"""
# contacts

a nickel laterite logged down each hole as ferricrete (FERR), limonite (LIM), saprolite (SAP) and bedrock (BRK).
should you estimate the horizons apart? grade against distance to a contact separates a hard boundary, where grade
jumps, from a soft one, where grade changes gradually and samples on one side inform the other.
"""

# %% [hidden]
import sys
from pathlib import Path

HERE = Path(__file__).parent if "__file__" in globals() else Path.cwd()
sys.path.insert(0, str(HERE.parents[1]))

# %%
import boitata as bt
import matplotlib.pyplot as plt
from common import save

data = bt.datasets.nickel_laterite_profile()
assays = bt.merge_intervals(data["assays"], data["horizons"])
points = bt.Drillholes(data["collars"], data["surveys"], assays).samples()
print(f"{len(points)} assays of 1 m in {len(data['collars'])} holes")

# %% [markdown]
# `contact` measures, down each hole, the distance from every sample of the two horizons to the nearest sample of the
# other (negative inside) and bins them. `plot.contact` draws the mean per bin with the counts as light bars. the same
# call runs on composites or on any points with hole ids and a domain column.

# %%
pairs = [("LIM", "SAP"), ("SAP", "BRK")]
tables = {
    (grade, inside): bt.contact(
        points,
        grade,
        domain_column="HORIZON",
        holes="HOLE_ID",
        inside=inside,
        outside=outside,
        max_distance=8.0,
        bin=1.0,
    )
    for grade in ["NI_PCT", "CO_PCT"]
    for inside, outside in pairs
}
for (grade, inside), table in tables.items():
    mean = dict(zip(table["distance"], table["mean"], strict=True))
    print(f"{grade} {inside} contact: {mean[-0.5]:.2f} % at -0.5 m, {mean[0.5]:.2f} % at +0.5 m")

fig, axes = plt.subplots(2, 2, figsize=(10, 6.4))
for ax, ((grade, inside), table) in zip(axes.flat, tables.items(), strict=True):
    outside = dict(pairs)[inside]
    bt.plot.contact(table, labels=(inside, outside), ax=ax)
    ax.set(title=f"{grade.split('_')[0].title()}, {inside} over {outside}", xlabel="Distance to contact (m)")
    ax.set_ylabel(f"Mean {grade.split('_')[0].title()} (%)")
fig.tight_layout()
save(fig, "contacts")

# %% [markdown]
# Ni is soft at the LIM/SAP contact: it rises from 1.10 to 1.32 % across it, then keeps climbing for several meters into
# the saprolite, so Ni in either horizon near the contact can borrow samples from the other. at the base of the
# saprolite it falls from 2.21 to 0.29 % within a meter, a hard boundary: keep bedrock samples out of the saprolite
# estimate. Co shows the other pattern, stepping from 0.12 to 0.04 % at the LIM/SAP contact and staying flat on each
# side. the same horizons can be a hard boundary for one grade and a soft one for another.