pstats

pstats is a Python standard library module for reading and analyzing profiling results produced by Python's profilers such as cProfile and profile. It provides tools to inspect, compare, and present statistics about function calls, execution time, and call relationships. The main entry point is the Stats class, which can load data from a file created by a profiler and expose methods to reorder and display the data.

Internally, pstats stores statistics as a mapping from function identifiers to a tuple of metrics. Each entry is keyed by a triple (filename, line number, function name) and the values typically include the number of calls, primitive calls, total time spent in the function, and cumulative time spent including subcalls, along with a callers map. The exact interpretation of the fields is documented in the module; in practice users sort and filter by time or by call counts.

Common operations include stripping directory paths for readability, sorting by cumulative time or self time, printing a compact top-N report, and displaying callers or callees to understand call relationships. pstats also supports combining data from multiple profiling runs, enabling cross-run comparisons. It is widely used for analyzing performance bottlenecks in Python programs.

See also cProfile, profile, and the pstats module in the Python standard library.