Dieser Browser wird nicht mehr unterstützt.
Führen Sie ein Upgrade auf Microsoft Edge durch, um die neuesten Features, Sicherheitsupdates und den technischen Support zu nutzen.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(86.4, 52%, 18%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ETA%3C/text%3E%3C/svg%3E)
Hi guys, we created a rag chatbot in teams. We used the template from teams and used our custom vector-database as datasource. our problem is that we dont want the chat to create a general answer from llm, when no chunks are returned from data source. How can we control that?
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(230.39999999999998, 60%, 32%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ESD%3C/text%3E%3C/svg%3E)
Hello,
I can suggest you to modify the query logic. You can modify the function that processes the user input. Ensure it incorporates this check before proceeding to invoke the llm for a general answer.
BR,
Hello,
the query logic is inside the framework function. We have currently no control on that:
prompts = PromptManager(PromptManagerOptions(prompts_folder=f"{os.getcwd()}/prompts"))
vector_db = VectorDataSource(
MyDataSourceOptions(
name="qdrant",
options=AzureOpenAIEmbeddingsOptions(
azure_api_key=config.AZURE_OPENAI_API_KEY,
azure_endpoint=config.AZURE_OPENAI_ENDPOINT,
azure_deployment=config.AZURE_OPENAI_EMBEDDING_MODEL_NAME
)
)
)
prompts.add_data_source(vector_db)
planner = ActionPlanner(
ActionPlannerOptions(model=model, prompts=prompts, default_prompt="chat")
)
# Define storage and application
storage = MemoryStorage()
bot_app = Application[TurnState](
ApplicationOptions(
bot_app_id=config.APP_ID,
storage=storage,
adapter=TeamsAdapter(config),
ai=AIOptions(planner=planner),
)
)