Custom reference links provided by chatbot pointing to public URLs.

Himanka 0 Reputation points
2024-07-02T05:07:57.8666667+00:00

Hello,

I am trying to build a text generation tool using resources deployed in Azure. My resources list in Azure is as follows:

  1. Azure OpenAI ada embedding, gpt3.5 and gpt4 models.
  2. Azure Blob storage which contains product files (.pdf, .html, etc.) imported from our Alfresco and Salesforce databases.
  3. Azure AI search index created from the above blob storage.

I am using the resources through endpoints in my local development environment and my python based application is being deployed in a private Windows server. While the setup is working fine and I am able to generate answers based on input prompts from user, the references are not pointing to the public URLs. The references are almost always pointing to the Blob storage documents, which should not be the case. The referencing should be done to the public version of the documents with public URLs, which are stored in the Alfresco and Salesforce databases.

Maybe this is happening because the files are being indexed from the blob storage and not the original source and somehow the original URLs are not accessed or indexed during the indexing process?

Please provide more insights and help me to solve this issue.

Azure AI Search
Azure AI Search
An Azure search service with built-in artificial intelligence capabilities that enrich information to help identify and explore relevant content at scale.
831 questions
Azure Blob Storage
Azure Blob Storage
An Azure service that stores unstructured data in the cloud as blobs.
2,575 questions
Azure OpenAI Service
Azure OpenAI Service
An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
2,536 questions
Azure AI services
Azure AI services
A group of Azure services, SDKs, and APIs designed to make apps more intelligent, engaging, and discoverable.
2,577 questions
0 comments

1 answer

Sort by: Most helpful
Most helpful Newest Oldest
  1. Amira Bedhiafi 18,501 Reputation points
    2024-07-03T14:02:23.0766667+00:00

    The Azure AI Search index you've created is likely indexing documents directly from Azure Blob storage. This means it's referencing the documents as they are stored there, which may not include the public URLs from Alfresco and Salesforce databases where the original documents reside.

    So you may need to modify your indexing process to fetch documents directly from Alfresco and Salesforce databases rather than Blob storage :

    • Update the Azure AI Search indexer configuration to point to the original data sources.
    • Verify that the documents fetched include metadata or fields that contain the public URLs.

    If fetching directly from Alfresco and Salesforce isn't feasible, consider adding a field to your Azure AI Search index that stores the public URLs. You can manually populate this field during the indexing process or through a batch update mechanism.

    0 comments