await this.DeleteFilePointsAsync(collectionName, file.FullName, token);
await this.DeleteFilePointsAsync(vectorStore, collectionName, file.FullName, token);
await this.SaveStateAsync(token);
this.logger.LogWarning(exception, "Failed to embed file '{FilePath}' for data source '{DataSourceName}'.", file.FullName, dataSource.Name);
logger.LogWarning(exception, "Failed to embed file '{FilePath}' for data source '{DataSourceName}'.", file.FullName, dataSource.Name);
this.UpsertStatus(this.CreateStatus(dataSource, DataSourceEmbeddingState.RUNNING, totalFiles, skippedFiles + completedFiles, failedFiles, file.Name, exception.Message));
}
}
this.UpsertStatus(this.CreateCompletedStatus(dataSource, totalFiles, skippedFiles + completedFiles, failedFiles, lastError));
this.logger.LogInformation(
logger.LogInformation(
"Finished background embeddings for data source '{DataSourceName}' ({DataSourceId}). Indexed={IndexedFiles}, Failed={FailedFiles}, Total={TotalFiles}.",
dataSource.Name,
dataSource.Id,
@ -298,6 +289,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
}
private async Task<int> IndexOneFileAsync(
IVectorStoreClient vectorStore,
IDataSource dataSource,
FileInfo file,
string fingerprint,
@ -307,11 +299,11 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
CancellationToken token)
{
var collectionName = this.GetCollectionName(dataSource.Id);
this.logger.LogDebug(
logger.LogDebug(
"Resetting stored embeddings for file '{FilePath}' in collection '{CollectionName}' before re-indexing.",
file.FullName,
collectionName);
await this.DeleteFilePointsAsync(collectionName, file.FullName, token);
await this.DeleteFilePointsAsync(vectorStore, collectionName, file.FullName, token);
var batch = new List<(string Text, int ChunkIndex)>(EMBEDDING_BATCH_SIZE);
var totalChunkCount = 0;
@ -322,16 +314,16 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
totalChunkCount++;
if (batch.Count >= EMBEDDING_BATCH_SIZE)
await this.FlushBatchAsync(dataSource, file, fingerprint, embeddingProvider, provider, manifest, collectionName, batch, token);
await this.FlushBatchAsync(vectorStore, dataSource, file, fingerprint, embeddingProvider, provider, manifest, collectionName, batch, token);
}
if (batch.Count > 0)
await this.FlushBatchAsync(dataSource, file, fingerprint, embeddingProvider, provider, manifest, collectionName, batch, token);
await this.FlushBatchAsync(vectorStore, dataSource, file, fingerprint, embeddingProvider, provider, manifest, collectionName, batch, token);
if (totalChunkCount == 0)
throw new InvalidOperationException($"The file '{file.Name}' did not yield any text chunks.");
this.logger.LogDebug(
logger.LogDebug(
"Generated {ChunkCount} chunks for file '{FilePath}' in data source '{DataSourceName}' ({DataSourceId}).",
totalChunkCount,
file.FullName,
@ -342,6 +334,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
}
private async Task FlushBatchAsync(
IVectorStoreClient vectorStore,
IDataSource dataSource,
FileInfo file,
string fingerprint,
@ -352,7 +345,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
List<(string Text, int ChunkIndex)> batch,
CancellationToken token)
{
this.logger.LogDebug(
logger.LogDebug(
"Requesting embeddings for batch of {ChunkCount} chunks from file '{FilePath}' in data source '{DataSourceName}' ({DataSourceId}).",
batch.Count,
file.FullName,
@ -360,7 +353,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
dataSource.Id);
var texts = batch.Select(item => item.Text).ToList();
var vectors = await provider.EmbedTextAsync(embeddingProvider.Model, this.settingsManager, token, texts);
var vectors = await provider.EmbedTextAsync(embeddingProvider.Model, settingsManager, token, texts);
if (vectors.Count != batch.Count)
throw new InvalidOperationException($"The embedding provider returned {vectors.Count} vectors for {batch.Count} text chunks.");
@ -374,9 +367,9 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
if (manifest.VectorSize == 0)
{
manifest.VectorSize = vectorSize;
await this.EnsureCollectionExistsAsync(collectionName, vectorSize, token);
await this.EnsureCollectionExistsAsync(vectorStore, collectionName, vectorSize, token);
await this.SaveStateAsync(token);
this.logger.LogInformation(
logger.LogInformation(
"Created embedding collection '{CollectionName}' with vector size {VectorSize} for data source '{DataSourceName}' ({DataSourceId}).",
collectionName,
vectorSize,
@ -385,6 +378,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
}
await this.UpsertPointsAsync(
vectorStore,
collectionName,
dataSource,
file,
@ -394,7 +388,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
this.TryGetRelativePath(dataSource, file),
token);
this.logger.LogDebug(
logger.LogDebug(
"Stored {ChunkCount} embedded chunks for file '{FilePath}' in collection '{CollectionName}'.",
batch.Count,
file.FullName,
@ -403,12 +397,13 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
batch.Clear();
}
private async Task EnsureCollectionExistsAsync(string collectionName, int vectorSize, CancellationToken token)
private async Task EnsureCollectionExistsAsync(IVectorStoreClient vectorStore, string collectionName, int vectorSize, CancellationToken token)
{
await this.embeddingStore.EnsureEmbeddingStoreExists(collectionName, vectorSize, token);
await vectorStore.EnsureVectorStoreExists(collectionName, vectorSize, token);
}
private async Task UpsertPointsAsync(
IVectorStoreClient vectorStore,
string collectionName,
IDataSource dataSource,
FileInfo file,
@ -419,7 +414,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
CancellationToken token)
{
var embeddedAtUtc = DateTime.UtcNow;
var points = batch.Select((item, index) => new EmbeddingStoragePoint(
var points = batch.Select((item, index) => new VectorStoragePoint(
this.CreatePointId(dataSource.Id, fingerprint, item.ChunkIndex),
vectors[index],
dataSource.Id,
@ -434,24 +429,31 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
file.LastWriteTimeUtc,
embeddedAtUtc)).ToList();
await this.embeddingStore.InsertEmbedding(collectionName, points, token);
await vectorStore.InsertEmbedding(collectionName, points, token);
}
private async Task DeleteFilePointsAsync(string collectionName, string filePath, CancellationToken token)
private async Task DeleteFilePointsAsync(IVectorStoreClient vectorStore, string collectionName, string filePath, CancellationToken token)
{
await this.embeddingStore.DeleteEmbeddingByFile(collectionName, filePath, token);
await vectorStore.DeleteEmbeddingByFile(collectionName, filePath, token);
}
private async Task DeleteCollectionAsync(string collectionName)
private async Task DeleteCollectionAsync(string collectionName, IVectorStoreClient? vectorStore, CancellationToken token)
{
await this.embeddingStore.DeleteEmbeddingStore(collectionName, CancellationToken.None);
vectorStore ??= await databaseClientProvider.GetVectorStoreAsync(token);
if (!vectorStore.IsAvailable)
{
logger.LogWarning("Could not delete embedding collection '{CollectionName}' because the vector store '{VectorStoreName}' is unavailable.", collectionName, vectorStore.Name);
return;
}
await vectorStore.DeleteVectorStore(collectionName, token);
}
private async Task WaitForInitialSettingsAndBootstrapAsync(CancellationToken token)
{
while (!token.IsCancellationRequested)
{
if (this.settingsManager.HasCompletedInitialSettingsLoad
if (settingsManager.HasCompletedInitialSettingsLoad
&& !string.IsNullOrWhiteSpace(SettingsManager.ConfigDirectory)
&& !string.IsNullOrWhiteSpace(SettingsManager.DataDirectory))
{
@ -463,7 +465,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
token.ThrowIfCancellationRequested();
this.logger.LogInformation("Embedding background service is ready. Checking whether automatic data source refresh is enabled.");
logger.LogInformation("Embedding background service is ready. Checking whether automatic data source refresh is enabled.");
await this.QueueAllInternalDataSourcesIfAutomaticRefreshAsync();
}
@ -474,26 +476,26 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
private bool TryResolveEmbeddingProvider(IDataSource dataSource, [NotNullWhen(true)] out EmbeddingProvider? embeddingProvider)
{
embeddingProvider = this.settingsManager.ConfigurationData.EmbeddingProviders.FirstOrDefault(provider =>
embeddingProvider = settingsManager.ConfigurationData.EmbeddingProviders.FirstOrDefault(provider =>
dataSource is IInternalDataSource internalDataSource &&
provider.Id.Equals(internalDataSource.EmbeddingId, StringComparison.OrdinalIgnoreCase));
return embeddingProvider != default && embeddingProvider.UsedLLMProvider is not LLMProviders.NONE;
}
private async Task<DataSourceEmbeddingManifest> EnsureCompatibleManifestAsync(IDataSource dataSource, EmbeddingProvider embeddingProvider, string collectionName, CancellationToken token)
private async Task<DataSourceEmbeddingManifest> EnsureCompatibleManifestAsync(IDataSource dataSource, EmbeddingProvider embeddingProvider, string collectionName, IVectorStoreClient vectorStore, CancellationToken token)
{
var embeddingSignature = this.BuildEmbeddingSignature(embeddingProvider);
var manifest = await this.GetManifestAsync(dataSource.Id, token);
if (!string.Equals(manifest.EmbeddingSignature, embeddingSignature, StringComparison.Ordinal))
{
this.logger.LogInformation(
logger.LogInformation(
"Embedding configuration changed for data source '{DataSourceName}' ({DataSourceId}). Resetting persisted state and collection '{CollectionName}'.",
dataSource.Name,
dataSource.Id,
collectionName);
await this.ResetPersistedStateAsync(dataSource.Id);
await this.ResetPersistedStateAsync(dataSource.Id, vectorStore, token);
manifest = await this.GetManifestAsync(dataSource.Id, token);
}
@ -509,6 +511,7 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
}
private async Task RemoveMissingFileEmbeddingsAsync(
IVectorStoreClient vectorStore,
IDataSource dataSource,
string collectionName,
DataSourceEmbeddingManifest manifest,
@ -521,9 +524,9 @@ public sealed partial class DataSourceEmbeddingService : BackgroundService
foreach (var removedFilePath in manifest.Files.Keys.Except(existingPaths, StringComparer.OrdinalIgnoreCase).ToList())
{
await this.DeleteFilePointsAsync(collectionName, removedFilePath, token);
await this.DeleteFilePointsAsync(vectorStore, collectionName, removedFilePath, token);
manifest.Files.Remove(removedFilePath);
this.logger.LogInformation(
logger.LogInformation(
"Removed stale embeddings for deleted file '{FilePath}' from data source '{DataSourceName}' ({DataSourceId}).",
removedFilePath,
dataSource.Name,

83
app/MindWork AI Studio/Tools/Services/RustService.Tokenizer.cs
View File

@ -9,6 +9,31 @@ public sealed partial class RustService
private string currentTokenizerPath = string.Empty;
private bool hasInitializedTokenizer;
private static TokenizerResponse CreateUnavailableTokenizerResponse(string message) => new(
false,
0,
message,
TokenizerStatus.UNAVAILABLE,
string.Empty);
public async Task<TokenizerResponse> GetTokenizerInfo(CancellationToken cancellationToken = default)
{
try
{
return await this.http.GetFromJsonAsync<TokenizerResponse>("/system/tokenizer/info", this.jsonRustSerializerOptions, cancellationToken);
}
catch (OperationCanceledException) when (cancellationToken.IsCancellationRequested)
{
this.logger?.LogWarning("Fetching tokenizer info from Rust service was cancelled by caller.");
return CreateUnavailableTokenizerResponse("Operation cancelled by caller.");
}
catch (Exception e)
{
this.logger?.LogError(e, "Error while fetching tokenizer info from Rust service.");
return CreateUnavailableTokenizerResponse(e.Message);
}
}
public async Task<TokenizerResponse> ValidateTokenizer(string filePath)
{
var result = await this.http.PostAsJsonAsync("/tokenizer/validate", new {
@ -18,15 +43,12 @@ public sealed partial class RustService
if (!result.IsSuccessStatusCode)
{
this.logger!.LogError($"Failed to validate the tokenizer '{result.StatusCode}'");
return new TokenizerResponse
{
Success = false,
Message = "An error occured while sending the path to the Rust framework for validation: "+result.StatusCode,
TokenCount = 0
};
return CreateUnavailableTokenizerResponse("An error occured while sending the path to the Rust framework for validation: "+result.StatusCode);
}
return await result.Content.ReadFromJsonAsync<TokenizerResponse>(this.jsonRustSerializerOptions);
var response = await result.Content.ReadFromJsonAsync<TokenizerResponse>(this.jsonRustSerializerOptions);
return response;
}
public async Task<TokenizerResponse> StoreTokenizer(string modelId, string filePath)
@ -40,11 +62,7 @@ public sealed partial class RustService
if (!result.IsSuccessStatusCode)
{
this.logger!.LogError($"Failed to store the tokenizer '{result.StatusCode}'");
return new TokenizerResponse{
Success = false,
Message = "An error occured while sending the path to the Rust framework for storing: "+result.StatusCode,
TokenCount = 0
};
return CreateUnavailableTokenizerResponse("An error occured while sending the path to the Rust framework for storing: "+result.StatusCode);
}
return await result.Content.ReadFromJsonAsync<TokenizerResponse>(this.jsonRustSerializerOptions);
@ -60,11 +78,7 @@ public sealed partial class RustService
if (!result.IsSuccessStatusCode)
{
this.logger!.LogError($"Failed to delete the tokenizer '{result.StatusCode}'");
return new TokenizerResponse{
Success = false,
Message = "An error occured while sending the tokenizer delete request to the Rust framework: "+result.StatusCode,
TokenCount = 0
};
return CreateUnavailableTokenizerResponse("An error occured while sending the tokenizer delete request to the Rust framework: "+result.StatusCode);
}
return await result.Content.ReadFromJsonAsync<TokenizerResponse>(this.jsonRustSerializerOptions);
@ -79,14 +93,15 @@ public sealed partial class RustService
if (!result.IsSuccessStatusCode)
{
this.logger!.LogError($"Failed to get the token count '{result.StatusCode}'");
return new TokenizerResponse{
Success = false,
Message = "Error while getting token count from Rust service: "+result.StatusCode,
TokenCount = 0
};
this.hasInitializedTokenizer = false;
return CreateUnavailableTokenizerResponse("Error while getting token count from Rust service: "+result.StatusCode);
}
return await result.Content.ReadFromJsonAsync<TokenizerResponse>(this.jsonRustSerializerOptions);
var response = await result.Content.ReadFromJsonAsync<TokenizerResponse>(this.jsonRustSerializerOptions);
if (response is not { Status: TokenizerStatus.AVAILABLE })
this.hasInitializedTokenizer = false;
return response;
}
public async Task<TokenizerResponse?> SetTokenizer(string providerName, string path)
@ -99,14 +114,15 @@ public sealed partial class RustService
if (!result.IsSuccessStatusCode)
{
this.logger!.LogError($"Failed to set the tokenizer '{result.StatusCode}'");
return new TokenizerResponse{
Success = false,
Message = "An error occured while sending the path to the Rust framework for setting a tokenizer: "+result.StatusCode,
TokenCount = 0
};
this.hasInitializedTokenizer = false;
return CreateUnavailableTokenizerResponse("An error occured while sending the path to the Rust framework for setting a tokenizer: "+result.StatusCode);
}
return await result.Content.ReadFromJsonAsync<TokenizerResponse>(this.jsonRustSerializerOptions);
var response = await result.Content.ReadFromJsonAsync<TokenizerResponse>(this.jsonRustSerializerOptions);
if (response is not { Success: true, Status: TokenizerStatus.AVAILABLE })
this.hasInitializedTokenizer = false;
return response;
}
public async Task<TokenizerResponse?> EnsureTokenizer(string providerName, string path)
@ -115,14 +131,19 @@ public sealed partial class RustService
try
{
if (this.hasInitializedTokenizer && this.currentTokenizerPath == path)
return new TokenizerResponse(true, 0, "Success");
return new TokenizerResponse(true, 0, string.Empty, TokenizerStatus.AVAILABLE);
var response = await this.SetTokenizer(providerName, path);
if (response is { Success: true })
if (response is { Success: true, Status: TokenizerStatus.AVAILABLE })
{
this.currentTokenizerPath = path;
this.hasInitializedTokenizer = true;
}
else
{
this.currentTokenizerPath = string.Empty;
this.hasInitializedTokenizer = false;
}
return response;
}

3
runtime/Cargo.toml
View File

@ -1,6 +1,6 @@
[package]
name = "mindwork-ai-studio"
version = "26.4.1"
version = "26.5.5"
edition = "2024"
description = "MindWork AI Studio"
authors = ["Thorsten Sommer"]
@ -48,6 +48,7 @@ tempfile = "3.27.0"
strum_macros = "0.28.0"
sysinfo = "0.39.1"
bytes = "1.11.1"
tokenizers = "0.23.1"
[target.'cfg(target_os = "windows")'.dependencies]
windows-registry = "0.6.1"

4
runtime/src/app_window.rs
View File

@ -28,7 +28,7 @@ use crate::pdfium::PDFIUM_LIB_PATH;
use crate::qdrant::{start_qdrant_server, stop_qdrant_server};
#[cfg(debug_assertions)]
use crate::dotnet::create_startup_env_file;
use crate::tokenizer::set_path_resolver;
use crate::tokenizer::set_default_tokenizer_path;
/// The Tauri main window.
pub static MAIN_WINDOW: Lazy<Mutex<Option<WebviewWindow>>> = Lazy::new(|| Mutex::new(None));
@ -151,7 +151,7 @@ pub fn start_tauri() {
start_qdrant_server(app.handle().clone());
set_path_resolver(app.path_resolver());
set_default_tokenizer_path(app.handle().clone());
info!(Source = "Bootloader Tauri"; "Reconfigure the file logger to use the app data directory {data_path:?}");
switch_to_file_logging(data_path).map_err(|e| error!("Failed to switch logging to file: {e}")).unwrap();

6
runtime/src/runtime_api.rs
View File

@ -33,6 +33,7 @@ pub fn start_runtime_api() {
.route("/system/dotnet/port", get(crate::dotnet::dotnet_port))
.route("/system/dotnet/ready", get(crate::dotnet::dotnet_ready))
.route("/system/qdrant/info", get(crate::qdrant::qdrant_port))
.route("/system/tokenizer/info", get(crate::tokenizer::tokenizer_info))
.route("/clipboard/set", post(crate::clipboard::set_clipboard))
.route("/events", get(crate::app_window::get_event_stream))
.route("/updates/check", get(crate::app_window::check_for_update))
@ -56,6 +57,11 @@ pub fn start_runtime_api() {
.route("/retrieval/fs/extract", get(crate::file_data::extract_data))
.route("/log/paths", get(crate::log::get_log_paths))
.route("/log/event", post(crate::log::log_event))
.route("/tokenizer/count", post(crate::tokenizer::token_count))
.route("/tokenizer/validate", post(crate::tokenizer::validate_tokenizer))
.route("/tokenizer/store", post(crate::tokenizer::store_tokenizer))
.route("/tokenizer/delete", post(crate::tokenizer::delete_tokenizer))
.route("/tokenizer/set", post(crate::tokenizer::set_tokenizer))
.route("/shortcuts/register", post(crate::app_window::register_shortcut))
.route("/shortcuts/validate", post(crate::app_window::validate_shortcut))
.route("/shortcuts/suspend", post(crate::app_window::suspend_shortcuts))

404
runtime/src/tokenizer.rs
View File

@ -1,24 +1,43 @@
use std::fs;
use std::path::PathBuf;
use std::sync::{OnceLock, RwLock};
use log::warn;
use rocket::post;
use rocket::serde::json::Json;
use rocket::serde::Serialize;
use serde::Deserialize;
use tauri::PathResolver;
use tokenizers::Error;
use tokenizers::tokenizer::{Tokenizer, Error as TokenizerError};
use std::sync::{Mutex, OnceLock, RwLock};
use axum::Json;
use log::{error, warn};
use once_cell::sync::Lazy;
use serde::{Deserialize, Serialize};
use tauri::path::BaseDirectory;
use tauri::Manager;
use tokenizers::tokenizer::Tokenizer;
use crate::api_token::APIToken;
use crate::environment::DATA_DIRECTORY;
static TOKENIZER: OnceLock<RwLock<Option<Tokenizer>>> = OnceLock::new();
const DEFAULT_TOKENIZER_RESOURCE_PATH: &str = "resources/tokenizers/tokenizer.json";
const NO_TOKENIZER_LOADED_MESSAGE: &str = "Tokenizer must be set before counting tokens.";
static TOKENIZER_PATH_RESOLVER: OnceLock<PathResolver> = OnceLock::new();
static TOKENIZER: OnceLock<RwLock<Option<Tokenizer>>> = OnceLock::new();
static DEFAULT_TOKENIZER_PATH: OnceLock<PathBuf> = OnceLock::new();
static TOKENIZER_STATUS: Lazy<Mutex<TokenizerStatusInfo>> = Lazy::new(|| Mutex::new(TokenizerStatusInfo::default()));
static TOKENIZER_OPERATION_LOCK: Lazy<Mutex<()>> = Lazy::new(|| Mutex::new(()));
#[derive(Clone, Copy, Default, Serialize, PartialEq, Eq)]
pub enum TokenizerStatus {
#[default]
Unavailable,
Running,
Available,
}
#[derive(Default)]
struct TokenizerStatusInfo {
status: TokenizerStatus,
unavailable_reason: Option<String>,
}
#[derive(Deserialize)]
pub struct SetTokenText {
pub text: String,
text: String,
}
#[derive(Clone, Deserialize)]
@ -42,94 +61,205 @@ pub struct TokenizerResponse {
success: bool,
token_count: usize,
message: String,
status: TokenizerStatus,
stored_path: String,
}
impl From<Result<usize, TokenizerError>> for TokenizerResponse {
fn from(result: Result<usize, TokenizerError>) -> Self {
match result {
Ok(count) => TokenizerResponse {
success: true,
token_count: count,
message: "Success".to_string(),
},
Err(e) => TokenizerResponse {
success: false,
token_count: 0,
message: e.to_string(),
},
impl TokenizerResponse {
fn available(token_count: usize) -> Self {
TokenizerResponse {
success: true,
token_count,
message: String::new(),
status: TokenizerStatus::Available,
stored_path: String::new(),
}
}
fn stored(stored_path: String) -> Self {
TokenizerResponse {
success: true,
token_count: 0,
message: String::new(),
status: TokenizerStatus::Available,
stored_path,
}
}
fn unavailable(reason: String) -> Self {
TokenizerResponse {
success: false,
token_count: 0,
message: reason,
status: TokenizerStatus::Unavailable,
stored_path: String::new(),
}
}
}
pub fn set_path_resolver(path_resolver: PathResolver) {
match TOKENIZER_PATH_RESOLVER.set(path_resolver) {
pub fn set_default_tokenizer_path(app_handle: tauri::AppHandle) {
let tokenizer_path = match app_handle
.path()
.resolve(DEFAULT_TOKENIZER_RESOURCE_PATH, BaseDirectory::Resource)
{
Ok(path) => path,
Err(e) => {
let reason = format!("The default tokenizer file '{DEFAULT_TOKENIZER_RESOURCE_PATH}' could not be resolved: {e}");
error!(Source = "Tokenizer"; "{reason}");
set_tokenizer_unavailable(reason);
return;
}
};
if !tokenizer_path.is_file() {
let reason = format!("The default tokenizer file was not found: {}", tokenizer_path.display());
error!(Source = "Tokenizer"; "{reason}");
set_tokenizer_unavailable(reason);
return;
}
match DEFAULT_TOKENIZER_PATH.set(tokenizer_path) {
Ok(_) => (),
Err(e) => warn!(Source = "Tokenizer"; "Could not set the path resolver: {:?}", e),
Err(e) => warn!(Source = "Tokenizer"; "Could not set the default tokenizer path: {:?}", e),
}
}
fn tokenizer_state() -> &'static RwLock<Option<Tokenizer>> {
TOKENIZER.get_or_init(|| RwLock::new(None))
pub async fn tokenizer_info(_token: APIToken) -> Json<TokenizerResponse> {
let status = TOKENIZER_STATUS.lock().unwrap();
match status.status {
TokenizerStatus::Available => Json(TokenizerResponse::available(0)),
TokenizerStatus::Running => Json(TokenizerResponse {
success: false,
token_count: 0,
message: String::new(),
status: TokenizerStatus::Running,
stored_path: String::new(),
}),
TokenizerStatus::Unavailable => Json(TokenizerResponse::unavailable(status.unavailable_reason.clone().unwrap_or_default())),
}
}
pub fn handle_tokenizer_set(path: &str) -> Result<(), Error> {
let tokenizer_path = if path.trim().is_empty() {
let relative_source_path = String::from("resources/tokenizers/tokenizer.json");
let path_resolver = TOKENIZER_PATH_RESOLVER
.get()
.ok_or_else(|| Error::from("Tokenizer path resolver is not initialized"))?;
path_resolver
.resolve_resource(relative_source_path)
.ok_or_else(|| Error::from("Failed to resolve default tokenizer resource path"))?
} else {
PathBuf::from(path)
};
pub async fn token_count(_token: APIToken, req: Json<SetTokenText>) -> Json<TokenizerResponse> {
match get_token_count(&req.text) {
Ok(count) => Json(TokenizerResponse::available(count)),
Err(e) => Json(TokenizerResponse::unavailable(e)),
}
}
let tokenizer = Tokenizer::from_file(tokenizer_path)?;
let mut tokenizer_guard = tokenizer_state()
.write()
.map_err(|_| Error::from("Tokenizer state lock is poisoned"))?;
*tokenizer_guard = Some(tokenizer);
pub async fn validate_tokenizer(_token: APIToken, payload: Json<TokenizerPath>) -> Json<TokenizerResponse> {
match handle_tokenizer_validate(&PathBuf::from(payload.file_path.clone())) {
Ok(count) => Json(TokenizerResponse::available(count)),
Err(e) => Json(TokenizerResponse::unavailable(e)),
}
}
pub async fn store_tokenizer(_token: APIToken, payload: Json<TokenizerStorage>) -> Json<TokenizerResponse> {
match handle_tokenizer_store(&payload) {
Ok(dest_path) => Json(TokenizerResponse::stored(dest_path)),
Err(e) => Json(TokenizerResponse::unavailable(e.to_string())),
}
}
pub async fn delete_tokenizer(_token: APIToken, payload: Json<TokenizerDelete>) -> Json<TokenizerResponse> {
match handle_tokenizer_delete(&payload) {
Ok(_) => Json(TokenizerResponse::stored(String::new())),
Err(e) => Json(TokenizerResponse::unavailable(e.to_string())),
}
}
pub async fn set_tokenizer(_token: APIToken, payload: Json<TokenizerPath>) -> Json<TokenizerResponse> {
match handle_tokenizer_set(&payload.file_path) {
Ok(_) => Json(TokenizerResponse::available(0)),
Err(e) => Json(TokenizerResponse::unavailable(e)),
}
}
pub fn handle_tokenizer_set(path: &str) -> Result<(), String> {
let _operation_guard = begin_tokenizer_operation()?;
set_tokenizer_running();
let tokenizer_path = resolve_tokenizer_path(path).map_err(|e| {
error!(Source = "Tokenizer"; "{e} Starting the app without a tokenizer.");
unavailable_with_status_update(&e)
})?;
let tokenizer = load_tokenizer_from_file(&tokenizer_path).map_err(|e| {
error!(Source = "Tokenizer"; "{e}");
unavailable_with_status_update(&e)
})?;
match tokenizer_state().write() {
Ok(mut tokenizer_guard) => *tokenizer_guard = Some(tokenizer),
Err(_) => return Err(unavailable_with_status_update("Tokenizer state lock is poisoned.")),
}
set_tokenizer_available();
Ok(())
}
fn handle_tokenizer_validate(path: &PathBuf) -> Result<usize, TokenizerError> {
if !path.is_file() {
return Err(TokenizerError::from(format!(
"Tokenizer file was not found: {}",
path.display()
)));
fn handle_tokenizer_validate(path: &PathBuf) -> Result<usize, String> {
let _operation_guard = begin_tokenizer_operation()?;
set_tokenizer_running();
let result = validate_tokenizer_file(path);
match tokenizer_state().read() {
Ok(tokenizer_guard) if tokenizer_guard.is_some() => set_tokenizer_available(),
Ok(_) => set_tokenizer_unavailable(NO_TOKENIZER_LOADED_MESSAGE.to_string()),
Err(_) => set_tokenizer_unavailable("Tokenizer state lock is poisoned.".to_string()),
}
let tokenizer = Tokenizer::from_file(path).map_err(|e| {
TokenizerError::from(format!(
"Failed to load tokenizer from '{}': {}",
path.display(),
e
))
})?;
result
}
pub fn get_token_count(text: &str) -> Result<usize, String> {
if text.trim().is_empty() {
return Ok(0);
}
let _operation_guard = begin_tokenizer_operation()?;
{
let status = TOKENIZER_STATUS.lock().unwrap();
if status.status != TokenizerStatus::Available {
return Err(status.unavailable_reason.clone().unwrap_or_else(|| NO_TOKENIZER_LOADED_MESSAGE.to_string()));
}
}
let tokenizer_guard = tokenizer_state()
.read()
.map_err(|_| unavailable_with_status_update("Tokenizer state lock is poisoned."))?;
let tokenizer = match tokenizer_guard.as_ref() {
Some(tokenizer) => tokenizer,
None => {
drop(tokenizer_guard);
return Err(unavailable_with_status_update("Tokenizer not initialized."));
}
};
let token_count = match tokenizer.encode(text, true) {
Ok(enc) => enc.len(),
Err(e) => {
let reason = format!("Failed to tokenize text: {e}");
drop(tokenizer_guard);
return Err(unavailable_with_status_update(&reason));
}
};
Ok(token_count)
}
fn validate_tokenizer_file(path: &PathBuf) -> Result<usize, String> {
let tokenizer = load_tokenizer_from_file(path)?;
let test_string = "Hello, world! This is a test string for tokenizer validation.";
let encoding = tokenizer.encode(test_string, true).map_err(|e| {
TokenizerError::from(format!(
"Tokenizer failed to encode validation string: {}",
e
))
})?;
let encoding = tokenizer
.encode(test_string, true)
.map_err(|e| format!("Tokenizer failed to encode validation string: {e}"))?;
let token_count = encoding.len();
if token_count == 0 {
return Err(TokenizerError::from(
"Tokenizer produced 0 tokens for test string. The tokenizer is likely invalid or misconfigured.",
));
return Err("Tokenizer produced 0 tokens for test string. The tokenizer is likely invalid or misconfigured.".to_string());
}
if encoding.get_tokens().iter().any(|t| t.is_empty()) {
return Err(TokenizerError::from(
"Tokenizer produced empty tokens. The tokenizer is invalid.",
));
return Err("Tokenizer produced empty tokens. The tokenizer is invalid.".to_string());
}
Ok(token_count)
@ -147,26 +277,25 @@ fn handle_tokenizer_store(payload: &TokenizerStorage) -> Result<String, std::io:
.file_name()
.and_then(|n| n.to_str())
.ok_or_else(|| std::io::Error::new(std::io::ErrorKind::InvalidInput, "Invalid tokenizer file path"))?;
let model_path = &base_path.join(&payload.model_id);
let destination_path = &model_path.join(source_name);
let model_path = base_path.join(&payload.model_id);
let destination_path = model_path.join(source_name);
if source_path.eq(destination_path) {
return Ok(destination_path.to_str().unwrap().to_string());
if source_path.eq(&destination_path) {
return Ok(destination_path.to_string_lossy().to_string());
}
match model_path.try_exists()? {
true => fs::remove_dir_all(model_path.clone())?,
false => (),
if model_path.try_exists()? {
fs::remove_dir_all(&model_path)?;
}
if payload.file_path.trim().is_empty() {
return Ok(String::from(""));
return Ok(String::new());
}
fs::create_dir_all(model_path)?;
fs::create_dir_all(&model_path)?;
fs::copy(&source_path, &destination_path)?;
Ok(destination_path.to_str().unwrap().to_string())
Ok(destination_path.to_string_lossy().to_string())
}
fn handle_tokenizer_delete(payload: &TokenizerDelete) -> Result<(), std::io::Error> {
@ -189,74 +318,61 @@ fn handle_tokenizer_delete(payload: &TokenizerDelete) -> Result<(), std::io::Err
Ok(())
}
pub fn get_token_count(text: &str) -> Result<usize, TokenizerError> {
if text.trim().is_empty() {
return Err(TokenizerError::from("Input text is empty"));
fn tokenizer_state() -> &'static RwLock<Option<Tokenizer>> {
TOKENIZER.get_or_init(|| RwLock::new(None))
}
fn begin_tokenizer_operation() -> Result<std::sync::MutexGuard<'static, ()>, String> {
TOKENIZER_OPERATION_LOCK
.lock()
.map_err(|_| unavailable_with_status_update("Tokenizer operation lock is poisoned."))
}
fn set_tokenizer_available() {
let mut status = TOKENIZER_STATUS.lock().unwrap();
status.status = TokenizerStatus::Available;
status.unavailable_reason = None;
}
fn set_tokenizer_running() {
let mut status = TOKENIZER_STATUS.lock().unwrap();
status.status = TokenizerStatus::Running;
status.unavailable_reason = None;
}
fn set_tokenizer_unavailable(reason: String) {
let mut status = TOKENIZER_STATUS.lock().unwrap();
status.status = TokenizerStatus::Unavailable;
status.unavailable_reason = Some(reason);
}
fn unavailable_with_status_update(reason: &str) -> String {
let reason = reason.to_string();
match tokenizer_state().write() {
Ok(mut tokenizer_guard) => *tokenizer_guard = None,
Err(_) => set_tokenizer_unavailable("Tokenizer state lock is poisoned.".to_string()),
}
let tokenizer = tokenizer_state()
.read()
.map_err(|_| TokenizerError::from("Tokenizer state lock is poisoned"))?
.clone()
.ok_or_else(|| TokenizerError::from("Tokenizer not initialized"))?;
let enc = tokenizer.encode(text, true)?;
Ok(enc.len())
set_tokenizer_unavailable(reason.clone());
reason
}
#[post("/tokenizer/count", data = "<req>")]
pub fn token_count(_token: APIToken, req: Json<SetTokenText>) -> Json<TokenizerResponse> {
Json(get_token_count(&req.text).into())
}
#[post("/tokenizer/validate", data = "<payload>")]
pub fn validate_tokenizer(_token: APIToken, payload: Json<TokenizerPath>) -> Json<TokenizerResponse> {
Json(handle_tokenizer_validate(&PathBuf::from(payload.file_path.clone())).into())
}
#[post("/tokenizer/store", data = "<payload>")]
pub fn store_tokenizer(_token: APIToken, payload: Json<TokenizerStorage>) -> Json<TokenizerResponse> {
match handle_tokenizer_store(&payload) {
Ok(dest_path) => Json(TokenizerResponse {
success: true,
token_count: 0,
message: dest_path,
}),
Err(e) => Json(TokenizerResponse {
success: false,
token_count: 0,
message: e.to_string(),
}),
fn resolve_tokenizer_path(path: &str) -> Result<PathBuf, String> {
if !path.trim().is_empty() {
return Ok(PathBuf::from(path));
}
DEFAULT_TOKENIZER_PATH
.get()
.cloned()
.ok_or_else(|| "Default tokenizer path is not initialized.".to_string())
}
#[post("/tokenizer/delete", data = "<payload>")]
pub fn delete_tokenizer(_token: APIToken, payload: Json<TokenizerDelete>) -> Json<TokenizerResponse> {
match handle_tokenizer_delete(&payload) {
Ok(_) => Json(TokenizerResponse {
success: true,
token_count: 0,
message: "Success".to_string(),
}),
Err(e) => Json(TokenizerResponse {
success: false,
token_count: 0,
message: e.to_string(),
}),
fn load_tokenizer_from_file(path: &PathBuf) -> Result<Tokenizer, String> {
if !path.is_file() {
return Err(format!("Tokenizer file was not found: {}", path.display()));
}
}
#[post("/tokenizer/set", data = "<payload>")]
pub fn set_tokenizer(_token: APIToken, payload: Json<TokenizerPath>) -> Json<TokenizerResponse> {
match handle_tokenizer_set(&payload.file_path) {
Ok(_) => Json(TokenizerResponse {
success: true,
token_count: 0,
message: "Success".to_string(),
}),
Err(e) => Json(TokenizerResponse {
success: false,
token_count: 0,
message: e.to_string(),
}),
}
Tokenizer::from_file(path)
.map_err(|e| format!("Failed to load tokenizer from '{}': {e}", path.display()))
}