Hi, Im bussy to explorer the AWS Bedrock Chatbot. In AWS i can test my chatbot with my own data set that i provided due a web crawl, but when i use it in the Mendix envorinment (https://marketplace.mendix.com/link/component/227926 starter app) i got errors which i dont know how to fix. I provided my knowledge base in the bot configurator. They will be found in the dropdown. But when i start a chat conversation i always get <error> back in the chat. In my Mendix console i see the following error: ProviderConfig_ExecuteAction: com.mendix.systemwideinterfaces.MendixRuntimeException: java.lang.NullPointerException: Cannot invoke "software.amazon.awssdk.services.bedrockagentruntime.model.RetrievalResultS3Location.uri()" because the return value of "software.amazon.awssdk.services.bedrockagentruntime.model.RetrievalResultLocation.s3Location()" is null at AmazonBedrockConnector.RetrieveAndGenerate (JavaAction : 'RetrieveAndGenerate') at MyFirstAIBot.ChatContext_BedrockRetrieveAndGenerate_ActionMicroflow (SubMicroflow : 'Retrieve And Generate') at {"name":"ConversationalUI.ProviderConfig_ExecuteAction","type":"JavaAction"} at ConversationalUI.ChatContext_ExecuteAction (JavaAction : 'ProviderConfig_ExecuteAction') I think something needs to be configured in AWS? Anyone got an idea to help me forward? When i work with only S3 bucket with a pdf uploaded it works fine. But when i add an extra datasource (webcrawler) i get this error.
asked
Sonny van Rein
2 answers
2
Hi Sonny,
there recently have been changes to the available knowledgebase sources on AWS side which resulted in a changed response body. We are currently working on upgrading the Bedrock Connector and GenAICommons module to incorporate those changes so that other knowledgebase sources such as a webcrawler (or sharepoint or confluence or ...) will be possible next to S3 soon. Thank you for bringing this to our attention!
Best regards,
Lina
answered
Strategic Alliance Development Team
1
Hi,
there has been a new release of bedrock but it does not yet support other data sources for the knowledge bases. We will let you know as soon as we have a release where there at least is no error when using a data source other than S3, or where the full range of possible data sources is available.