Module schema_evolution 

Autonomous schema evolution

Automatically detects schema changes from ingested events and manages schema evolution without manual intervention. Works alongside the existing SchemaRegistry to:

1. Infer schemas from event payloads on first ingestion of a new event type
2. Detect drift when new events have fields not in the registered schema
3. Auto-register backward-compatible schema versions (new optional fields)
4. Flag breaking changes that require manual review (type changes, removed required fields)

How it works

On each event ingestion, the SchemaEvolutionManager compares the event’s payload structure against the latest registered schema for that event type. If differences are detected:

• New optional fields → auto-registered as a new schema version
• Type changes or missing required fields → flagged as breaking, not auto-registered

Limitations

• Only analyzes top-level and one level of nested fields
• Type inference is basic: string, number, boolean, array, object, null
• Does not handle polymorphic types (field sometimes string, sometimes number)
• No automatic data migration — schemas describe shape, don’t transform data

Structs

EvolutionRecord

A record of a schema evolution event

FieldSchema

An inferred field schema

FieldTypeChange

A field type change (potential breaking change)

InferredSchema

Schema shape inferred from a payload

SchemaDiff

Diff between two schemas

SchemaEvolutionManager

Schema evolution manager — tracks inferred schemas and evolution history

SchemaEvolutionStats

Schema evolution statistics

Enums

EvolutionAction

Schema evolution event types

InferredType

Inferred JSON type for a field

Functions

compute_diff

Compute the diff between an existing schema and a new payload’s inferred schema

infer_schema

Infer a schema shape from a JSON payload

to_json_schema

Convert an inferred schema to a JSON Schema document