kbeaugrand / SemanticKernel.Connectors.Memory.SqlServer

SQL Server connector for Semantic Kernel plugin and Kernel Memory
MIT License
46 stars 10 forks source link

SQL Server Connection successful but when I ask always receive 'INFO NOT FOUND' message #107

Closed patorbor closed 7 months ago

patorbor commented 7 months ago

Hi @kbeaugrand ,

The connection to SQL Server is successful, but when I ask always receive "INFO NOT FOUND" message.

What I'm missing.

For example, in the database I have a table called 'Operarios'. when I ask "How much operarios are?" "INFO NOT FOUND" message is received

I test with this code (OpenAI api key and SQL sa password are hidden for security):

        memory = new KernelMemoryBuilder()
            .WithOpenAIDefaults("<OpenAI api key>")
            .WithSqlServerMemoryDb("Server=(local)\\DISCOM;Database=Visdeltex;User Id=sa;Password=<SAPassword>;TrustServerCertificate=True;")
            .Build<MemoryServerless>();
kbeaugrand commented 7 months ago

Hello,

These connectors are not NL 2 SQL features, they are supposed to be used by Retrieval Augmented Generation. You will therefore need to vectorize and index your data to use SemanticKernel or KernelMemory.

Take a look at KernelMemory to see what it's all about.