dataguy.ContextManager

class ContextManager(max_code_history=100)

Tracks a concise Python execution context for LLM prompts:

• Non-private module imports

• Relevant variables (DataFrames, arrays, primitives, lists, dicts)

• Recent executed code history

__init__(max_code_history=100)

Methods

__init__([max_code_history])
add_code(code)	Appends executed code lines to history (up to max length).
get_context_summary()	Returns a markdown-like summary:
update_from_globals(globals_dict)	Rebuilds the context from the given globals dictionary, keeping only non-private modules and meaningful variables.