Crashes with Pandas 2.2+ and NumPy 2.0+ due to strict type enforcement

[andrewendlinger]

Describe the bug

Running pyAMARES with recent versions of pandas (>= 2.2.0) and numpy (>= 2.0.0) causes multiple crashes. This happens because these newer versions removed legacy behaviors like implicit upcasting (causing LossySetitemError) and introduced strict typing for empty DataFrames (causing .str accessor AttributeErrors).

Expected behavior

The package should reliably process prior knowledge and fit data without crashing in modern Python environments.

Proposed Solution

1. Short-term: Pin pandas < 2.2.0 and numpy < 2.0.0 in the package requirements to immediately restore stability.
2. Long-term: Refactor the codebase to properly handle strict type casting and empty DataFrames to support modern dependency versions.