fix(security): harden artifact serialization and loading

[XianBW (XianBW)]

Summary

• replace HIST stock-index object NPY metadata with validated JSON
• replace pandas/numpy module-prefix trust with an explicit restricted-pickle global allowlist
• make unrestricted MLflow artifact loading opt-in through trusted=True with a warning
• constrain HighFreqProvider artifact paths to an explicit root and stop mutating input configs

Security impact

This blocks the reported HIST object-deserialization path and the pandas.read_pickle/namespace bypass in RestrictedUnpickler. Arbitrary Python pickle artifacts can still be loaded for compatibility, but only through an explicit trusted path.

Compatibility

• HIST stock_index must now be JSON; the bundled mapping is migrated without changing its 735 entries.
• callers loading arbitrary model objects from MLflow must explicitly pass trusted=True. DataFrame, Series, ndarray, and supported Qlib data artifacts continue through restricted loading.
• HighFreqProvider paths default to the current working directory as artifact_root; callers using absolute paths outside it must set the intended root.

Validation

• py_compile for changed Python files
• git diff --check
• verified migrated HIST mapping exactly matches the previous NPY mapping
• verified ndarray, Series, and DataFrame restricted-pickle round trips
• regression tests reject pandas.read_pickle, numpy.load, os.system, subprocess.Popen, and reduce payloads

Full project pytest is delegated to CI because the local base environment does not contain Qlib's complete optional dependency set.

[XianBW (XianBW)]

Local verification update:

• 23 security regression tests passed
• handler serialization plus full train/backtest pipeline: 4 passed
• repository-wide Black, Flake8, and Pylint passed
• HIST JSON mapping verified identical to all 735 entries in the previous object NPY
• added explicit trusted=True to the pipeline test that intentionally restores an arbitrary model pickle

The branch also contains minimal CI compatibility fixes for current Node/commitlint, Plotly, MLflow, and Windows/Python 3.8 OSQP behavior.

[XianBW (XianBW)]

Final verification: all 76 GitHub checks pass on commit 9d8d63df/latest branch state, including title, CLA, pip compatibility, source builds, notebook execution, full pytest, and slow pytest matrices.