System.ArgumentException: Attempt to insert un-indexed tag field: '__part_n', will not be able to filter on it, please adjust the tag settings in your Redis Configuration

[hemantkathuria]

I am trying to set up Kernel-Memory with Semantic Kernel and Redis in Blazor application. I have created enterprise cache at Azure.

I'm trying to import one PDF first so that vectors are stored in Redis cache.

  await KernelMemory.ImportDocumentAsync(path, e.File.Name);

Getting the below error.

info: Microsoft.KernelMemory.Pipeline.InProcessPipelineOrchestrator[0]
      Queueing upload of 1 files for further processing [request employee_handbook.pdf]
info: Microsoft.KernelMemory.Pipeline.InProcessPipelineOrchestrator[0]
      File uploaded: employee_handbook.pdf, 142977 bytes
info: Microsoft.KernelMemory.Pipeline.InProcessPipelineOrchestrator[0]
      Handler 'extract' processed pipeline 'default/employee_handbook.pdf' successfully
info: Microsoft.KernelMemory.Pipeline.InProcessPipelineOrchestrator[0]
      Handler 'partition' processed pipeline 'default/employee_handbook.pdf' successfully
info: Microsoft.KernelMemory.Pipeline.InProcessPipelineOrchestrator[0]
      Handler 'gen_embeddings' processed pipeline 'default/employee_handbook.pdf' successfully
fail: Microsoft.KernelMemory.MemoryDb.Redis.RedisMemory[0]
      Attempt to insert un-indexed tag field: '__part_n', will not be able to filter on it, please adjust the tag settings in your Redis Configuration
fail: Microsoft.KernelMemory.Pipeline.InProcessPipelineOrchestrator[0]
      Pipeline start failed
      System.ArgumentException: Attempt to insert un-indexed tag field: '__part_n', will not be able to filter on it, please adjust the tag settings in your Redis Configuration
         at Microsoft.KernelMemory.MemoryDb.Redis.RedisMemory.UpsertAsync(String index, MemoryRecord record, CancellationToken cancellationToken)
         at Microsoft.KernelMemory.Handlers.SaveRecordsHandler.SaveRecordAsync(DataPipeline pipeline, IMemoryDb db, MemoryRecord record, HashSet`1 createdIndexes, CancellationToken cancellationToken)
         at Microsoft.KernelMemory.Handlers.SaveRecordsHandler.InvokeAsync(DataPipeline pipeline, CancellationToken cancellationToken)
         at Microsoft.KernelMemory.Pipeline.InProcessPipelineOrchestrator.RunPipelineAsync(DataPipeline pipeline, CancellationToken cancellationToken)
         at Microsoft.KernelMemory.Pipeline.BaseOrchestrator.ImportDocumentAsync(String index, DocumentUploadRequest uploadRequest, IContext context, CancellationToken cancellationToken)

This how I am setting up

builder.Services.AddAzureOpenAIChatCompletion(ApplicationConstants.C_DeploymentChatName, ApplicationConstants.C_Endpoint, new DefaultAzureCredential());

builder.Services.AddKernelMemory((builder) => {
    builder.WithAzureOpenAITextEmbeddingGeneration(MemoryHelper.GetEmbeddingConfig(ApplicationConstants.C_DeploymentEmbeddingName,ApplicationConstants.C_Endpoint));
    builder.WithAzureOpenAITextGeneration(MemoryHelper.GetEmbeddingConfig(ApplicationConstants.C_DeploymentChatName, ApplicationConstants.C_Endpoint));
    builder.WithRedisMemoryDb("connectionstring");  


});

[hemantkathuria]

Any thoughts on the above? @dluc

[dluc]

see @magols answer below :-)

[magols]

Had similar error with the __part_n.
I had to add the following to the tags:

"Tags": {
"__part_n": ",",
"collection": ","
}

Now it works for me without warnings

[dluc]

correct, the Redis implementation requires to configure the tags being used. Hopefully this limitation will be removed in a future refactoring.

[kapilkhub]

Had similar error with the __part_n. I had to add the following to the tags:

"Tags": { "__part_n": ",", "collection": "," }

Now it works for me without warnings

hi @magols could you please paste the solution, how you modified redis configuratuon

[kapilkhub]

"Tags": { "__part_n": ",", "collection": "," }

I think I know how :)