Save Pareto front history per iteration with configurable export

[Copilot]

Implements per-iteration Pareto front archiving. Previously, self.history['pareto_front'] was overwritten each step; only the final state was preserved.

Changes

• Per-iteration Pareto storage: Keys changed from pareto_front to pareto_front_{iteration}, preserving full optimization trajectory
• CSV export per iteration: Added history/pareto_iteration{i}.csv alongside existing particle exports
• Control flag: FileManager.saving_history_enabled gates all per-iteration I/O (particles + Pareto)
• Data structure fix: Replaced failing numpy structured arrays with list-of-dicts, resolving ValueError with mixed parameter types (int/float/bool)
• Zarr export: Added Pickle codec support for object arrays

Usage

pso = patatune.MOPSO(...)
pso.optimize(num_iterations=100)

# Access evolution of Pareto front
for i in range(100):
    pareto = pso.history[f'pareto_front_{i}']
    # Each entry: {'position': array, 'fitness': array}

# Disable if storage is a concern
patatune.FileManager.saving_history_enabled = False

Fixes the structured array bug that prevented history saving from working at all.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Save history of archived solutions</issue_title>
<issue_description>Save the pareto of each iteration and allow to export it (with a flag to enable the feature)</issue_description>

Comments on the Issue (you are @copilot in this section)

@rsreds #43 adds the history of each step in the output. Needs to add the history of the pareto too

• Fixes Save history of archived solutions #45

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.