RenderContext and BaseRenderer Documentation

RenderContext

RenderContext is a Pydantic model that defines the context required for rendering strings. It includes the following attributes:

Attributes

• viewer (str):

  • Description: The viewer of the rendered string.
  • Default: "human"
  • Constraints: Must be one of 'human', 'environment', or 'agent_i' where i is a non-negative integer representing the index of the agent in the episode log.

• verbose (bool):

  • Description: Whether to render the verbose version of the string.
  • Default: False

• tags_to_render (list[str]):

  • Description: The special tags to render.
  • Default: []

Validators

• viewer_must_be_valid:
  • Ensures that the viewer field follows the required constraints.
  • Raises a ValueError if viewer is not one of the valid options.

Example

from pydantic import ValidationError
 
try:
    context = RenderContext(viewer="agent_1", verbose=True, tags_to_render=["tag1", "tag2"])
except ValidationError as e:
    print(e)

BaseRenderer

BaseRenderer is a base class that represents a simple string renderer.

Methods

• __call__(self, input_string: str, context: RenderContext) -> str:
  • Description: Renders the given input_string based on the provided context. Currently, it simply returns the input_string without any modification.
  • Parameters:
    • input_string (str): The string to be rendered.
    • context (RenderContext): The rendering context that includes viewer, verbosity, and tags.
  • Returns: The input string (str).

Example

context = RenderContext(viewer="human", verbose=False, tags_to_render=["highlight"])
renderer = BaseRenderer()
rendered_string = renderer("This is a test string.", context)
print(rendered_string)  # Output: This is a test string.