basilisk-engine 0.1.34__py3-none-any.whl → 0.1.36__py3-none-any.whl

basilisk/mesh/mesh.py

@@ -1,234 +1,234 @@
-import numpy as np
-import glm
-import os
-# from pyobjloader import load_model
-from .model import load_model
-from .narrow_bvh import NarrowBVH
-from ..generic.matrices import compute_inertia_moment, compute_inertia_product
-from ..generic.meshes import get_extreme_points_np, moller_trumbore
-from .mesh_from_data import from_data
-import time
-
-
-class Mesh():
-    data: np.ndarray
-    """The mesh vertex data stored as a 4-byte float numpy array. Format will be [position.xyz, uv.xy, normal.xyz, tangent.xyz, bitangent.xyz]"""
-    points: np.ndarray
-    """All the unique points of the mesh given by the model file"""  
-    indices: np.ndarray
-    """Indices of the triangles corresponding to the points array"""  
-    bvh: any
-    """Data structure allowing the access of closest points more efficiently"""
-    volume: float
-    """The volume of the unscaled mesh"""
-    geometric_center: glm.vec3
-    """The geometric center of the mesh"""
-    center_of_mass: glm.vec3
-    """The center of mass of the mesh calculated from the inertia tensor algorithm"""
-    half_dimensions: glm.vec3
-    """The aligned half dimensions to the untransformed mesh"""
-    bvh: NarrowBVH
-    """BVH for accessing triangle intersections with a line"""
-
-    def __init__(self, data: str | os.PathLike | np.ndarray, custom_format:bool=False) -> None:
-        """
-        Mesh object containing all the data needed to render an object and perform physics/collisions on it
-        Args:
-            data: str
-                path to the .obj file of the model or an array of the mesh data
-            custom_format: bool
-                makes expected changes to the given data if false. Leaves data as given if true
-        """
-        
-        # Verify the path type
-        if isinstance(data, str) or isinstance(data, os.PathLike):  # Load the model from file
-            model = load_model(data, calculate_tangents=True)
-
-             # Get the vertex data
-            if len(model.vertex_data[0]) == 8:
-                self.data = model.vertex_data.copy()
-            else:
-                self.data = np.zeros(shape=(len(model.vertex_data), 8))
-                self.data[:,:3] = model.vertex_data[:,:3]
-                self.data[:,5:] = model.vertex_data[:,3:]
-            
-            # Get tangent data
-            if len(model.tangent_data[0]) == 6:
-                self.data = np.hstack([self.data, model.tangent_data])
-            else:
-                tangents = np.zeros(shape=(len(self.data), 6))
-                tangents[:,:] += [1.0, 0.0, 0.0, 0.0, 1.0, 0.0]
-                self.data = np.hstack([self.data, tangents])
-
-        elif isinstance(data, np.ndarray):
-            model = from_data(data, custom_format)      
-            self.data = model.vertex_data
-            
-        else:                                                       # Invalid data type
-            raise TypeError(f'Invalid path type: {type(data)}. Expected a string or os.path')
-        
-        # Mesh points and triangles used for physics/collisions
-        self.points = model.vertex_points.copy()
-        self.indices = model.point_indices.copy()
-
-        self.hash = hash(str(self.data))
-        self.custom = custom_format
-
-        # generate edges from faces
-        edges = [set() for _ in range(len(self.points))]
-        for face in self.indices:
-            
-            # add points to the dictionary since each point on a face is adjacent to the others
-            for i in range(3): edges[face[i]].update([int(face[(i + 1) % 3]), int(face[(i + 2) % 3])])
-        self.edges = [tuple(adjacent) for adjacent in edges]
-
-        # Model will no longer be used
-        del model
-        
-        # generate geometric data
-        maximum, minimum = get_extreme_points_np(self.points)
-        self.geometric_center = (glm.vec3(maximum) + glm.vec3(minimum)) / 2
-        self.half_dimensions  = maximum - self.geometric_center
-        
-        # volume and center of mass
-        self.volume = 0
-        self.center_of_mass = glm.vec3(0.0)
-        for triangle in self.indices:
-            pts = [glm.vec3(self.points[t]) for t in triangle]
-            det_j = glm.dot(pts[0], glm.cross(pts[1], pts[2]))
-            tet_volume = det_j / 6
-            self.volume += tet_volume
-            self.center_of_mass += tet_volume * (pts[0] + pts[1] + pts[2]) / 4
-        self.center_of_mass /= self.volume
-        
-        # data structrues
-        self.bvh = NarrowBVH(self)
-        
-    def get_inertia_tensor(self, scale: glm.vec3) -> glm.mat3x3:
-        """
-        Gets the inertia tensor of the mesh with the given scale and mass 1
-        """
-        # scale variables
-        center_of_mass = self.center_of_mass * scale
-        volume = self.volume * scale.x * scale.y * scale.z
-        
-        # uses density = 1 to calculate variables, should be the same for mass = 1 since they are only spatial variables
-        points = self.points.copy()
-        points[:, 0] *= scale.x
-        points[:, 1] *= scale.y
-        points[:, 2] *= scale.z
-        
-        ia = ib = ic = iap = ibp = icp = 0
-        for triangle in self.indices:
-            pts = [points[t] for t in triangle]
-            det_j = glm.dot(pts[0], glm.cross(pts[1], pts[2]))
-            
-            ia += det_j * (compute_inertia_moment(pts, 1) + compute_inertia_moment(pts, 2))
-            ib += det_j * (compute_inertia_moment(pts, 0) + compute_inertia_moment(pts, 2))
-            ic += det_j * (compute_inertia_moment(pts, 0) + compute_inertia_moment(pts, 1))
-            iap += det_j * compute_inertia_product(pts, 1, 2)
-            ibp += det_j * compute_inertia_product(pts, 0, 1)
-            icp += det_j * compute_inertia_product(pts, 0, 2)
-            
-        # since tensor was calc with density = 1. we say mass = density / volume = 1 / volume
-        ia = ia / volume / 60 - volume * (center_of_mass[1] ** 2 + center_of_mass[2] ** 2)
-        ib = ib / volume / 60 - volume * (center_of_mass[0] ** 2 + center_of_mass[2] ** 2)
-        ic = ic / volume / 60 - volume * (center_of_mass[0] ** 2 + center_of_mass[1] ** 2)
-        iap = iap / volume / 120 - volume * center_of_mass[1] * center_of_mass[2]
-        ibp = ibp / volume / 120 - volume * center_of_mass[0] * center_of_mass[1]
-        icp = icp / volume / 120 - volume * center_of_mass[0] * center_of_mass[2]
-        
-        return glm.mat3x3(
-            ia, -ibp, -icp,
-            -ibp, ib, -iap,
-            -icp, -iap, ic
-        )
-        
-        # return glm.mat3x3(
-        #     ic, -iap, -icp,
-        #     -iap, ib, -ibp,
-        #     -ibp, -iap, ia
-        # )
-        
-    def get_best_triangle(self, point: glm.vec3, vec: glm.vec3) -> int:
-        """
-        Gets the triangle with the closest intersection, -1 if no intersection is found
-        """
-        indices = self.bvh.get_possible_triangles(point, vec)
-        best_distance = -1
-        best_index = -1
-        
-        point = glm.vec3(point)
-        vec = glm.vec3(vec)
-
-        for triangle in indices:
-            
-            # check if triangle intersects
-            intersection = moller_trumbore(point, vec, [self.points[t] for t in self.indices[triangle]])
-            if not intersection: continue
-            
-            # check if triangle is on correct side of line
-            difference = intersection - self.geometric_center
-            if glm.dot(difference, vec) < 0: continue
-            
-            # determine best distance
-            distance = glm.length(difference)
-            if best_distance < 0 or distance < best_distance: 
-                best_distance = distance
-                best_index = triangle
-                
-        return best_index
-        
-    def get_best_dot(self, vec: glm.vec3) -> int:
-        """
-        Gets the point with the highest normalized dot product to the given vector
-        """
-        triangle = self.bvh.get_best_dot(vec)
-        if triangle == -1: return None
-        index = max(self.indices[triangle], key=lambda t: glm.dot(glm.normalize(self.points[t]), vec))
-        return index
-    
-    def get_best_dot_hill_climbing(self, vec: glm.vec3) -> int:
-        """
-        Gets the point with the highest dot product to the given vector using a hill climbing algorithm. This function is only effective for convex models.
-        """
-        best_index = 0
-        best_dot = glm.dot(self.points[best_index], vec)
-        
-        while True:
-            
-            best_changed = False
-            for index in self.edges[best_index]:
-                
-                dot = glm.dot(self.points[index], vec)
-                if dot > best_dot:
-                    best_dot = dot
-                    best_changed = True
-                    best_index = index
-                    
-            if not best_changed: break
-        
-        return best_index
-    
-    def get_line_collided(self, position: glm.vec3, forward: glm.vec3) -> list[tuple[int, int, int]]:
-        """
-        Determines which triangles are intersecting with the given line segment. Returns the indices of the triangle contained in the mesh points list
-        """
-        return self.bvh.get_line_collided(position, forward)
-
-    def __repr__(self) -> str:
-        size = (self.data.nbytes + self.points.nbytes + self.indices.nbytes) / 1024 / 1024
-        return f'<Basilisk Mesh | {len(self.data)} vertices, {size:.2} mb>'
-    
-    @property
-    def top_right(self): return self.bvh.root.top_right
-    @property
-    def bottom_left(self): return self.bvh.root.bottom_left
-    @property
-    def aabb_points(self): 
-        x1, y1, z1 = self.top_right
-        x2, y2, z2 = self.bottom_left
-        return [glm.vec3(x, y, z) for z in (z1, z2) for y in (y1, y2) for x in (x1, x2)]
-
-    def __hash__(self):
+import numpy as np
+import glm
+import os
+# from pyobjloader import load_model
+from .model import load_model
+from .narrow_bvh import NarrowBVH
+from ..generic.matrices import compute_inertia_moment, compute_inertia_product
+from ..generic.meshes import get_extreme_points_np, moller_trumbore
+from .mesh_from_data import from_data
+import time
+
+
+class Mesh():
+    data: np.ndarray
+    """The mesh vertex data stored as a 4-byte float numpy array. Format will be [position.xyz, uv.xy, normal.xyz, tangent.xyz, bitangent.xyz]"""
+    points: np.ndarray
+    """All the unique points of the mesh given by the model file"""  
+    indices: np.ndarray
+    """Indices of the triangles corresponding to the points array"""  
+    bvh: any
+    """Data structure allowing the access of closest points more efficiently"""
+    volume: float
+    """The volume of the unscaled mesh"""
+    geometric_center: glm.vec3
+    """The geometric center of the mesh"""
+    center_of_mass: glm.vec3
+    """The center of mass of the mesh calculated from the inertia tensor algorithm"""
+    half_dimensions: glm.vec3
+    """The aligned half dimensions to the untransformed mesh"""
+    bvh: NarrowBVH
+    """BVH for accessing triangle intersections with a line"""
+
+    def __init__(self, data: str | os.PathLike | np.ndarray, custom_format:bool=False) -> None:
+        """
+        Mesh object containing all the data needed to render an object and perform physics/collisions on it
+        Args:
+            data: str
+                path to the .obj file of the model or an array of the mesh data
+            custom_format: bool
+                makes expected changes to the given data if false. Leaves data as given if true
+        """
+        
+        # Verify the path type
+        if isinstance(data, str) or isinstance(data, os.PathLike):  # Load the model from file
+            model = load_model(data, calculate_tangents=True)
+
+             # Get the vertex data
+            if len(model.vertex_data[0]) == 8:
+                self.data = model.vertex_data.copy()
+            else:
+                self.data = np.zeros(shape=(len(model.vertex_data), 8))
+                self.data[:,:3] = model.vertex_data[:,:3]
+                self.data[:,5:] = model.vertex_data[:,3:]
+            
+            # Get tangent data
+            if len(model.tangent_data[0]) == 6:
+                self.data = np.hstack([self.data, model.tangent_data])
+            else:
+                tangents = np.zeros(shape=(len(self.data), 6))
+                tangents[:,:] += [1.0, 0.0, 0.0, 0.0, 1.0, 0.0]
+                self.data = np.hstack([self.data, tangents])
+
+        elif isinstance(data, np.ndarray):
+            model = from_data(data, custom_format)      
+            self.data = model.vertex_data
+            
+        else:                                                       # Invalid data type
+            raise TypeError(f'Invalid path type: {type(data)}. Expected a string or os.path')
+        
+        # Mesh points and triangles used for physics/collisions
+        self.points = model.vertex_points.copy()
+        self.indices = model.point_indices.copy()
+
+        self.hash = hash(str(self.data))
+        self.custom = custom_format
+
+        # generate edges from faces
+        edges = [set() for _ in range(len(self.points))]
+        for face in self.indices:
+            
+            # add points to the dictionary since each point on a face is adjacent to the others
+            for i in range(3): edges[face[i]].update([int(face[(i + 1) % 3]), int(face[(i + 2) % 3])])
+        self.edges = [tuple(adjacent) for adjacent in edges]
+
+        # Model will no longer be used
+        del model
+        
+        # generate geometric data
+        maximum, minimum = get_extreme_points_np(self.points)
+        self.geometric_center = (glm.vec3(maximum) + glm.vec3(minimum)) / 2
+        self.half_dimensions  = maximum - self.geometric_center
+        
+        # volume and center of mass
+        self.volume = 0
+        self.center_of_mass = glm.vec3(0.0)
+        for triangle in self.indices:
+            pts = [glm.vec3(self.points[t]) for t in triangle]
+            det_j = glm.dot(pts[0], glm.cross(pts[1], pts[2]))
+            tet_volume = det_j / 6
+            self.volume += tet_volume
+            self.center_of_mass += tet_volume * (pts[0] + pts[1] + pts[2]) / 4
+        self.center_of_mass /= self.volume
+        
+        # data structrues
+        self.bvh = NarrowBVH(self)
+        
+    def get_inertia_tensor(self, scale: glm.vec3) -> glm.mat3x3:
+        """
+        Gets the inertia tensor of the mesh with the given scale and mass 1
+        """
+        # scale variables
+        center_of_mass = self.center_of_mass * scale
+        volume = self.volume * scale.x * scale.y * scale.z
+        
+        # uses density = 1 to calculate variables, should be the same for mass = 1 since they are only spatial variables
+        points = self.points.copy()
+        points[:, 0] *= scale.x
+        points[:, 1] *= scale.y
+        points[:, 2] *= scale.z
+        
+        ia = ib = ic = iap = ibp = icp = 0
+        for triangle in self.indices:
+            pts = [points[t] for t in triangle]
+            det_j = glm.dot(pts[0], glm.cross(pts[1], pts[2]))
+            
+            ia += det_j * (compute_inertia_moment(pts, 1) + compute_inertia_moment(pts, 2))
+            ib += det_j * (compute_inertia_moment(pts, 0) + compute_inertia_moment(pts, 2))
+            ic += det_j * (compute_inertia_moment(pts, 0) + compute_inertia_moment(pts, 1))
+            iap += det_j * compute_inertia_product(pts, 1, 2)
+            ibp += det_j * compute_inertia_product(pts, 0, 1)
+            icp += det_j * compute_inertia_product(pts, 0, 2)
+            
+        # since tensor was calc with density = 1. we say mass = density / volume = 1 / volume
+        ia = ia / volume / 60 - volume * (center_of_mass[1] ** 2 + center_of_mass[2] ** 2)
+        ib = ib / volume / 60 - volume * (center_of_mass[0] ** 2 + center_of_mass[2] ** 2)
+        ic = ic / volume / 60 - volume * (center_of_mass[0] ** 2 + center_of_mass[1] ** 2)
+        iap = iap / volume / 120 - volume * center_of_mass[1] * center_of_mass[2]
+        ibp = ibp / volume / 120 - volume * center_of_mass[0] * center_of_mass[1]
+        icp = icp / volume / 120 - volume * center_of_mass[0] * center_of_mass[2]
+        
+        return glm.mat3x3(
+            ia, -ibp, -icp,
+            -ibp, ib, -iap,
+            -icp, -iap, ic
+        )
+        
+        # return glm.mat3x3(
+        #     ic, -iap, -icp,
+        #     -iap, ib, -ibp,
+        #     -ibp, -iap, ia
+        # )
+        
+    def get_best_triangle(self, point: glm.vec3, vec: glm.vec3) -> int:
+        """
+        Gets the triangle with the closest intersection, -1 if no intersection is found
+        """
+        indices = self.bvh.get_possible_triangles(point, vec)
+        best_distance = -1
+        best_index = -1
+        
+        point = glm.vec3(point)
+        vec = glm.vec3(vec)
+
+        for triangle in indices:
+            
+            # check if triangle intersects
+            intersection = moller_trumbore(point, vec, [self.points[t] for t in self.indices[triangle]])
+            if not intersection: continue
+            
+            # check if triangle is on correct side of line
+            difference = intersection - self.geometric_center
+            if glm.dot(difference, vec) < 0: continue
+            
+            # determine best distance
+            distance = glm.length(difference)
+            if best_distance < 0 or distance < best_distance: 
+                best_distance = distance
+                best_index = triangle
+                
+        return best_index
+        
+    def get_best_dot(self, vec: glm.vec3) -> int:
+        """
+        Gets the point with the highest normalized dot product to the given vector
+        """
+        triangle = self.bvh.get_best_dot(vec)
+        if triangle == -1: return None
+        index = max(self.indices[triangle], key=lambda t: glm.dot(glm.normalize(self.points[t]), vec))
+        return index
+    
+    def get_best_dot_hill_climbing(self, vec: glm.vec3) -> int:
+        """
+        Gets the point with the highest dot product to the given vector using a hill climbing algorithm. This function is only effective for convex models.
+        """
+        best_index = 0
+        best_dot = glm.dot(self.points[best_index], vec)
+        
+        while True:
+            
+            best_changed = False
+            for index in self.edges[best_index]:
+                
+                dot = glm.dot(self.points[index], vec)
+                if dot > best_dot:
+                    best_dot = dot
+                    best_changed = True
+                    best_index = index
+                    
+            if not best_changed: break
+        
+        return best_index
+    
+    def get_line_collided(self, position: glm.vec3, forward: glm.vec3) -> list[tuple[int, int, int]]:
+        """
+        Determines which triangles are intersecting with the given line segment. Returns the indices of the triangle contained in the mesh points list
+        """
+        return self.bvh.get_line_collided(position, forward)
+
+    def __repr__(self) -> str:
+        size = (self.data.nbytes + self.points.nbytes + self.indices.nbytes) / 1024 / 1024
+        return f'<Basilisk Mesh | {len(self.data)} vertices, {size:.2} mb>'
+    
+    @property
+    def top_right(self): return self.bvh.root.top_right
+    @property
+    def bottom_left(self): return self.bvh.root.bottom_left
+    @property
+    def aabb_points(self): 
+        x1, y1, z1 = self.top_right
+        x2, y2, z2 = self.bottom_left
+        return [glm.vec3(x, y, z) for z in (z1, z2) for y in (y1, y2) for x in (x1, x2)]
+
+    def __hash__(self):
         return self.hash