Introduce support for `CancellationToken` across various methods to allow for task cancellation and improve responsiveness.
- Update `DecodeAsync` method in `DocxContentDecoder.cs`, `PdfContentDecoder.cs`, `TextContentDecoder.cs`, and `IContentDecoder.cs` to include an optional `CancellationToken` parameter.
- Modify endpoint handlers in `Program.cs` to accept and pass `CancellationToken` parameters.
- Update methods in `ChatService.cs` to include `CancellationToken` parameters.
- Update methods in `DocumentService.cs` to include `CancellationToken` parameters.
- Update methods in `VectorSearchService.cs` to include `CancellationToken` parameters.
These changes ensure that long-running operations can be canceled if needed, improving the application's ability to handle cancellation requests gracefully.
The response format for the `/api/ask-streaming` endpoint has been simplified by removing multiple individual elements that contained parts of the answer.
The README.md file was updated to correct the terminology from "tag" to "property" for the *streamState* attribute.
Updated README.md to document new `/api/documents`, `/api/ask`, and `/api/ask-streaming` endpoints, including features like conversation history, token usage, and response streaming. Enhanced Program.cs by adding new endpoints for asking questions and streaming responses, with additional metadata. Updated `documentsApiGroup` to include new document management endpoints.
Refactored Program.cs to use AddAzureSql with new options. Added VectorSearchService and DocumentService as scoped services. Updated documentsApiGroup to use DocumentService for document operations and added a delete document endpoint. Moved document-related methods from VectorSearchService to new DocumentService for better separation of concerns.
- Replace `TotalTokenCount` with `EmbeddingTokenCount` in `ImportDocumentResponse`.
- Add `OriginalQuestion` and `ReformulatedQuestion` fields to `QuestionResponse` and a new constructor.
- Add a new constructor to `TokenUsageResponse` to initialize `Question`.
- Add `TextChunkerService` to service collection in `Program.cs`.
- Clarify prompt and update token counting in `ChatService`.
- Differentiate token counting in `TokenizerService` with `CountChatCompletionTokens` and `CountEmbeddingTokens`.
- Update `VectorSearchService` to use `TextChunkerService` and new token counting methods.
- Introduce `TextChunkerService` for text splitting and tokenization.
Upgraded Microsoft.SemanticKernel package from version 1.34.0 to 1.35.0 in SqlDatabaseVectorSearch.csproj. This update includes new features, bug fixes, and other improvements provided in the latest version.
Updated README.md to reflect support for PDF, DOCX, and TXT files.
Removed commented-out code in DocxContentDecoder.cs.
Added TextContentDecoder service in Program.cs and updated exception handling middleware.
Updated document upload endpoint description in Program.cs.
Modified VectorSearchService to throw NotSupportedException for unsupported content types.
Added TextContentDecoder class in TextContentDecoder.cs.
- Added `using` statements in `Program.cs` for new content decoding.
- Registered new content decoder services in `builder.Services`.
- Modified `documentsApiGroup.MapPost` to pass `file.ContentType`.
- Refactored `VectorSearchService` to use `IServiceProvider` and handle content types.
- Added `DocumentFormat.OpenXml` package reference.
- Created `DocxContentDecoder` and `PdfContentDecoder` classes.
- Created `IContentDecoder` interface.
- Clarified comment in `ChatService.cs`.
- Added `TokenizerService` and `ILogger` parameters to `VectorSearchService` constructor.
- Updated paragraph splitting to use `tokenizerService.CountTokens`.
- Added logging for token count of each paragraph in `VectorSearchService`.
- Updated `ModelId` comment in `appsettings.json` to include "gpt-4o-mini".
- Changed `MaxTokensPerParagraph` in `appsettings.json` from 1024 to 1000.
Eliminated the 50ms delay in the asynchronous streaming response
logic of the /api/ask-streaming endpoint by removing the line
`await Task.Delay(50);`. This change improves the response time
by removing unnecessary latency after each delta in the response
stream.
- Updated `Response` record class to allow nullable `Question` and `Answer` properties; moved `StreamState` enum to a new file.
- Added `JsonStringEnumConverter` in `Program.cs` for better enum serialization.
- Corrected terminology in document upload endpoint description.
- Introduced `/api/ask-streaming` endpoint for streaming question responses.
- Added `AskStreamingAsync` method in `VectorSearchService` for handling streaming logic.
- Created `StreamState.cs` to define `StreamState` enum with `Start`, `Append`, and `End` values.
Updated `Response` record in `Response.cs` to include an optional `StreamState` property, which can be `Start`, `Append`, or `End`. Added a new `StreamState` enum to `Response.cs`.
In `ChatService.cs`, added new methods `AskQuestionAsync` and `AskStreamingAsync` to handle asking questions and streaming responses, respectively. Refactored `CreateChatAsync` to return a `ChatHistory` object.
In `VectorSearchService.cs`, added a new `AskQuestionAsync` method to handle questions using `ChatService`. Updated `CreateContextAsync` to return a tuple with the reformulated question and chunks. Removed the previous implementation of `AskQuestionAsync` and replaced it with the new method utilizing `ChatService`.
Upgraded the Microsoft.SemanticKernel package from version 1.33.0 to 1.34.0 in the SqlDatabaseVectorSearch.csproj file. This update may include bug fixes, new features, or other improvements.
Modified Program.cs to use AddAzureSql for ApplicationDbContext with UseVectorSearch and NoTracking options. Updated EFCore.SqlServer.VectorSearch package to version 9.0.0-preview.2 in SqlDatabaseVectorSearch.csproj.
- Added HTTP client defaults configuration in Program.cs.
- Adjusted OpenAPI configuration order in Program.cs.
- Changed MaxTokensPerParagraph in AppSettings.cs to 1000.
- Added Microsoft.Extensions.Http.Resilience package in SqlDatabaseVectorSearch.csproj.
- Reordered and added ModelId properties in appsettings.json.
Updated `Program.cs` to always execute OpenAPI setup, including `app.MapOpenApi()` and `app.UseSwaggerUI`, and added `options.RemoveServerList()` to `builder.Services.AddOpenApi`. Simplified `app.UseSwaggerUI` call.
In `SqlDatabaseVectorSearch.csproj`, updated `Microsoft.ML.Tokenizers` to `1.0.1`, `Microsoft.ML.Tokenizers.Data.Cl100kBase` and `Microsoft.ML.Tokenizers.Data.O200kBase` to `1.0.1`, `MinimalHelpers.OpenApi` to `2.1.3`, and `TinyHelpers.AspNetCore` to `4.0.15`.
- Updated EFCore.SqlServer.VectorSearch from 0.2.0 to 9.0.0-preview.1
- Updated Microsoft.AspNetCore.OpenApi from 9.0.0 to 9.0.1
- Updated Microsoft.EntityFrameworkCore.SqlServer from [8.0.11,9.0.0) to 9.0.1
- Updated Microsoft.Extensions.Caching.Hybrid from 9.0.0-preview.9.24556.5 to 9.1.0-preview.1.25064.3
The `ChatService` class now includes a private readonly field `appSettings`, initialized via the constructor. A new method `CreateQuestionAsync` has been added to handle question creation asynchronously, leveraging chat history. Additionally, a comment has been updated for clarity regarding chunk limitations.
Updated Program.cs to use ConfigureAndGet for settings and added TokenizerService singleton. Modified ChatService to include TokenizerService and updated prompt logic for token limits. Added TokenizerService class using TiktokenTokenizer. Updated AppSettings and AzureOpenAISettings with new token and model settings. Updated SqlDatabaseVectorSearch.csproj with new package references and version bump. Updated appsettings.json with new settings.
Enhanced comments in ChatService for token counting clarity.
Updated Microsoft.SemanticKernel to 1.33.0 in SqlDatabaseVectorSearch.csproj.
Added possible values comment for ModelId in appsettings.json.
Upgraded Swashbuckle.AspNetCore.SwaggerUI from 7.1.0 to 7.2.0.
Upgraded TinyHelpers.AspNetCore from 4.0.5 to 4.0.6.
These updates may include bug fixes, performance improvements,
or new features.
Updated Program.cs to replace Swagger with OpenApi and MemoryCache with HybridCache. Refactored ChatService.cs to use HybridCache asynchronously. Removed MessageLimit from AppSettings.cs and appsettings.json. Updated SqlDatabaseVectorSearch.csproj to include HybridCache package and update dependencies.
The `AppSettings` class in `AppSettings.cs` has been modified to
remove the `MessageLimit` and `MessageExpiration` properties.
Correspondingly, the `appsettings.json` file has been updated
to remove the `MessageLimit` and `MessageExpiration` settings
under the `AppSettings` section to maintain consistency.
Updated Program.cs to use HybridCache and improved Swagger setup.
Refactored ChatService to simplify chat history management.
Updated SqlDatabaseVectorSearch.csproj to target net9.0 and
updated package references.
Added EnableRetryOnFailure option to the AddSqlServer method in Program.cs. This change configures the SQL Server connection to automatically retry on failure, improving the application's resilience to transient faults.
Enhanced ApplicationDbContext configuration by adding EnableRetryOnFailure to improve resilience against transient SQL Server failures. Existing configurations for UseVectorSearch and NoTracking remain unchanged.
Adjusted the version range specification for the `Microsoft.EntityFrameworkCore.SqlServer` package in `SqlDatabaseVectorSearch.csproj` by removing the space between the version numbers and the comma. This change improves the readability and uniformity of the code without altering the functional behavior of the version constraint.
Updated Microsoft.SemanticKernel from 1.30.0 to 1.31.0.
Updated Swashbuckle.AspNetCore from 7.0.0 to 7.1.0.
Updated TinyHelpers.AspNetCore from 3.1.19 to 4.0.4.
Added TinyHelpers.AspNetCore.Swashbuckle version 4.0.5.
Updated Microsoft.SemanticKernel from 1.30.0 to 1.31.0.
Updated Swashbuckle.AspNetCore from 7.0.0 to 7.1.0.
Updated TinyHelpers.AspNetCore from 3.1.19 to 4.0.2.
Added TinyHelpers.AspNetCore.Swashbuckle version 4.0.3.
- Updated .editorconfig with new C# style preferences.
- Modified using directives in VectorSearchService.cs.
- Changed tuple element order in VectorSearchService.cs foreach loop.
- Updated SqlDatabaseVectorSearch.csproj to target .NET 9.0.
- Updated package references and removed TinyHelpers package.
Update .editorconfig, fix VectorSearchService, upgrade packages
Updated .editorconfig with new C# style preferences. Removed `TinyHelpers.Extensions` using directive and corrected variable order in `VectorSearchService.cs`. Upgraded target framework to `net9.0` and updated several package references in `SqlDatabaseVectorSearch.csproj`.
Updated README.md to include a note on Vector support in Azure SQL Database/Managed Instance (EAP) and corrected "Dimension" to "Dimensions". Adjusted prompt formatting in ChatService.cs by replacing "---" with "=====" and improved text chunk appending. Added new package reference for `Microsoft.ML.Tokenizers.Data.Cl100kBase` in SqlDatabaseVectorSearch.csproj.
Updated README.md to include a note on Vector support in Azure SQL Database and corrected "Dimension" to "Dimensions". Modified ChatService.cs to append separator "---" only once at the end of the prompt.
- Changed ChatService registration in Program.cs from scoped to singleton.
- Modified ChatHistory initialization in ChatService.cs by removing unnecessary line breaks.
- Renamed loop variable in ChatService.cs from 'result' to 'text' for better readability.
- Updated Microsoft.SemanticKernel package version from 1.26.0 to 1.27.0 in SqlDatabaseVectorSearch.csproj.
Updated Program.cs to use ConfigureAndGet method for settings,
changed ChatService to singleton, and added TokenizerService
singleton. Modified ChatService to use TokenizerService for
token counting. Updated AppSettings and AzureOpenAISettings
with new properties. Added new package references in
SqlDatabaseVectorSearch.csproj. Updated appsettings.json with
new properties. Added TokenizerService class for token counting.
Swapped the order of `IsRequired` and `HasColumnType("vector(1536)")` method calls for the `Embedding` property in the `ApplicationDbContext` class. This change improves readability and adheres to a preferred coding convention without altering functionality.
The README.md file has been updated to specify that Vectors are saved and retrieved using the `EFCore.SqlServer.VectorSearch` library with Entity Framework Core. Additionally, the note about using straight SQL has been moved to a separate section and rephrased for better clarity.
Marco Minerva