Pseudocode Extractor

Real code has type annotations, error handling, logging, defensive checks, and language-specific idioms. Pseudocode has the algorithm. Extraction is deciding what's essential vs. what's machinery.

Compare → verified-pseudocode-extractor: that one preserves pre/post/invariants from verified code. This one works on any code and focuses on algorithmic clarity.

Keep vs. strip

Source code element	In pseudocode
Core control flow (if/while/for)	KEEP — this is the algorithm
Data transformations	KEEP — the what
Variable types	STRIP usually; KEEP if the type is the point (e.g., "priority queue")
Error handling (try/catch)	STRIP unless the error path is algorithmic
Logging, metrics, tracing	STRIP
Defensive null checks	STRIP (or note as precondition)
Resource management (open/close)	STRIP (becomes "with resource R: ...")
Memory management	STRIP entirely
Language idioms (`__enter__`, RAII)	STRIP — translate to intent