docs: add Python deploy guide and runnable example for Datum compute

[scotwells]

Summary

Adds a step-by-step guide for deploying a Python web service on Datum compute, plus a ready-to-deploy example so a developer can go from clone to a live Instance without copy-pasting from prose.

• Guide now lives at docs/guides/deploy-a-python-app.md (moved out of docs/compute/guides/). It walks through packaging a Python HTTP service as a Unikraft unikernel, publishing it to the metro registry with kraft, and deploying it with datumctl compute.
• Runnable example at examples/hello-python/ materializes everything the guide describes:
  • server.py — a stdlib-only HTTP service (/ and /healthz)
  • Dockerfile — builds the trimmed shared-library closure so the rootfs fits the unikernel's in-RAM filesystem
  • Kraftfile — Unikraft Cloud packaging spec
  • workload.yaml — the Datum compute Workload manifest
  • README.md — one-line description, link back to the guide, and the key build + deploy commands

Why this matters

Python is the most-requested first language for new users, and getting a dynamically linked interpreter to boot as a unikernel is the hardest part of the on-ramp. This pairs a narrative guide with a folder you can build and deploy as-is, turning "Hello from Datum" into a five-minute experience and giving us a verified reference for the trimmed-closure pattern.

🤖 Generated with Claude Code