docs: add a Vectors how-to page #118

The Antora component has no page about vectors. Vector search is one of the main uses of the bindings, and what is specific to Python is not written down in this repository: the two ways to create an index and their different defaults (quantization is NONE in CREATE INDEX ... LSM_VECTOR and INT8 in create_vector_index; addHierarchy is false and True), which of them builds the graph at creation, how NumPy arrays and lists cross as vectors, find_nearest with allowed_rids, get_stats(), warm_up(), and what mutations_before_rebuild does (a floor, not an override).

The old guide at docs.humem.ai/arcadedb has this content, with the old import name and with measured tables that were taken on 26.10.x wheels. This issue adds a "Vectors" how-to page that ports the Python-specific parts, states behavior and no numbers, and is SQL and openCypher first. Every Python block of the page runs in a new test (tests/test_docs_vectors_page.py) on both packages.

Part of #19 (the vector docs of humemai/arcadedb-embedded-python#499).