Add tolerant convert.parse for JSON-like cell values

[ebhills]

Problem

Cell values often contain JSON, Python literals, YAML-like mappings/lists, or accidentally quoted structures. Recipes need an explicit general-purpose parser for these values while preserving the existing JSON/YAML conversion wrangles.

Scope

• Add convert.parse with any, dictionary, list, and scalar expected categories, per-column defaults, and missing-value handling.
• Return JSON-compatible values, including normalized NumPy values and independent mutable defaults.
• Reject YAML anchors, aliases, explicit tags, and unsupported output/default values; preserve YAML-only numeric forms such as 00123 and 12:34 as text.
• Keep from_json, from_yaml, to_json, and to_yaml behavior unchanged. Migrating other wrangles or adding recursive parsing is outside this issue.

Acceptance criteria

• PR Add convert.parse object parsing wrangle #1187 is reviewed and merged with required CI passing.
• Focused tests cover supported representations, quoted input, defaults, expected categories, YAML safety, numeric identifiers, and compatibility.
• Generated recipe schema validates the scalar and array forms of expected.

Implementation and validation are tracked in #1187.