MindWorkAI/AI-Studio
MindWorkAI/AI-Studio
1
0
Fork 0
mirror of https://github.com/MindWorkAI/AI-Studio.git synced 2025-04-28 08:19:47 +00:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

Fixed the chat thread we use for the data retrieval

Browse Source
This commit is contained in:
Thorsten Sommer 2025-03-08 11:32:38 +01:00
parent df01ce188e
commit 51bf077485
Signed by: tsommer
GPG Key ID: 371BBA77A02C0108
2 changed files with 10 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
app/MindWork AI Studio/Tools/RAG/RAGProcesses/AISrcSelWithRetCtxVal.cs
View File

@ -79,12 +79,19 @@ public sealed class AISrcSelWithRetCtxVal : IRagProcess
var dataContexts = new List<IRetrievalContext>();
if (proceedWithRAG)
{
//
// At this point in time, the chat thread contains already the
// last block, which is the waiting AI block. We need to remove
// this block before we call the RAG process:
//
var chatThreadWithoutWaitingAIBlock = chatThread with { Blocks = chatThread.Blocks[..^1] };
//
// We kick off the retrieval process for each data source in parallel:
//
var retrievalTasks = new List<Task<IReadOnlyList<IRetrievalContext>>>(selectedDataSources.Count);
foreach (var dataSource in selectedDataSources)
retrievalTasks.Add(dataSource.RetrieveDataAsync(lastPrompt, chatThread, token));
retrievalTasks.Add(dataSource.RetrieveDataAsync(lastPrompt, chatThreadWithoutWaitingAIBlock, token));
//
// Wait for all retrieval tasks to finish:

3
app/MindWork AI Studio/wwwroot/changelog/v0.9.32.md
View File

@ -3,4 +3,5 @@
- Improved the ERI client to expect JSON responses and send JSON requests using camel case.
- Improved the ERI client to raise an error when the server responds with additional JSON data that is not expected.
- Improved the error handling in the ERI data source info dialog in cases where servers respond with an invalid message.
- Improved the error handling for the entire RAG process.
- Improved the error handling for the entire RAG process.
- Fixed the chat thread we use for the data retrieval by removing the last block, which is meant to be for the final AI answer.