Klongpy

DuckDb to 0.9.2 (still limited to Python 3.11, 0.10.0 will have support for 3.12)
Min version of Python to 3.9 (because NumPy 1.26.4 is 3.9 to 3.12)

A few upgrades since last release: Pandas 2.2.2 and DuckDB 1.0.

Some important bug fixes and simplified the internals for groupby.

Most notably, KlongPy can now call into python code that uses multiprocessing.
 
[fix async for remote connections · briangu/klongpy8ef2186 (github.com)](https://github.com/briangu/klongpy/commit/8ef2186a420d4aa4f643692913f49cb6e4b9f306)
[fix multiprocessing for dynamic modules · briangu/klongpy794a5d1 (github.com)](https://github.com/briangu/klongpy/commit/794a5d1a276def4ac849e4fe6a832a2ba73b695c)
[simplify the groupby · briangu/klongpy4c73131 (github.com)](https://github.com/briangu/klongpy/commit/4c73131c4ce1186b3388959feecf4164131adb06)

fixed negate which was creating a NumPy object array when negating an array - should be the base type in the array.
[vectorize black-scholes model · briangu/klongpy989dcc0 (github.com)](https://github.com/briangu/klongpy/commit/989dcc016282db83fc8e53490beb4cded639b2e2)

fixed format2 and form broadcasting bugs.
some minor dyad perf optimizations
fixed kgpy printing of strings, particularly annoying when embedded in arrays and you couldn't tell if an element was a string or integer.

added some initial examples for docker / k8s deploy