[SPARK-56864][INFRA][PYTHON] Consolidate python-ps-minimum image into python-minimum

[zhengruifeng]

What changes were proposed in this pull request?

This PR consolidates the python-ps-minimum Docker image and its CI workflow into the existing python-minimum image, eliminating a near-duplicate.

Specifically:

• Updates the label on dev/spark-test-image/python-minimum/Dockerfile to cover both PySpark and Pandas API on Spark.
• Deletes dev/spark-test-image/python-ps-minimum/Dockerfile.
• Deletes .github/workflows/build_python_ps_minimum.yml.
• Adds "pyspark-pandas": "true" to .github/workflows/build_python_minimum.yml so Pandas API on Spark minimum-deps coverage is preserved.
• Drops the python-ps-minimum entries from .github/workflows/build_infra_images_cache.yml (the paths trigger and the build/push step).
• Removes the build_python_ps_minimum.yml badge from README.md.

Why are the changes needed?

To save CI resources. The two Dockerfiles were nearly identical. The only functional differences were in BASIC_PIP_PKGS:

Package	python-minimum	python-ps-minimum
numpy	pinned ==1.22.4	unpinned
scikit-learn	included	omitted

Everything else (base image, apt packages, Python version, venv setup, CONNECT_PIP_PKGS) was the same. Maintaining both images doubles the image build/cache cost and runs a duplicate scheduled workflow without commensurate test value. Reusing python-minimum (which has the stricter pin and a superset of packages) for the Pandas API on Spark minimum-deps job keeps coverage while halving the image footprint and the associated CI runtime.

Does this PR introduce any user-facing change?

No. CI-only change.

How was this patch tested?

Existing CI. The merged build_python_minimum.yml now runs both pyspark and pyspark-pandas jobs against the python-minimum image.

Was this patch authored or co-authored using generative AI tooling?

Generated-by: Claude Code (model: claude-opus-4-7)

[dongjoon-hyun]

+1, LGTM.

cc @peter-toth