How to use SK (.Net/C#) to connect to a local Chroma database, and then pass the question&result to the LLM model??

[orchardc]

Using Python and LangChain, I can connect to a local Chroma database, and then pass the question&result to the LLM model.

Question: How to use SK to do the same thing (code below) in .Net/C# ?

vectordb = Chroma(persist_directory=persist_directory, embedding_function=embeddings)

llm = ChatOpenAI(model_name="gpt-3.5-turbo-0301", temperature=0)
qa_chain = RetrievalQA.from_chain_type(llm=llm,
                                    chain_type="stuff",
                                    retriever=vectordb.as_retriever(),
                                    return_source_documents=True,
                                    chain_type_kwargs={
                                        "verbose": False,
                                        "prompt": prompt,
                                        "memory": memory
                                    })

I want to use SK (.Net/C#) to replace Python/LangChain but found it's not straightforward.
Thanks!

[EvanChaki_MSFT (evchaki)]

Check out this sample - https://github.com/microsoft/semantic-kernel/blob/main/dotnet/samples/KernelSyntaxExamples/Example15_TextMemoryPlugin.cs

[Federico Laurianti (Laurianti)]

I know this discussion is old and closed, but since people still land here, a pointer for today: Semantic Kernel works with the Microsoft.Extensions.VectorData abstractions, and the Chroma implementation is CommunityToolkit.VectorData.Chroma, in the AI Community Toolkit. The case above, a local Chroma as the source the model answers from, is this sample; this variant runs fully local with Ollama.