https://github.com/rickcau/semantic-kernel-console-demos
My playground for Semantic Kernel Console Demos
https://github.com/rickcau/semantic-kernel-console-demos
assistants function-calling plugin semantic-kernel sk speech-to-text streaming
Last synced: 2 months ago
JSON representation
My playground for Semantic Kernel Console Demos
- Host: GitHub
- URL: https://github.com/rickcau/semantic-kernel-console-demos
- Owner: Rickcau
- License: mit
- Created: 2024-02-14T18:13:29.000Z (over 1 year ago)
- Default Branch: master
- Last Pushed: 2025-01-01T19:27:20.000Z (9 months ago)
- Last Synced: 2025-04-23T17:11:32.356Z (6 months ago)
- Topics: assistants, function-calling, plugin, semantic-kernel, sk, speech-to-text, streaming
- Language: C#
- Homepage:
- Size: 83 KB
- Stars: 4
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Semantic Kernel Console App Example Playground
In this Solution / Repo I have several Console Apps that I demo various SK examples. I often leverage console apps to learn new concepts and to play with different patterns as it's a quick way to get up an running. I hope you find this repo useful.
## LICENSE
The [MIT License](./LICENSE) applies to all the solutions / projects / examples included in this repository.
## Console-SK-DeFi-Assistant
In this example, I create a custom SK plugin for the Uniswap V3 Subgraph. I also make use of ToolCallBehavior.AutoInvokeKernelFunctions setting so the Kernel will automatically determine if there is a Native Function it needs to invoke.
When you run the example, you can ask general questions like you would with any ChatBot, but if you ask a question that aligns to the description of the SK Function the Kernel will invoke the function passing in a GraphQL query. What is interesting about this, is the the AI will general the GraphQL query for you, or you can actually provide a GraphQL query.
In this example, I am not streaming the Chat Completion so you have to wait for the AI to finish. If you are interesting in seeing a Streaming example, take a look at the next example.
## Console-SK-DeFi-Assistant-Streaming
This is the same example as the above, but I have implementing Streaming logic, which results in a better Chat experience, and would be the recommended approach for a production solution.
## Console-SK-Multiple-Function-Calling-Assistant
In this example, I demo how you can nest multiple Native Functions and have the Kernel automatically call each one based on the prompt. If you have a use case that has very specific tasks that need to be invoked depending on the scenario, this is a super easy to understand example that will get you headed in the right direction.
## Notes
This concept could be used with any GraphQL endpoint. I will explore with other GraphQL endpoints when I have time, but the UniSwap V3 GraphQL endpoint is very well know in the Crypto space.
## Requirements for this example
1. Rename the App.config.bak to App.config
2. Modify the App.config with your LLM/AI details.
~~~
~~~