How azure-samples/semantic-kernel-advanced-usage works
Project type and purpose
• This is not a single application but a collection of self-contained “scenario” templates demonstrating advanced features of the Microsoft Semantic Kernel framework.
• Each folder under templates/ is an independent example (e.g., Dapr orchestration, authentication context, Copilot Studio skill, natural-language-to-SQL, Opentelemetry, etc.).
• The README lists the scenarios and explains their goals: showcasing orchestration patterns, hosting (Dapr), state machines, telemetry, Copilot integration, MS Graph APIs, and custom multi-agent strategies.
File structure
• Root level: repository metadata (LICENSE, CONTRIBUTING, CI, linter configs).
• templates/: contains directories for each scenario, each with its own source (src), infra (Bicep templates), Dockerfiles, requirements, and README.
• Patterns per scenario:
• An infra/ folder with Azure Bicep files provisioning resources (Cosmos DB, ACR, App Service, Actor hosting, OpenAI keys, UAMI).
• A src/ folder with Python modules organized by feature (agents, chat, API, process steps, utilities).
• Dockerfiles and .env.sample for local/docker deployments.
• Some scenarios include external clients (Chainlit, Flask/Starlette server, Copilot/Teams Bot, front-end HTML, Jupyter notebooks).
README for additional insights
• High-level overview of SK capabilities demonstrated.