emerge 0.4.7__py3-none-any.whl → 0.4.9__py3-none-any.whl

emerge/_emerge/selection.py

@@ -0,0 +1,596 @@
+# EMerge is an open source Python based FEM EM simulation module.
+# Copyright (C) 2025  Robert Fennis.
+
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU General Public License
+# as published by the Free Software Foundation; either version 2
+# of the License, or (at your option) any later version.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, see
+# <https://www.gnu.org/licenses/>.
+
+from __future__ import annotations
+import gmsh
+import numpy as np
+from scipy.spatial import ConvexHull
+from .cs import Axis, CoordinateSystem, _parse_vector, Plane
+from typing import Callable, TypeVar
+
+def align_rectangle_frame(pts3d: np.ndarray, normal: np.ndarray) -> dict[str, np.ndarray]:
+    """Tries to find a rectangle as convex-hull of a set of points with a given normal vector.
+
+    Args:
+        pts3d (np.ndarray): The points (N,3)
+        normal (np.ndarray): The normal vector.
+
+    Returns:
+        dict[str, np.ndarray]: The output data
+    """
+    
+    # 1. centroid
+    Omat = np.squeeze(np.mean(pts3d, axis=0))
+
+    # 2. build e_x, e_y
+    n = np.squeeze(normal/np.linalg.norm(normal))
+    seed = np.array([1.,0.,0.])
+    if abs(seed.dot(n)) > 0.9:
+        seed = np.array([0.,1.,0.])
+    e_x = seed - n*(seed.dot(n))
+    e_x /= np.linalg.norm(e_x)
+    e_y = np.cross(n, e_x)
+
+    # 3. project into 2D
+    pts2d = np.vstack([[(p-Omat).dot(e_x), (p-Omat).dot(e_y)] for p in pts3d])
+
+    # 4. convex hull
+    hull = ConvexHull(pts2d)
+    hull_pts = pts2d[hull.vertices]
+
+    # 5. rotating calipers: find min-area rectangle
+    best = (None, np.inf, None)  # (angle, area, (xmin,xmax,ymin,ymax))
+    for i in range(len(hull_pts)):
+        p0 = hull_pts[i]
+        p1 = hull_pts[(i+1)%len(hull_pts)]
+        edge = p1 - p0
+        theta = -np.arctan2(edge[1], edge[0])  # rotate so edge aligns with +X
+        R = np.array([[np.cos(theta), -np.sin(theta)],
+                      [np.sin(theta),  np.cos(theta)]])
+        rot = hull_pts.dot(R.T)
+        xmin, ymin = rot.min(axis=0)
+        xmax, ymax = rot.max(axis=0)
+        area = (xmax-xmin)*(ymax-ymin)
+        if area < best[1]:
+            best = (theta, area, (xmin,xmax,ymin,ymax), R)
+
+    theta, _, (xmin,xmax,ymin,ymax), R = best
+
+    # 6. rectangle axes in 3D
+    u =  np.cos(-theta)*e_x + np.sin(-theta)*e_y
+    v = -np.sin(-theta)*e_x + np.cos(-theta)*e_y
+
+    # corner points in 3D:
+    corners = []
+    for a in (xmin, xmax):
+        for b in (ymin, ymax):
+            # back-project to the original 2D frame:
+            p2 = np.array([a, b]).dot(R)  # rotate back
+            P3 = Omat + p2[0]*e_x + p2[1]*e_y
+            corners.append(P3)
+
+    return {
+      "origin": Omat,
+      "axes": (u, v, n),
+      "corners": np.array(corners).reshape(4,3)
+    }
+
+TSelection = TypeVar("TSelection", bound="Selection")
+
+# Your custom alphabet
+ALPHABET = (
+    "abcdefghijklmnopqrstuvwxyz"
+    "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
+    "012345%#"
+)
+# Map char → int and int → char
+CHAR_TO_VAL = {ch: i for i, ch in enumerate(ALPHABET)}
+VAL_TO_CHAR = {i: ch for i, ch in enumerate(ALPHABET)}
+
+def encode_data(values: tuple[float,...]) -> str:
+    """
+    Convert a tuple of floats into a custom base64-like encoded string
+    using 16-bit float representation and a 64-character alphabet.
+    """
+    # Convert floats to 16-bit half-floats
+    arr = np.array(values, dtype=np.float16)
+    # Get raw bytes
+    byte_data = arr.tobytes()
+    # Convert bytes to a bitstring
+    bitstring = ""
+    for byte in byte_data:
+        bitstring += f"{byte:08b}"
+    
+    # Pad the bitstring to a multiple of 6 bits
+    pad_len = (6 - len(bitstring) % 6) % 6
+    bitstring += "0" * pad_len
+    
+    # Encode 6 bits at a time
+    encoded = ""
+    for i in range(0, len(bitstring), 6):
+        chunk = bitstring[i:i+6]
+        val = int(chunk, 2)
+        encoded += VAL_TO_CHAR[val]
+    
+    # Optionally store how many pad bits we added
+    # so we can remove them during decoding.
+    # We'll prepend it as one character encoding pad_len (0-5).
+    encoded = VAL_TO_CHAR[pad_len] + encoded
+    return encoded
+
+def decode_data(encoded: str) -> tuple[float,...]:
+    """
+    Decode a string produced by floats_to_custom_string
+    back into a tuple of float64 values.
+    """
+    # The first character encodes how many zero bits were padded
+    pad_char = encoded[0]
+    pad_len = CHAR_TO_VAL[pad_char]
+    data = encoded[1:]
+    
+    # Convert each char back to 6 bits
+    bitstring = ""
+    for ch in data:
+        val = CHAR_TO_VAL[ch]
+        bitstring += f"{val:06b}"
+    
+    # Remove any padding bits at the end
+    if pad_len > 0:
+        bitstring = bitstring[:-pad_len]
+    
+    # Split into bytes
+    byte_data = []
+    for i in range(0, len(bitstring), 8):
+        byte_chunk = bitstring[i:i+8]
+        if len(byte_chunk) < 8:
+            break
+        byte_val = int(byte_chunk, 2)
+        byte_data.append(byte_val)
+    
+    byte_array = bytes(byte_data)
+    # Recover as float16
+    arr = np.frombuffer(byte_array, dtype=np.float16)
+    # Convert back to float64 for higher precision
+    return tuple(np.array(arr, dtype=np.float64))
+
+class Selection:
+    """A generalized class representing a slection of tags.
+
+    """
+    dim: int = -1
+    def __init__(self, tags: list[int] = None):
+
+        self._tags: set[int] = {}
+
+        if tags is not None:
+            if not isinstance(tags, (list,set,tuple)):
+                raise TypeError(f'Argument tags must be of type list, tuple or set, instead its {type(tags)}')
+            self._tags = set(tags)
+    
+    @staticmethod
+    def from_dim_tags(dim: int, tags: list[int]) -> DomainSelection | PointSelection | FaceSelection | EdgeSelection:
+        if dim==0:
+            return PointSelection(tags)
+        elif dim==1:
+            return EdgeSelection(tags)
+        elif dim==2:
+            return FaceSelection(tags)
+        elif dim==3:
+            return DomainSelection(tags)
+        raise ValueError(f'Dimension must be 0,1,2 or 3. Not {dim}')
+    
+    @property
+    def tags(self) -> list[int]:
+        return list(self._tags)
+    
+    @property
+    def color_rgb(self) -> tuple[int,int,int]:
+        return (0.5,0.5,1.0)
+    
+    @property
+    def centers(self) -> list[tuple[float, float, float],]:
+        return [gmsh.model.occ.get_center_of_mass(self.dim, tag) for tag in self.tags]
+    
+    @property
+    def opacity(self) -> float:
+        return 0.6
+    ####### DUNDER METHODS
+    def __repr__(self) -> str:
+        return f'{type(self).__name__}({self.tags})'
+    
+    ####### PROPERTIES
+    @property
+    def dimtags(self) -> list[tuple[int,int]]:
+        return [(self.dim, tag) for tag in self.tags]
+    
+    @property
+    def center(self) -> np.ndarray | list[np.ndarray]:
+        if len(self.tags)==1:
+            return gmsh.model.occ.getCenterOfMass(self.dim, self.tags[0])
+        else:
+            return [gmsh.model.occ.getCenterOfMass(self.dim, tag) for tag in self.tags]
+    
+    @property
+    def points(self) -> np.ndarray | list[np.ndarray]:
+        '''A list of 3D coordinates of all nodes comprising the selection.'''
+        points = gmsh.model.get_boundary(self.dimtags, recursive=True)
+        coordinates = [gmsh.model.getValue(*p, []) for p in points]
+        return coordinates
+    
+    @property
+    def bounding_box(self) -> tuple[np.ndarray, np.ndarray]:
+        if len(self.tags)==1:
+            return gmsh.model.occ.getBoundingBox(self.dim, self.tags[0])
+        else:
+            minx = miny = minz = 1e10
+            maxx = maxy = maxz = -1e10
+            for tag in self.tags:
+                x0, y0, z0, x1, y1, z1 = gmsh.model.occ.getBoundingBox(self.dim, tag)
+                minx = min(minx, x0)
+                miny = min(miny, y0)
+                minz = min(minz, z0)
+                maxx = max(maxx, x1)
+                maxy = max(maxy, y1)
+                maxz = max(maxz, z1)
+            return (minx, miny, minz), (maxx, maxy, maxz)
+        
+    def exclude(self, xyz_excl_function: Callable = lambda x,y,z: True, plane: Plane = None, axis: Axis = None) -> Selection:
+        """Exclude points by evaluating a function(x,y,z)-> bool
+
+        This modifies the selection such that the selection does not contain elements
+        of this selection of which the center of mass is excluded by the exclusion function.
+
+        Args:
+            xyz_excl_function (Callable): A callable for (x,y,z) that returns True if the point should be excluded.
+
+        Returns:
+            Selection: This Selection modified without the excluded points.
+        """
+        include = [~xyz_excl_function(*gmsh.model.occ.getCenterOfMass(*tag)) for tag in self.dimtags]
+        
+        if axis is not None:
+            norm = axis.np
+            include2 = [abs(gmsh.model.getNormal(tag, np.array([0,0]))@norm)<0.9 for tag in self.tags]
+            include = [i1 for i1, i2 in zip(include, include2) if i1 and i2]
+        self._tags = [t for incl, t in zip(include, self._tags) if incl]
+        return self
+    
+    def isolate(self, xyz_excl_function: Callable = lambda x,y,z: True, plane: Plane = None, axis: Axis = None) -> Selection:
+        """Include points by evaluating a function(x,y,z)-> bool
+
+        This modifies the selection such that the selection does not contain elements
+        of this selection of which the center of mass is excluded by the exclusion function.
+
+        Args:
+            xyz_excl_function (Callable): A callable for (x,y,z) that returns True if the point should be excluded.
+
+        Returns:
+            Selection: This Selection modified without the excluded points.
+        """
+        include1 = [xyz_excl_function(*gmsh.model.occ.getCenterOfMass(*tag)) for tag in self.dimtags]
+        
+        if axis is not None:
+            norm = axis.np
+            include2 = [(gmsh.model.getNormal(tag, np.array([0,0]))@norm)>0.99 for tag in self.tags]
+            include1 = [i1 for i1, i2 in zip(include1, include2) if i1 and i2]
+        self._tags = [t for incl, t in zip(include1, self._tags) if incl]
+        return self
+
+    def __operable__(self, other: Selection) -> None:
+        if not self.dim == other.dim:
+            raise ValueError(f'Selection dimensions must be equal. Trying to operate on dim {self.dim} and {other.dim}')
+        pass
+
+    def __add__(self, other: TSelection) -> TSelection:
+        self.__operable__(other)
+        return Selection.from_dim_tags(self.dim, self._tags + other._tags)
+    
+    def __and__(self, other: TSelection) -> TSelection:
+        self.__operable__(other)
+        return Selection.from_dim_tags(self.dim, self._tags.intersection(other._tags))
+    
+    def __or__(self, other: TSelection) -> TSelection:
+        self.__operable__(other)
+        return Selection.from_dim_tags(self.dim, self._tags.union(other._tags))
+    
+    def __sub__(self, other: TSelection) -> TSelection:
+        self.__operable__(other)
+        return Selection.from_dim_tags(self.dim, self._tags.difference(other.tags))
+    
+
+
+class PointSelection(Selection):
+    """A Class representing a selection of points.
+
+    """
+    dim: int = 0
+    def __init__(self, tags: list[int] = None):
+        super().__init__(tags)
+
+class EdgeSelection(Selection):
+    """A Class representing a selection of edges.
+
+    """
+    dim: int = 1
+    def __init__(self, tags: list[int] = None):
+        super().__init__(tags)
+
+class FaceSelection(Selection):
+    """A Class representing a selection of Faces.
+
+    """
+    dim: int = 2
+    def __init__(self, tags: list[int] = None):
+        super().__init__(tags)
+
+    @property
+    def normal(self) -> np.ndarray:
+        ''' Returns a 3x3 coordinate matrix of the XY + out of plane basis matrix defining the face assuming it can be projected on a flat plane.'''
+        ns = [gmsh.model.getNormal(tag, np.array([0,0])) for tag in self.tags]
+        return ns[0]
+    
+    def rect_basis(self) -> tuple[CoordinateSystem, tuple[float, float]]:
+        ''' Returns a dictionary with keys: origin, axes, corners. The axes are the 3D basis vectors of the rectangle. The corners are the 4 corners of the rectangle.
+        
+        Returns:
+            cs: CoordinateSystem: The coordinate system of the rectangle.
+            size: tuple[float, float]: The size of the rectangle (width [m], height[m])
+        '''
+        if len(self.tags) != 1:
+            raise ValueError('rect_basis only works for single face selections')
+        
+        pts3d = self.points
+        normal = self.normal
+        data = align_rectangle_frame(pts3d, normal)
+        plane = data['axes'][:2]
+        origin = data['origin']
+
+        cs = CoordinateSystem(Axis(plane[0]), Axis(plane[1]), Axis(data['axes'][2]), origin)
+
+        size1 = np.linalg.norm(data['corners'][1] - data['corners'][0])
+        size2 = np.linalg.norm(data['corners'][2] - data['corners'][0])
+
+        if size1>size2:
+            cs.swapxy()
+            return cs, (size1, size2)
+        else:
+            return cs, (size2, size1)
+            
+    def sample(self, Npts: int) -> tuple[np.ndarray, np.ndarray, np.ndarray] | list[tuple[np.ndarray, np.ndarray, np.ndarray]]:
+        ''' Sample the surface at the compiler defined parametric coordinate range.
+        This function usually returns a square region that contains the surface.
+        
+        Returns:
+        --------
+            X: np.ndarray
+                a NxN numpy array of X coordinates.
+            Y: np.ndarray
+                a NxN numpy array of Y coordinates.
+            Z: np.ndarray
+                a NxN numpy array of Z coordinates'''
+        coordset = []
+        for tag in self.tags:
+            tags, coord, param = gmsh.model.mesh.getNodes(2, tag, includeBoundary=True)
+            
+            uss = param[0::2]
+            vss = param[1::2]
+
+            umin = min(uss)
+            umax = max(uss)
+            vmin = min(vss)
+            vmax = max(vss)
+
+            us = np.linspace(umin, umax, Npts)
+            vs = np.linspace(vmin, vmax, Npts)
+
+            U, V = np.meshgrid(us, vs, indexing='ij')
+
+            shp = U.shape
+
+            uax = U.flatten()
+            vax = V.flatten()
+            
+            pcoords = np.zeros((2*uax.shape[0],))
+
+            pcoords[0::2] = uax
+            pcoords[1::2] = vax
+
+            coords = gmsh.model.getValue(2, tag, pcoords).reshape(-1,3).T
+           
+            coordset.append((coords[0,:].reshape(shp), 
+                             coords[1,:].reshape(shp), 
+                             coords[2,:].reshape(shp)))
+            
+        X = [c[0] for c in coordset]
+        Y = [c[1] for c in coordset]
+        Z = [c[2] for c in coordset]
+        if len(X) == 1:
+            X = X[0]
+            Y = Y[0]
+            Z = Z[0]
+        return X, Y, Z
+    
+class DomainSelection(Selection):
+    """A Class representing a selection of domains.
+
+    """
+    dim: int = 3
+    def __init__(self, tags: list[int] = None):
+        super().__init__(tags)
+
+SELECT_CLASS: dict[int, type[Selection]] = {
+    0: PointSelection,
+    1: EdgeSelection,
+    2: FaceSelection,
+    3: DomainSelection
+}
+
+######## SELECTOR
+
+class Selector:
+    """A class instance with convenient methods to generate selections using method chaining.
+
+    Use the specific properties and functions in a "language" like way to make selections.
+
+    To specify what to select, use the .node, .edge, .face or .domain property.
+    These properties return the Selector after which you can say how to execute a selection.
+
+
+    """
+    def __init__(self):
+        self._current_dim: int = -1
+    
+    ## DIMENSION CHAIN
+    @property
+    def node(self) -> Selector:
+        self._current_dim = 0
+        return self
+
+    @property
+    def edge(self) -> Selector:
+        self._current_dim = 1
+        return self
+    
+    @property
+    def face(self) -> Selector:
+        self._current_dim = 2
+        return self
+
+    @property
+    def domain(self) -> Selector:
+        self._current_dim = 3
+        return self
+    
+    def near(self,
+             x: float,
+             y: float,
+             z: float = 0) -> Selection | PointSelection | EdgeSelection | FaceSelection | DomainSelection:
+        """Returns a selection of the releative dimeions by which of the instances is most proximate to a coordinate.
+
+        Args:
+            x (float): The X-coordinate
+            y (float): The Y-coordinate
+            z (float, optional): The Z-coordinate. Defaults to 0.
+
+        Returns:
+            Selection | PointSelection | EdgeSelection | FaceSelection | DomainSelection: The resultant selection.
+        """
+        dimtags = gmsh.model.getEntities(self._current_dim)
+        
+        
+        dists = [np.linalg.norm(np.array([x,y,z]) - gmsh.model.occ.getCenterOfMass(*tag)) for tag in dimtags]
+        index_of_closest = np.argmin(dists)
+
+        return SELECT_CLASS[self._current_dim]([dimtags[index_of_closest][1],])
+    
+    def inlayer(self, 
+                x: float,
+                y: float,
+                z: float,
+                vector: tuple[float, float, float] | np.ndarray | Axis) -> FaceSelection | EdgeSelection | DomainSelection:
+        '''Returns a list of selections that are in the layer defined by the plane normal vector and the point (x,y,z)
+        
+        The layer is specified by two infinite planes normal to the provided vector. The first plane is originated
+        at the provided origin. The second plane is placed such that it contains the point origin+vector.
+
+        Args:
+            x (float): The X-coordinate from which to select
+            y (float): The Y-coordinate from which to select
+            z (float): The Z-coordinate from which to select
+            vector (np.ndarray, tuple, Axis): A vector with length in (meters) originating at the origin.
+
+        Returns:
+            Selection | PointSelection | EdgeSelection | FaceSelection | DomainSelection: The resultant selection.
+
+        '''
+        vector = _parse_vector(vector)
+        
+        dimtags = gmsh.model.getEntities(self._current_dim)
+
+        coords = [gmsh.model.occ.getCenterOfMass(*tag) for tag in dimtags]
+
+        L = np.linalg.norm(vector)
+        vector = vector / L
+
+        output = []
+        for i, c in enumerate(coords):
+            c_local = c - np.array([x,y,z])
+            if 0 < np.dot(vector, c_local) < L:
+                output.append(dimtags[i][1])
+        return SELECT_CLASS[self._current_dim](output)
+    
+    def inplane(self,
+                x: float,
+                y: float,
+                z: float,
+                nx: float,
+                ny: float,
+                nz: float,
+                tolerance: float = 1e-6) -> FaceSelection:
+        """Returns a FaceSelection for all faces that lie in a provided infinite plane
+        specified by an origin plus a plane normal vector.
+
+        Args:
+            x (float): The plane origin X-coordinate
+            y (float): The plane origin Y-coordinate
+            z (float): The plane origin Z-coordinate
+            nx (float): The plane normal X-component
+            ny (float): The plane normal Y-component
+            nz (float): The plane normal Z-component
+            tolerance (float, optional): An in plane tolerance (displacement and normal dot product). Defaults to 1e-6.
+
+        Returns:
+            FaceSelection: All faces that lie in the specified plane
+        """
+        orig = np.array([x,y,z])
+        norm = np.array([nx,ny,nz])
+        norm = norm/np.linalg.norm(norm)
+        
+        dimtags = gmsh.model.getEntities(2)
+        coords = [gmsh.model.occ.getCenterOfMass(*tag) for tag in dimtags]
+        normals = [gmsh.model.get_normal(t, (0,0)) for d,t, in dimtags]
+        tags = []
+        for (d,t), o, n in zip(dimtags, coords, normals):
+            normdist = np.abs((o-orig)@norm)
+            dotnorm = np.abs(n@norm)
+            if normdist < tolerance and dotnorm > 1-tolerance:
+                tags.append(t)
+        return FaceSelection(tags)
+    
+    def code(self, code: str):
+        nums1 = decode_data(code)
+        
+        dimtags = gmsh.model.getEntities(2)
+        
+        scoring = dict()
+        for dim, tag in dimtags:
+            x1, y1, z1 = gmsh.model.occ.getCenterOfMass(2, tag)
+            xmin2, ymin2, zmin2, xmax2, ymax2, zmax2 = gmsh.model.occ.getBoundingBox(dim, tag)
+            nums2 = [x1, y1, z1, xmin2, ymin2, zmin2, xmax2, ymax2, zmax2]
+            score = np.sqrt(sum([(a-b)**2 for a,b in zip(nums1, nums2)]))
+            scoring[tag] = score
+        
+        min_val = min(scoring.values())
+
+        # Find all keys whose value == min_val
+        candidates = [k for k, v in scoring.items() if v == min_val]
+
+        # Pick the lowest key
+        lowest_key = min(candidates)
+        return FaceSelection([lowest_key,])
+
+SELECTOR_OBJ = Selector()