Suggest dbt tests in print_profile_schema based on profile results

[stumelius]

Overview

print_profile_schema already generates a schema.yml skeleton with profile metrics stored in meta. It could go further and emit dbt test suggestions based on what the profile reveals — turning a profile run into a data contract bootstrap tool.

Motivation

Today print_profile_schema outputs:

columns:
  - name: id
    meta:
      is_unique: true
      not_null_proportion: 1.0
      distinct_count: 1000

A user then has to manually translate this into tests. The mapping is mechanical and well-defined, so the macro can do it automatically.

Suggested rules

Profile result	Suggested test
is_unique = true	- unique
not_null_proportion = 1.0	- not_null
distinct_count <= N (configurable, e.g. 20)	- accepted_values: values: [...] populated from the actual values (requires value frequency feature)
min and max present	- dbt_utils.expression_is_true: expression: "col between {min} and {max}"

API design

New optional parameter on print_profile_schema:

dbt_profiler.print_profile_schema(
    relation_name="my_model",
    suggest_tests=true,          -- off by default to avoid breaking existing usage
    max_values_for_accepted=20   -- threshold for emitting accepted_values test
)

Notes

• The accepted_values suggestion depends on the value frequency feature (Question: add TOP x column values and distribution. #65 / a future issue) to know the actual distinct values
• Suggestions should be clearly marked as suggestions (e.g. a comment in the YAML) so users know to review them before committing
• This makes dbt-profiler a lightweight contract bootstrap tool: profile once → get a tested schema.yml ready to review