Fix Python 3.12 and numpy 2.x compatibility

[SaeedTaghavi]

Problem

Closes #84
Installing pyspike fails on Python 3.11+ and numpy 2.x because:

• The bundled pre-generated .c files were compiled with an old
  version of Cython incompatible with Python 3.12's C API
• numpy 2.0 removed internal structs the old .c files depended on
• pkg_resources is no longer available in newer setuptools

Solution

• Add pyproject.toml declaring Cython and numpy as build dependencies
• Simplify setup.py to always compile from .pyx files using
  cythonize(), removing reliance on bundled .c files entirely
• Replace pkg_resources with importlib.metadata in __init__.py
• Fix invalid escape sequences in docstrings causing SyntaxWarnings
• Update .gitignore to exclude build artifacts

Testing

Tested successfully on Python 3.12.3 with numpy 2.4.4 on Ubuntu WSL2.