Issue #11 – Profiling, Benchmarking & Performance Tuning

[D4R102004]

Description: Conduct a comprehensive performance audit of the Pathos AI library. This involves using profiling tools to identify bottlenecks in the Node.expand logic and the heavy-lifting optimization loops. We will establish "Performance Baselines" to ensure future updates don't make the library slower.

Objectives:

Identify "Hot Spots" in the code (functions that take the most time).

Compare the memory overhead of the Node class with and without __slots__.

Verify that Numpy-based algorithms (GA, PSO) scale better than pure Python loops as problem size increases.

Create a benchmarks/ suite to track execution time across different Python versions.

Tasks:

[ ] Implement Profiling Scripts:

    Use cProfile and line_profiler to analyze a deep A* search and a large Genetic Algorithm run.

[ ] Memory Auditing:

    Use sys.getsizeof or tracemalloc to measure the memory footprint of 1,000,000 Node objects.

[ ] The "Numpy vs. Loop" Showdown:

    Create a test that solves the same problem using a pure Python loop vs. a vectorized Numpy operation to document the speedup factor (e.g., 10x, 50x).

[ ] Optimization Refactoring:

    Based on profiling results, refactor any slow sections in core.py (e.g., optimizing the f_score property).

[ ] Automated Benchmarking:

    Create scripts/run_benchmarks.py that outputs a markdown table of results for the README.

Acceptance Criteria:

[ ] A profiling report is generated showing that Node.expand is efficient.

[ ] Memory usage for 1 million nodes is documented (should be significantly lower thanks to __slots__).

[ ] A performance regression test is added to the CI/CD pipeline (optional but recommended).

Dependencies:

Issue #2 through Issue #10 (Requires all modules to be complete).