SemanticKernel-Basics

Code examples of Semantic Kernel with Java SDK

Detailed Blog is in here

Pre reqs to run the samples

1. Check out Semantic Kernel Java Branch and run mvn install. 2. The samples are based on the checkout from 8-Aug-2023
2. Create Azure Open AI resource in Azure portal and deploy 'gpt-35-turbo' and 'text-embedding-ada-002' models there.
3. Optionally, create Azure Congnitive Search resource, if you want to run the cognitive search sample (Example06_SKwithCognitiveSearch.class).
4. Rename src/main/resources/conf.properties.example to src/main/resources/conf.properties and update it with the details from Step #2 and #3.

Samples

1. Example01_InlineFunction.class > Demonstrates, how to initialize a Kernel and use an Inline function with Azure Open AI(AOAI).
2. Example02_SummarizerAsPrompt.class > Demonstrates, how to use a Skill and Function from skills directory with a skprompt.txt and config.json.
3. Example03_SummarizerAsPlugin.class > Demonstrates, how to use a Skill and Function defined as plugin.
4. Example04_SkillsPipeline.class > Demonstrates, how to use a Skill and Function in a pipeline or chain, passing output from one function to another as input.
5. Example05_SKwithMemory.class > Demonstrates, how to use a volatile memory (local simulation of a vector DB) with AOAI.
6. Example06_SKwithCognitiveSearch.class > Demonstrates, how to use a Azure Cognitive Search with AOAI.
7. Example08_SequentialPlanner.class > Demonstrates, how to use a bunch of Skills together and instruct the LLM to pick intelligently a or combination of Skills depending on the input provided.

Bonus Sample

Example07_DesignThinking.class

Sample of Stanford University's approach to "Design Thinking" with phases: Empathize, Define, Ideate, Prototype, Test.

Inspired from SKRecipes

Documentation/Further reading

1. Semantic Kernel Official Docs
2. Semantic Kernel Github Repo
3. In Depth Tutorials in Python & C#
4. Sample apps (not in Java yet)