Make Azure AI Foundry the primary AI provider across Parts 1-6 (#492)

[jongalloway]

Summary

Closes #492.

GitHub Models is fully retired on July 30, 2026 (brownouts July 16 & 23), so it can no longer be the default provider for the July 27 workshop. This PR makes Azure AI Foundry (Azure OpenAI) the primary provider across Parts 1–6 and demotes GitHub Models to a clearly-labeled legacy fallback.

Because the app already uses AddAzureOpenAIClient (and the IChatClient / IEmbeddingGenerator abstractions), this is a documentation + connection-string change — no application code rewrite.

Changes

• Part 2 — Azure AI Foundry is now the primary path:
  • Visual Studio flow selects "Azure OpenAI"; CLI uses --provider azureopenai
  • New "Set the Azure AI Foundry connection string" section (endpoint/key + gpt-4o-mini / text-embedding-3-small), linking to the Part 4 deployment steps
  • GitHub Models moved into a collapsible legacy fallback block with a retirement [!WARNING]
  • Updated "What You've Learned" and the connection-troubleshooting entry to Azure
• Part 3 — exploration code comment shows the Azure endpoint as primary, GitHub Models as legacy
• Root README — architecture diagram node, "Development → Production" flow, intro, lab-overview bullet, and prerequisites all lead with Azure AI Foundry (GitHub Models labeled legacy/retiring)
• Part 5 — provider mentions reordered (Azure first)
• Sample AppHost.cs (Parts 2, 4, 5, 6) — user-secrets comment shows the Azure connection string as primary with a GitHub Models legacy note

Verification

• ✅ Part 2 solution builds (0 warnings, 0 errors)
• ✅ No errors in edited .cs files
• ✅ Parts 5–6 functionally unchanged (comment-only code edits)

Scope notes

Left for their own milestone issues to keep this PR focused:

• Pre-workshop Azure setup checklist → Add mandatory pre-workshop Azure AI Foundry setup checklist #493
• Retirement banners in Part 1 → Add GitHub Models retirement banners and legacy labeling #494
• Reframing Part 4 into alternative/local providers → Reframe Part 4 as alternative & local providers (Foundry Local + Ollama) #495

Follow-up: Part 2 currently links forward to Part 4 for resource/model deployment. Once #493 (pre-work checklist) lands, that reference should point at the pre-work section instead.