DRY GFQL call validation safelist entries

[lmeyerov]

Summary

• audited the PR feat(gfql): strengthen axis diagnostics #1592 axis/ring extraction and confirmed call/validation.py no longer owns duplicate axis/ring payload validation bodies; it adapts canonical graphistry.validate.validate_axis payload errors into GFQL call validator results
• DRYed repeated private safelist entry definitions with compact helper ownership in graphistry.compute.gfql.call.support
• DRYed repeated encoding, metadata, filter-by-dict, degree, and schema-effect entry definitions while preserving runtime metadata shape
• added a Development changelog entry

Refs #1058.

Axis/ring classification

Validator surface	Relationship to validate_axis.py	Decision
encode_axis.rows	delegates to documented_axis_rows_payload_error() after preserving exact GFQL list/dict message precheck	intentionally split adapter, now inlined in safelist
ring_continuous_layout.axis	delegates to ring_continuous_axis_payload_error() with None allowed	intentionally split adapter, now inlined in safelist
ring_categorical_layout.axis	delegates to ring_categorical_axis_payload_error() with None allowed	intentionally split adapter, now inlined in safelist

Sample axis migration:

'rows': _validate_payload_error(
    'rows',
    documented_axis_rows_payload_error,
    list_of_dicts_message=True,
)

LOC buckets

• Production target file: graphistry/compute/gfql/call/validation.py 1243 -> 1109 LOC (-134)
• Production total: +320/-403 (net -83) across validation.py and support.py
• Tests: +116/-1 LOC across focused helper-contract and schema-effect pressure tests
• Docs/changelog: +1 LOC in CHANGELOG.md

Compatibility / compiler-plan surface

• Compiler-plan surface touched: no
• No logical/physical plan routes, dispatch semantics, executor behavior, source-span metadata, or public compatibility shims changed
• Public viz payload compatibility preserved; unknown axis/ring extension subtypes continue to be handled by canonical validate_axis.py behavior

Source-path fidelity

• Preserved the feat(gfql): add encode call parity #1584/feat(gfql): strengthen axis diagnostics #1592 anchored diagnostics by keeping the GFQL call adapter behavior and routing payload details through canonical helpers
• Focused tests cover exact rows[0], axis[0], and axis[0].bounds['min'] detail paths before remote post

Validation

• old-vs-new safelist behavioral comparator against origin/master: function keys equal (51), metadata failures 0, sampled validator diffs 0, sampled schema-effect diffs 0
• python3 -m pytest -q graphistry/tests/compute/test_gfql_call_validation.py graphistry/tests/compute/test_call_schema_validation.py graphistry/tests/compute/test_call_operations.py graphistry/tests/test_gfql_remote_metadata.py graphistry/tests/test_validate_settings.py -> 129 passed, 2 skipped
• python3 -m pytest -q graphistry/tests/compute/test_astcall_chains.py graphistry/tests/compute/gfql/test_layout_chain_predicate.py graphistry/tests/compute/gfql/layouts/test_call_layout_radial.py graphistry/tests/layout/ring/test_continuous.py graphistry/tests/layout/ring/test_ring_categorical.py graphistry/tests/test_dataset_id_invalidation.py -> 77 passed, 3 skipped
• broad non-GPU GFQL core CI surface, serial locally because pytest-xdist is unavailable: 3110 passed, 20 skipped, 230 deselected, 15 xfailed
• full broad GFQL core surface without exclusions in this shell fails on cuDF/UMAP tests with cudaErrorNoDevice because cuDF/cupy are installed but no CUDA device is visible; PR CI remains the authoritative full gate
• ./bin/ruff.sh graphistry/compute/gfql/call/validation.py graphistry/compute/gfql/call/support.py graphistry/tests/compute/test_gfql_call_validation.py graphistry/tests/compute/test_call_schema_validation.py graphistry/tests/test_gfql_remote_metadata.py graphistry/tests/test_validate_settings.py -> green
• ./bin/typecheck.sh graphistry/compute/gfql/call/validation.py graphistry/compute/gfql/call/support.py -> green
• python3 -m py_compile graphistry/compute/gfql/call/validation.py graphistry/compute/gfql/call/support.py graphistry/tests/compute/test_gfql_call_validation.py graphistry/tests/compute/test_call_schema_validation.py -> green
• ./bin/test-minimal-lite.sh -> local environment failure only: cuDF/cupy are installed in this shell but no CUDA device is visible, causing cudaErrorNoDevice failures in GPU/UMAP/text tests. Non-GPU targeted validation is green; PR CI is the authoritative gate.

Review

• Review skill wave 1 converged locally. The initial PR was too high-addition; follow-up slim pass reduced production net from -24 to -83 and target-file shrink from -97 to -134. Pressure pass added helper-contract/schema-effect tests for the mechanical deletion surfaces. No unresolved blocker/important findings.

DGX

• DGX RAPIDS 25.02/26.02 not run for this diff: no DataFrame/cuDF/GPU runtime path changed. The local broad sentinel CUDA failures are environmental and unrelated to the validation-only code path.