RAIL (Reliable AI Language) is a dialect of XML that allows users to
specify guardrails for large language models (LLMs).
A RAIL file contains a root element called
<rail version="x.y">
that contains the following elements as children:
1. <input strict=True/False>, which contains the input schema
2. <output strict=True/False>, which contains the output schema
3. <prompt>, which contains the prompt to be passed to the LLM
4. <instructions>, which contains the instructions to be passed to the LLM
input_schema: Optional[Schema]
output_schema
instance-attribute
instructions
instance-attribute
instructions: Optional[Instructions]
prompt
instance-attribute
version
class-attribute
instance-attribute
from_file
classmethod
from_file(file_path: str) -> Rail
from_string
classmethod
from_string(string: str) -> Rail
from_xml
classmethod
from_xml(xml: ET._Element)
from_pydantic
classmethod
from_pydantic(output_class: Type[BaseModel], prompt: Optional[str] = None, instructions: Optional[str] = None, reask_prompt: Optional[str] = None, reask_instructions: Optional[str] = None)
from_string_validators
classmethod
from_string_validators(validators: List[Validator], description: Optional[str] = None, prompt: Optional[str] = None, instructions: Optional[str] = None, reask_prompt: Optional[str] = None, reask_instructions: Optional[str] = None)