Repository navigation
Early Access 2
Pre-releaseThis release introduces the class hierarchy of Generator. A Generator is a producer of action or creator of artifact from the output of a Movement. In this release, 3 examples of a Generator is in this release:
PythonCodeRunner: extracts from a Movement output code section demarcated by ```python and executes the code within, capturing the stdout output into the output of the current Movement.
PDFGenerator: takes its input from the Movement prompt (prompt_str attribute), and creates a best guess PDF document from it.
PromptContextGenerator: a special generator that collects input from the user to build a context for the current prompt, giving it intent and accuracy.
Together with this release is a tests/ask_with_context.py code that exercises all 3 Generator types to demo their capabilities.