Azure AI content filter problem

[szoke120]

Hi!
As I wrote in this discussion: link, when processing some emails, the content filter simply turns on and I don't get a response, just an error message:

java.lang.IllegalArgumentException: text cannot be null or blank
	at dev.langchain4j.internal.Exceptions.illegalArgument(Exceptions.java:19) ~[langchain4j-core-0.28.0.jar:na]
	at dev.langchain4j.internal.ValidationUtils.ensureNotBlank(ValidationUtils.java:98) ~[langchain4j-core-0.28.0.jar:na]
	at dev.langchain4j.data.message.AiMessage.<init>(AiMessage.java:30) ~[langchain4j-core-0.28.0.jar:na]
	at dev.langchain4j.data.message.AiMessage.from(AiMessage.java:100) ~[langchain4j-core-0.28.0.jar:na]
	at dev.langchain4j.data.message.AiMessage.aiMessage(AiMessage.java:127) ~[langchain4j-core-0.28.0.jar:na]
	at dev.langchain4j.model.azure.InternalAzureOpenAiHelper.aiMessageFrom(InternalAzureOpenAiHelper.java:247) ~[langchain4j-azure-open-ai-0.28.0.jar:na]
	at dev.langchain4j.model.azure.AzureOpenAiChatModel.generate(AzureOpenAiChatModel.java:250) ~[langchain4j-azure-open-ai-0.28.0.jar:na]
	at dev.langchain4j.model.azure.AzureOpenAiChatModel.generate(AzureOpenAiChatModel.java:207) ~[langchain4j-azure-open-ai-0.28.0.jar:na]
	at dev.langchain4j.service.DefaultAiServices$1.invoke(DefaultAiServices.java:129) ~[langchain4j-0.28.0.jar:na]
	at jdk.proxy2/jdk.proxy2.$Proxy123.chat(Unknown Source) ~[na:na]

Today I added an extra content filter to this deployment and something interesting happened.
So far, it has filtered out this particular email as follows: {"filtered":true,"severity":"medium"}
Now that I have raised the filter threshold to "high", it filters it out as follows: {"filtered":true,"severity":"high"}

What should I change so that the content filtering does not start or what should I do to get a response besides it?

[jdubois]

Hi, this should be supported in the next release thanks to #764