"""
Read an IFC building
====================

IFC (Industry Foundation Classes) is the dominant interchange format
for buildings and infrastructure (BIM). This example loads a real BIM
model, the buildingSMART "Single-family house" sample (CC-BY 4.0,
buildingSMART International), and confirms that every product becomes
one block in a :class:`pyvista.MultiBlock` with ``cad.guid``,
``cad.ifc_type``, ``cad.label``, and ``cad.material`` field data
attached.

Why this matters: a typical BIM workflow starts by filtering an IFC by
``IfcType`` (walls vs. slabs vs. mechanical) before any further
analysis. The ``.cad`` accessor's ``find`` is built for exactly that.
"""

# %%

import pandas as pd
import pyvista as pv

import pyvista_cad
from pyvista_cad.examples import downloads

# %%
# Read the real IFC building.

mb = pyvista_cad.read_ifc(downloads.ifc_building_path())
mb

# %%
# Real BIM exports carry stray markers: this file has a
# ``geo-reference`` point and a survey ``origin``, both placed far
# from the house. They have no reliable IFC class (generic
# ``IfcBuildingElementProxy``), so ``drop_spatial_outliers`` removes
# them by a robust MAD test on block position instead of by name.

mb = mb.cad.drop_spatial_outliers()

# %%
# Filter by IFC type using the ``.cad`` accessor: the canonical BIM
# pattern: pull every wall out of an arbitrary-depth building.

walls = mb.cad.find(ifc_type='IfcWall')
pd.DataFrame({'path': [str(p) for p, _ in walls]})

# %%
# Render with one color per element.

pl = pv.Plotter()
pl.add_mesh(mb, multi_colors=True, show_edges=True, edge_color='black')
pl.show()