Package forecasting tools with CLI entry points

[amarakramali]

What changed

• add standards-based pyproject.toml build and project metadata
• install six aep-* console commands for the complete forecasting workflow
• document editable and regular installation paths
• build wheel and source distributions in CI and smoke-test every installed command

Why

The repository previously required running modules from a source checkout and
could not be installed as a Python project. This made setup less conventional
and left packaging regressions untested.

User impact

Users can now install the project with python -m pip install . (or editable
mode) and run the workflow through stable console commands such as aep-demo
and aep-forecast.

Validation

• python -m pip install --editable ".[dev]"
• python -m ruff check src tests streamlit_app.py
• python -m pytest -q (65 passed)
• python -m compileall -q src tests streamlit_app.py
• python -m build (wheel and source distribution)
• all six installed console commands smoke-tested with --help