fix(describegpt): keep jsonschema examples valid against property type

[jqnatividad]

Problem

describegpt --format jsonschema produced a schema that failed JSON Schema Draft 2020-12 validation. The schema document itself passed the meta-schema, but its examples arrays contained values that did not validate against their own property's type:

[X Coordinate (State Plane)] examples value 'Other…'  -> not of type 'integer', 'null'
[Y Coordinate (State Plane)] examples value '(NULL)…' -> not of type 'integer', 'null'
[Latitude]                   examples value 'Other…'  -> not of type 'number', 'null'
[Longitude]                  examples value '(NULL)…' -> not of type 'number', 'null'
... (8 total on a NYC 311 sample)

Root cause

For numeric columns, the frequency aggregation-bucket rows — the "Other" long-tail bucket and the "(NULL)" bucket, rendered as Other… / (NULL)… — were landing in examples. coerce_value cannot parse those as a number, so it falls back to a JSON string, and that string leaked into an integer/number-typed property's examples array.

The spec recommends that each examples member validate against the schema, and validators that enforce this rejected the output.

Fix

In src/cmd/describegpt/formatters.rs:

• Add value_matches_json_type() — checks a coerced value against the property's declared JSON Schema scalar type.
• build_property_schema filters examples through it, so a value is emitted only when it matches the property's type. Numeric/boolean properties drop the Other…/(NULL)… sentinels; string properties keep all their (type-valid) examples. The existing "skip examples if empty" guard handles the case where filtering removes everything.

enum/const were already safe — generate_code_based_dictionary only populates enumeration for columns with no bucket rows.

Testing

• New unit tests: jsonschema_drops_non_numeric_examples_from_numeric_property, jsonschema_keeps_examples_for_string_property — pass.
• All 70 describegpt integration tests pass.
• cargo clippy clean for the changed file; binary builds clean.

🤖 Generated with Claude Code

[codacy-production]

Up to standards ✅

🟢 Issues 0 issues

Results:
0 new issues

View in Codacy

🟢 Metrics 4 complexity

Metric	Results
Complexity	4

View in Codacy

_{NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer
TIP This summary will be updated as you push new changes.}