forgeo-gmlib 0.6.2__cp312-cp312-win_amd64.whl

forgeo/gmlib/GeologicalModel3D.py

@@ -0,0 +1,758 @@
+#
+# This file is part of gmlib. It is free software.
+# You can redistribute it and/or modify it under the terms of the GNU Affero General Public License version 3.
+#
+
+from collections import namedtuple
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Any
+
+import numpy as np
+import yaml
+
+from forgeo.gmlib import geomodeller_project
+from forgeo.gmlib import pypotential3D as pypotential
+from forgeo.gmlib.common import CovarianceData
+from forgeo.gmlib.geomodeller_data import GradientData
+from forgeo.gmlib.pypotential3D import (
+    ConstantElevationSurface,
+    ElevationRaster,
+    ImplicitTopography,
+    Polyline,
+    VerticalSection,
+)
+from forgeo.gmlib.topography_reader import ImplicitDTM, ImplicitHorizontalPlane
+from forgeo.gmlib.utils.tools import BBox3
+
+
+@dataclass
+class Intersection:
+    point: Any
+    field: Any
+    value: float = 0.0
+    fault: str = None
+
+
+PileInfo = namedtuple("PileInfo", ["name", "color", "relation"])
+Box = namedtuple("Box", ["xmin", "ymin", "zmin", "xmax", "ymax", "zmax"])
+CRSInfo = namedtuple("CRSInfo", ["geomodeller", "qgis"])
+SeriesInfo = namedtuple(
+    "SeriesInfo", ["field", "drifts", "active_faults", "interfaces"]
+)
+
+scalardt = pypotential.scalar_type()
+
+
+# FIMXE: to be put elsewhere
+def image_ratio(section, nu):
+    assert section.umax > section.umin
+    nv = int(((section.vmax - section.vmin) / (section.umax - section.umin)) * nu)
+    assert nv > 0, (
+        f"z ratio is to small: {(section.vmax - section.vmin) / (section.umax - section.umin)}"
+    )
+    return nu, nv
+
+
+def covariance_data(potdata):
+    covmodel = potdata.covariance_model
+    return CovarianceData(
+        covmodel.gradient_variance,
+        covmodel.range,
+        covmodel.gradient_nugget,
+        covmodel.potential_nugget,
+    )
+
+
+def gradient_data(potdata):
+    graddata = potdata.gradients
+    return pypotential.gradient_data(graddata.locations, graddata.values)
+
+
+def interface_data(potdata):
+    return pypotential.interface_data(potdata.interfaces)
+
+
+def drift_basis(potdata):
+    drift_order = potdata.covariance_model.drift_order
+    return pypotential.drift_basis(drift_order)
+
+
+# FIXME this should be set elsewhere
+# Version with np.array here is suboptimal
+def point_between(p1, p2, field, v, precision=0.01):
+    squared_precision = precision**2
+    #    Point = pypotential.Point
+    #    p1 = pypotential.Point(p1)
+    #    p2 = pypotential.Point(p2)
+    p1 = np.asarray(p1, dtype=scalardt)
+    p2 = np.asarray(p2, dtype=scalardt)
+    assert len(p1.shape) == 1
+    assert len(p2.shape) == 1
+    # print('-> point between', p1, p2)
+    v1, v2 = field(p1), field(p2)
+    if v1 == v:
+        return p1
+    if v2 == v:
+        return p2
+    # print('field values:', v1, v2, 'looking for', v)
+    # FIXME: minimum can't be found (might scan between potentials)
+    if (v - v1) * (v - v2) > 0:
+        return None
+    #    previous = Point(p1) # copy is mandatory here
+    previous = np.copy(p1)  # copy is mandatory here
+    while True:
+        # (v - v1)/(v2 - v1)<1 so p is in [p1, p2] and we have convergence
+        # print('-> TEST:', p1)
+        # print('-> TEST:', v1, v2, p2- p1)
+        # if v1==v:
+        # return p1
+        # if v2==v:
+        # return p2
+        # if v1==v2==v:
+        # return None
+        p = p1 + float((v - v1) / (v2 - v1)) * (p2 - p1)
+        #      squared_length = (p - previous).squared_length
+        squared_length = np.sum((p - previous) ** 2)
+        if squared_length < squared_precision:
+            # print('-> FOUND', p)
+            return p
+        vp = field(p)
+        # print('field value:', vp)
+        #      previous = Point(p) # copy is mandatory here
+        previous = np.copy(p)  # copy is mandatory here
+        if (v - vp) * (v - v2) <= 0:
+            # Look for between vp and v2
+            v1 = vp
+            p1 = p
+        elif (v - vp) * (v - v1) <= 0:
+            # Look for between v1 and vp
+            v2 = vp
+            p2 = p
+        else:
+            raise AssertionError()
+            return None
+    return None
+
+
+def distance(p1, p2):
+    x = p1[0] - p2[0]
+    y = p1[1] - p2[1]
+    z = p1[2] - p2[2]
+    return np.sqrt(x**2 + y**2 + z**2)
+
+
+def extract_data_from_legacy_geomodeller_file(
+    filename, scalardt, skip_topography=False
+):
+    return geomodeller_project.extract_project_data(
+        filename, scalardt, skip_topography=skip_topography
+    )
+
+
+def make_potential(cache, name, potdata, drifts=None):
+    params = (
+        covariance_data(potdata),
+        gradient_data(potdata),
+        interface_data(potdata),
+        drifts or drift_basis(potdata),
+    )
+    cached = cache[name]
+    if cached is None:
+        field = pypotential.potential_field(*params)
+        cache[name] = field.data().beta()
+    else:
+        field = pypotential.potential_field(*params, cached)
+    return field
+
+
+def fault_potential(cache, name, data):
+    potdata = data.potential_data
+    assert len(potdata.interfaces) == 1, "inconsistent fault potential field"
+    return make_potential(cache, name, potdata)
+
+
+def finite_fault(cache, name, data):
+    assert not data.infinite
+    potdata = data.potential_data
+    assert len(potdata.interfaces) == 1, "inconsistent fault potential field"
+    points = np.reshape(potdata.interfaces[0], (-1, 3))
+    center = None
+    if data.center_type == "mean_center":
+        center = np.mean(points, axis=0)
+    elif data.center_type == "databox_center":
+        center = np.array(
+            [0.5 * (points[:, axis].min() + points[:, axis].max()) for axis in range(3)]
+        )
+    else:
+        assert type(data.center_type) is tuple, "unknown fault center type" + str(
+            data.center_type
+        )
+        center = np.array(data.center_type)
+    center = np.asarray(center, dtype=scalardt)
+    assert center is not None
+    field = fault_potential(cache, name, data)
+    # The following is as in GeoModeller: might be improved?
+    # it is based on the idea that the fault geometry is close
+    # to a (planar!) disk
+    g = pypotential.gradient(field)(center)
+    g.shape = (3,)
+    assert np.linalg.norm(g) > 0
+    g /= np.linalg.norm(g)
+    u = v = None
+    if g[0] == g[1] == 0:  # tangent plane is horizontal
+        u = np.array([1, 0, 0], dtype=scalardt)  # somewhat arbitrary
+    else:
+        vertical = np.array([0, 0, 1], dtype=scalardt)
+        u = np.cross(vertical, g)
+        u /= np.linalg.norm(u)
+    v = np.cross(u, g)  # the basis is not direct!
+    v /= np.linalg.norm(v)
+    g *= data.influence_radius
+    u *= data.lateral_extent
+    v *= data.vertical_extent
+    ellipsoid = pypotential.Ellipsoid(
+        pypotential.Point(*center),
+        (pypotential.Vector(*g), pypotential.Vector(*u), pypotential.Vector(*v)),
+    )
+    return field, ellipsoid
+
+
+def recast(data):
+    box, pile, faults_data, topography, formations = data
+    box_center = np.array(
+        [0.5 * (box[s + "min"] + box[s + "max"]) for s in ("X", "Y", "Z")],
+        dtype=scalardt,
+    )
+    L = max([box[s + "max"] - box[s + "min"] for s in ("X", "Y", "Z")])
+    assert L > 0
+
+    def recast_axis(i, x):
+        return (x - box_center[i]) / L
+
+    def recast_z(z):
+        return recast_axis(2, z)
+
+    def recast_P(P):
+        return (np.asarray(P, dtype=scalardt) - box_center) / L
+
+    if type(topography) is ImplicitHorizontalPlane:
+        topography = ImplicitHorizontalPlane(recast_z(topography.z))
+    else:
+        assert type(topography) is ImplicitDTM
+        topography = ImplicitDTM(
+            recast_P(topography.origin), topography.steps / L, recast_z(zmap)
+        )
+    new_box = {}
+    for i, s in enumerate(("X", "Y", "Z")):
+        new_box[s + "min"] = recast_axis(i, box[s + "min"])
+        new_box[s + "max"] = recast_axis(i, box[s + "max"])
+
+    def rescale_potential_data(handle):
+        if handle.potential_data is not None:
+            potdata = handle.potential_data
+            potdata.interfaces = [
+                recast_P(interface) for interface in potdata.interfaces
+            ]
+            gradients = potdata.gradients
+            gradients = GradientData(
+                recast_P(gradients.locations),
+                gradients.values,  # * L, # FIXME: scale effect *L
+            )
+            potdata.covariance_model.range /= L
+
+    for serie in pile.all_series:
+        if serie.potential_data is not None:
+            rescale_potential_data(serie)
+    for _name, fault in faults_data.items():
+        rescale_potential_data(fault)
+    return box, pile, faults_data, topography, formations
+
+
+def compute_fault_stops_on(model):
+    relations = {}
+    for fault, data in model.faults_data.items():
+        assert len(data.potential_data.interfaces) == 1
+        points = data.potential_data.interfaces[0]
+        limits = []
+        for limit in data.stops_on:
+            # the potential field associated to the limit
+            # is used to determine where lie the points defining the fault
+            v = np.mean(model.faults[limit](points))
+            assert v != 0, f"inconsitent limit potential in {fault} stops on {limit}"
+            limits.append((limit, -1 if v < 0 else 1))
+        relations[fault] = limits
+    return relations
+
+
+class CacheDir:
+    def __init__(self, data):
+        filepath = data["filepath"]
+        path = filepath.parent / (filepath.stem + ".cache")
+        path.mkdir(exist_ok=True)
+        origin = path / "origin"
+        valid = origin.exists()
+        if valid:
+            assert origin.is_file()
+            with origin.open() as f:
+                try:
+                    _path = Path(f.readline().strip())
+                    _timestamp = int(f.readline().strip())
+                    valid = _path == filepath and _timestamp == data["timestamp"]
+                except Exception:
+                    valid = False
+        if not valid:
+            with origin.open("w") as f:
+                print(data["filepath"].as_posix(), file=f)
+                print(data["timestamp"], file=f)
+        self.path = path
+        self.valid = valid
+
+    def _source_from_key(self, key):
+        return self.path / f"{key}.npy"
+
+    def __getitem__(self, key):
+        if not self.valid:
+            return None
+        source = self._source_from_key(key)
+        if not source.is_file():
+            return None
+        return np.load(source)
+
+    def __setitem__(self, key, a):
+        np.save(self._source_from_key(key), a)
+
+
+class GeologicalModel:
+    @staticmethod
+    def extract_data(path, *, skip_topography=False, convert_to_yaml=None):
+        # print()
+        # print('Loading model from:', filename)
+        # print()
+        path = Path(path)
+        assert path.is_file(), f"could not find {path.as_posix()}"
+        if path.suffix == ".xml":
+            data = extract_data_from_legacy_geomodeller_file(
+                path, scalardt, skip_topography=skip_topography
+            )
+            if convert_to_yaml is not None:
+                with open(convert_to_yaml, "w") as f:
+                    print(yaml.dump(data), file=f)
+        elif path.suffix == ".yaml":
+            with path.open() as f:
+                data = yaml.load(f)
+        else:
+            msg = "Unknown file extension."
+            raise OSError(msg)
+        assert "path" not in data
+        data["filepath"] = path
+        assert "timestamp" not in data
+        # last modification time in nanoseconds
+        data["timestamp"] = path.stat().st_mtime_ns
+        return data
+
+    def __init__(self, data, **kwargs):
+        if type(data) is str:
+            data = GeologicalModel.extract_data(data, **kwargs)
+        cache = CacheDir(data)
+        self.box = data["box"]
+        self.crs = CRSInfo(*data["crs"])
+        self.pile = data["pile"]
+        assert self.pile.reference in ("top", "base")
+        self.faults_data = data["faults_data"]
+        self.topography = data["topography"]
+        formations = data["formations"]
+        faults = {}
+        fault_ellipsoids = {}
+        fault_drifts = {}
+        faults_data = data["faults_data"]
+        for name, data in faults_data.items():
+            # print('Registering fault:', name)
+            field = None
+            ellipsoid = None
+            if data.infinite:
+                field = fault_potential(cache, f"fault-{name}", data)
+            else:
+                field, ellipsoid = finite_fault(cache, f"fault-{name}", data)
+            faults[name] = pypotential.Fault(field)
+            if ellipsoid is not None:
+                faults[name].stops_on(ellipsoid)
+                fault_ellipsoids[name] = ellipsoid
+        for name, data in faults_data.items():
+            fault = faults[name]
+            filtered_limits = []
+            for limit_name in data.stops_on:
+                if limit_name not in faults:
+                    pass
+                else:
+                    filtered_limits.append(limit_name)
+            data.stops_on = filtered_limits
+            for limit_name in data.stops_on:
+                fault.stops_on(faults[limit_name])
+        for name, fault in faults.items():
+            if name in fault_ellipsoids:
+                fault_drifts[name] = pypotential.make_finite_drift(
+                    fault, fault_ellipsoids[name]
+                )
+            else:
+                fault_drifts[name] = pypotential.make_drift(fault)
+        fields = []
+        values = []
+        relations = []
+        formation_names = [f.name for f in formations]
+        pile_formations = []
+        series_info = {}
+
+        def register_pile_formation(formation):
+            assert formation in formation_names, (
+                "Unknown formation " + formation + " in stratigraphic column!"
+            )
+            pile_formations.append(formation)
+
+        found_dummy_formation = False
+        pile = self.pile
+        for Sk, serie in enumerate(pile.all_series):
+            potdata = serie.potential_data
+            if potdata:
+                active_faults = {}
+                drifts = drift_basis(potdata)
+                if serie.influenced_by_fault:
+                    for name in serie.influenced_by_fault:
+                        try:
+                            drifts.append(fault_drifts[name])
+                            active_faults[name] = len(drifts) - 1
+                        except KeyError:
+                            pass
+                # print('Field for serie', serie.name, 'has', len(drifts), 'drifts.')
+                # print('with covariance:', covariance_data(potdata))
+                field = make_potential(cache, serie.name, potdata, drifts)
+                interfaces = []
+                for i, interface in enumerate(potdata.interfaces):
+                    if len(interface) == 0:
+                        pass
+                    else:
+                        mean_field_value = np.mean(field(interface))
+                        values.append(mean_field_value)
+                        formation = serie.formations[i]
+                        interfaces.append(
+                            (pile.reference + "-" + formation, mean_field_value)
+                        )
+                        fields.append(field)
+                        register_pile_formation(formation)
+                        relations.append(serie.relation)
+                series_info[serie.name] = SeriesInfo(
+                    field, drifts, active_faults, interfaces
+                )
+            else:
+
+                def register_single_formation():
+                    assert len(serie.formations) == 1
+                    # print('registering dummy formation', serie.formations)
+                    register_pile_formation(serie.formations[0])
+
+                if (pile.reference == "base" and Sk == 0) or (
+                    pile.reference == "top" and Sk == len(pile.all_series) - 1
+                ):
+                    register_single_formation()
+                    found_dummy_formation = True
+                else:
+                    pass
+        if not found_dummy_formation:
+            dummy_formations = [f for f in formations if f.is_dummy]
+            # when importing a GeoModeller project the DefaultCover dummy formation is not exported
+            if len(dummy_formations) == 0:
+                assert pile.reference in {"top", "base"}, "Unknown pile reference"
+                dummy = geomodeller_project.Formation(
+                    name="Default" + {"top": "Cover", "base": "Base"}[pile.reference],
+                    color=(0.3,) * 3,
+                    is_dummy=True,
+                )
+                formations.append(dummy)
+                dummy_formations = [dummy]
+            assert len(dummy_formations) == 1, "A dummy formation is needed!"
+            if pile.reference == "base":
+                pile_formations.insert(0, dummy_formations[0].name)
+            else:
+                pile_formations.append(dummy_formations[0].name)
+        assert len(fields) == len(relations)
+        assert len(fields) + 1 == len(pile_formations)
+        self.fields = fields
+        self.formations = formations
+        self.formation_colors = {f.name: f.color for f in formations}
+        self.pile_formations = pile_formations
+        self.series_info = series_info
+        self.values = values
+        self.relations = relations
+        self.faults = faults
+        self.fault_ellipsoids = fault_ellipsoids
+        self.fault_drifts = fault_drifts
+        self.fault_stops_on = compute_fault_stops_on(self)
+
+    def collect_pile_information(self):
+        relations = list(self.relations)
+        if self.pile.reference == "base":
+            relations.insert(0, "onlap")
+        else:
+            relations.append("onlap")
+        return [
+            PileInfo(formation, self.formation_colors[formation], relation)
+            for formation, relation in zip(self.pile_formations, relations)
+        ]
+
+    def nbformations(self):
+        return len(self.pile_formations)
+
+    def nbcontacts(self):
+        # FIXME: this should have to be define rigorously...
+        return len(self.fields)
+
+    def getbox(self):
+        xmin, xmax = (self.box["Xmin"], self.box["Xmax"])
+        ymin, ymax = (self.box["Ymin"], self.box["Ymax"])
+        zmin, zmax = (self.box["Zmin"], self.box["Zmax"])
+        return Box(xmin, ymin, zmin, xmax, ymax, zmax)
+
+    def bbox(self):
+        xmin, xmax = (self.box["Xmin"], self.box["Xmax"])
+        ymin, ymax = (self.box["Ymin"], self.box["Ymax"])
+        zmin, zmax = (self.box["Zmin"], self.box["Zmax"])
+        return BBox3(xmin, xmax, ymin, ymax, zmin, zmax)
+
+    def domain(self, x, y, z):
+        return self.rank((x, y, z))
+
+    def rank_without_topography(self, p):
+        n = len(self.fields)
+        j1 = 0
+        j2 = n
+        # parcourir les erode decroissants
+        erosion_surfaces = [i for i in range(n) if self.relations[i] == "erode"]
+        for i in reversed(erosion_surfaces):
+            field = self.fields[i]
+            vp = field(p)
+            vi = self.values[i]
+            if vp > vi:
+                j1 = i
+                break
+            j2 = i
+        # parcourir les onlap croissants dans l'intervalle j1 j2
+        rank = j1
+        assert j1 == 0 or (
+            self.relations[j1] == "erode" and self.fields[j1](p) > self.values[j1]
+        )
+        rank = j1
+        for i in range(j1, j2):
+            field = self.fields[i]
+            vp = field(p)
+            vi = self.values[i]
+            # when i=j1>0 we are sure that vp<vi (previous test)
+            if vp < vi:
+                break
+            rank += 1
+        # print ("Which domain:", p[0], p[1], p[2], rank+1)
+        return rank + 1
+
+    def rank(self, p, consider_topography=True):
+        if consider_topography and self.topography(p) > 0:
+            return 0
+        return self.rank_without_topography(p)
+
+    def intersect(
+        self,
+        p1,
+        p2,
+        consider_formations=True,
+        consider_faults=False,
+        consider_topography=True,
+        precision=0.01,
+    ):
+        # print('-> intersect', p1, p2)
+        p = point_between(p1, p2, self.topography, 0, precision)
+        if p is not None:
+            return Intersection(p, self.topography)
+        # print('            no topo')
+        # test fault potential fields
+        if consider_faults:
+            for name, fault in self.faults.items():
+                p = point_between(p1, p2, fault, 0, precision)
+                if p is not None:
+                    if self.is_fault_point_valid(p, name, consider_topography):
+                        return Intersection(p, fault, fault=name)
+        # test potential fields
+        # print('            no faults')
+        if consider_formations:
+            n = len(self.fields)
+            for i in range(n):
+                p = point_between(p1, p2, self.fields[i], self.values[i], precision)
+                if p is not None:
+                    if self.is_valid(p, i, consider_topography):
+                        # print('            Found:', n, p, 'shape:', p.shape)
+                        return Intersection(p, self.fields[i], self.values[i])
+            # print('            Nothing!')
+        return None
+
+    def is_valid(self, p, rank, with_topography=True):
+        if with_topography and self.topography(p) > 0:
+            return False
+        n = len(self.fields)
+        for i in range(rank + 1, n):
+            Ri = self.relations[i]
+            if Ri == "erode":
+                field = self.fields[i]
+                vp = field(p)
+                vi = self.values[i]
+                if vp > vi:
+                    return False
+        R = self.relations[rank]
+        if R == "onlap":
+            for i in reversed(range(rank)):
+                field = self.fields[i]
+                vp = field(p)
+                vi = self.values[i]
+                if vp < vi:
+                    return False
+                Ri = self.relations[i]
+                if Ri == "erode":
+                    break
+        return True
+
+    def is_fault_point_valid(self, p, fault, with_topography=True):
+        if with_topography and self.topography(p) > 0:
+            return False
+        if self.is_finite_fault(fault) and self.fault_ellipsoids[fault](p) >= 1:
+            return False
+        for limit_info in self.fault_stops_on[fault]:
+            limit, side = limit_info
+            if self.faults[limit](p) * side < 0:
+                return False
+        return True
+
+    def rank_colors(self):
+        for formation in self.pile_formation:
+            result.append(self.formation_colors[formation])
+        return result
+
+    def stats(self):
+        result = []
+        minimum_distances = []
+
+        def register(name, data):
+            ng = data.nb_gradients
+            nc = data.nb_contact_points
+            result.append(
+                f"{name} has {ng:d} gradients values "
+                f"and {nc:d} contact points which make {ng + 3 * nc} dof"
+            )
+            mdg, mdc = data.minimum_distances()
+            result.append(f"\tminimum distances {mdg:f} {mdc:f}")
+            minimum_distances.append((mdg, mdc))
+
+        for serie, info in self.series_info.items():
+            if info.field is None:
+                result.append(serie + "has no potential field")
+            else:
+                register(f"serie {serie}", info.field.data())
+        for fault, info in self.faults.items():
+            register(f"fault {fault}", info.potential_field.data())
+        md = np.array(minimum_distances)
+        mdg = md[:, 0]
+        mdc = md[:, 1]
+        if np.any(mdg >= 0):
+            result.append(
+                f"Minimum distance between gradients: {np.min(mdg[mdg >= 0]):f}"
+            )
+        if np.any(mdg < 0):
+            result.append(
+                f"There are {np.sum(mdg < 0):d} series with a single gradient data!"
+            )
+        if np.any(mdg == 0):  # This shall not happen (kriging matrix would be singular)
+            result.append(
+                f"There are {np.sum(mdg == 0):d} with confunded gradient data!"
+            )
+        if np.any(mdc >= 0):
+            result.append(
+                f"Minimum distance between contacts: {np.min(mdc[mdc >= 0]):f}"
+            )
+        if np.any(mdc < 0):
+            result.append(
+                f"There are {np.sum(mdc < 0):d} series with a single contact data!"
+            )
+        if np.any(mdc == 0):  # This shall not happen (kriging matrix would be singular)
+            result.append(
+                f"There are {np.sum(mdc == 0):d} series with a condunded contact data!"
+            )
+        return "\n".join(result)
+
+    def diagonal_section(self, flip=False):
+        box = self.getbox()
+        if flip:
+            diagonal = Polyline([[box.xmin, box.ymax, 0], [box.xmax, box.ymin, 0]])
+        else:
+            diagonal = Polyline([[box.xmin, box.ymin, 0], [box.xmax, box.ymax, 0]])
+        return VerticalSection(diagonal, box.zmin, box.zmax)
+
+    def x_section(self, x=None, flip=False):
+        box = self.getbox()
+        x = x or 0.5 * (box.xmin + box.xmax)
+        if flip:
+            path = Polyline([[x, box.ymax, 0], [x, box.ymin, 0]])
+        else:
+            path = Polyline([[x, box.ymin, 0], [x, box.ymax, 0]])
+        return VerticalSection(path, box.zmin, box.zmax)
+
+    def y_section(self, y=None, flip=False):
+        box = self.getbox()
+        y = y or 0.5 * (box.ymin + box.ymax)
+        if flip:
+            path = Polyline([[box.xmax, y, 0], [box.xmin, y, 0]])
+        else:
+            path = Polyline([[box.xmin, y, 0], [box.xmax, y, 0]])
+        return VerticalSection(path, box.zmin, box.zmax)
+
+    def ranks_to_rgb_picture(self, ranks, atmopshere_color=(0, 0, 0)):
+        picture = np.zeros((*ranks.shape, 3), dtype=np.uint8)
+        pile = self.collect_pile_information()
+        rank_color = [atmopshere_color] + [
+            [int(col * 255) for col in formation.color] for formation in pile
+        ]
+        assert np.all((ranks >= 0) & (ranks < len(rank_color))), (
+            "Inconsistency in domain indexing!"
+        )
+        for fi, color in enumerate(rank_color):
+            picture[ranks == fi] = color
+        return picture
+
+    def is_finite_fault(self, name):
+        return name in self.fault_ellipsoids
+
+    @property
+    def has_finite_faults(self):
+        return any(name in self.fault_ellipsoids for name in self.faults)
+
+    def rgb_picture(
+        self,
+        section,
+        width,
+        height=None,
+        atmopshere_color=(0, 0, 0),
+        return_ranks=False,
+    ):
+        if height is None:
+            width, height = image_ratio(section, width)
+        ranks = np.array([self.rank(p) for p in section.grid(width, height)])
+        ranks.shape = width, height
+        ranks = np.transpose(ranks)[::-1]
+        picture = self.ranks_to_rgb_picture(ranks, atmopshere_color)
+        if return_ranks:
+            return picture, ranks
+        return picture
+
+    def topography_as_elevation_surface(self):
+        topo = self.topography
+        if hasattr(topo, "origin"):
+            zmap = np.transpose(topo.z)[::-1]
+            return ElevationRaster(topo.origin, topo.steps, zmap)
+        return ConstantElevationSurface(topo.z)
+
+    def implicit_topography(self):
+        return ImplicitTopography(self.topography_as_elevation_surface())