Information Extraction Schema

ie-schema provides a flexible schema specification and parser for information extraction tasks.

Contents

• Usage Guide
  • Load a schema from JSON text
  • Load from a JSON Schema string
  • Load from dataclass or Pydantic model
  • Load a schema from disk
  • Iterate planned tasks
  • Render prompt debug string
• Python API Reference
  • Classes
  • Module Contents
• Rust Crate Reference
  • Where to go deeper

Installation

uv add ie_schema

Quick start

from dataclasses import dataclass
from ie_schema import IESchema

@dataclass
class BusinessRecord:
    business_name: str
    address_line1: str

# Native JSON schema, or a dataclass / Pydantic v2 model class/instance (install `pydantic`):
schema = IESchema.loads(BusinessRecord)
# schema = IESchema.loads('{"json_structures": ... }')

print(schema.prompt())

Install pydantic (e.g. uv add 'ie_schema[model]') to use IESchema.loads with dataclass/Pydantic model classes or instances.