Build numpy without a fixed SIMD baseline in the Docker image

[Robbie1977]

v1.22.40 fails to start on the VFB Rancher/k8s hosts:
RuntimeError: NumPy was built with baseline optimizations: (X86_V2) but your machine doesn't support: (X86_V2)

No VFBquery code changed. vfb_connect 2.4.1/2.4.2 (21 Aug) dropped the numpy<2.0.0 pin it carried in 2.3.x, so the image now resolves numpy 2.4.x. NumPy 2.4 raised its default x86 baseline to x86-64-v2 (SSE4.2/POPCNT/SSSE3), and every VFB host (parsley, buttermilk, vfbk8s10) is a QEMU VM on the generic qemu64 CPU model, which exposes none of those flags.

This builds numpy from source in the Dockerfile with -Dcpu-baseline=none before the rest of the dependencies are installed. The result has no fixed baseline but keeps X86_V2/V3/V4 as runtime-dispatched targets, so there is no performance cost on hosts that have them. Verified: numpy 2.4.6 built this way imports cleanly and survives the pip install -e . that follows (pandas 3.0.5, vfb_connect 2.4.2). Adds roughly 3 minutes per Docker build. NUMPY_SPEC build-arg allows pinning a version if ever needed.

The underlying fix is the hypervisor CPU type for the VM fleet; once that's done this block can go.

[github-actions]

⚠️ 167 test(s) skipped — VFB backend was unreachable

The full suite ran, but 167 test(s) were skipped because the VFB backend (Neo4j / SOLR / Owlery) did not answer during this run.

These are not failures and not a problem with this branch — but those queries went unverified, so a green check here is an incomplete run.

To re-run: open this workflow run and click Re-run all jobs once the backend is healthy (re-running the “Run completeness” check itself does nothing — it has no job behind it).

523 passed, 167 skipped, 338 warnings in 704.89s (0:11:44)

_{Posted automatically. This comment is removed once a run completes with zero skips.}