Merge pull request #22 from sigma-epsilon/17-customizable-cell-interp…

…olators

17 customizable cell interpolators

README.md

@@ -72,6 +72,10 @@ If you are a developer and want to install the library in development mode, the
 >>> pip install "-e .[test, dev]"
 ```
 
+## Versioning
+
+The project follows the rules of [semantic versioning](https://semver.org/).
+
 ## How to contribute?
 
 Contributions are currently expected in any the following ways:

docs/source/notebooks/interpolation.ipynb → docs/source/notebooks/approximation.ipynb

@@ -49,10 +49,10 @@
     }
    ],
    "source": [
-    "from sigmaepsilon.mesh.cells import LagrangianCellInterpolator\n",
+    "from sigmaepsilon.mesh.cells import LagrangianCellApproximator\n",
     "\n",
-    "interpolator = H8.interpolator()\n",
-    "isinstance(interpolator, LagrangianCellInterpolator)"
+    "approximator = H8.approximator()\n",
+    "isinstance(approximator, LagrangianCellApproximator)"
    ]
   },
   {
@@ -83,14 +83,14 @@
     "source_values = [1, 2, 3, 4, 5, 6, 7, 8]\n",
     "target_coordinates = master_coordinates\n",
     "\n",
-    "interpolator(source=source_coordinates, values=source_values, target=target_coordinates)"
+    "approximator(source=source_coordinates, values=source_values, target=target_coordinates)"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "It is possible to pass the source coordinates to the factory function. This is useful if we plan to reuse the interpolator with the same source points and can save a little time. In this case only the source values and the target points need to be provided."
+    "It is possible to pass the source coordinates to the factory function. This is useful if we plan to reuse the approximator with the same source points and can save a little time. In this case only the source values and the target points need to be provided."
    ]
   },
   {
@@ -114,15 +114,15 @@
     "source_values = [1, 2, 3, 4, 5, 6, 7, 8]\n",
     "target_coordinates = master_coordinates\n",
     "\n",
-    "interpolator = H8.interpolator(source_coordinates)\n",
-    "interpolator(values=source_values, target=target_coordinates)"
+    "approximator = H8.approximator(source_coordinates)\n",
+    "approximator(values=source_values, target=target_coordinates)"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "As noted in the documentation, if the number of source coorindates does not match the number of nodes (and hence the number of shape functions) of the master element of the class, the interpolation is gonna be under or overdetermined and the operation involves the calculation of a generalized inverse. This means, that for instance feeding a 4-noded quadrilateral with 9 source points and data values is more information than what the class is normally able to precisely handle and the resulting interpolator will represent a fitting function. In that case, if you want a precise interpolation, you would want to use a 9-node quadrilateral, or accept the loss of information."
+    "As noted in the documentation, if the number of source coorindates does not match the number of nodes (and hence the number of shape functions) of the master element of the class, the approximation is gonna be under or overdetermined and the operation involves the calculation of a generalized inverse. This means, that for instance feeding a 4-noded quadrilateral with 9 source points and data values is more information than what the class is normally able to precisely handle and the resulting approximator will represent a fitting function. In that case, if you want a precise approximation, you would want to use a 9-node quadrilateral, or accept the loss of information."
    ]
   },
   {
@@ -150,8 +150,8 @@
     "source_values = [i + 1 for i in range(9)]\n",
     "target_coordinates = master_coordinates\n",
     "\n",
-    "interpolator = Q4.interpolator()\n",
-    "interpolator(source=source_coordinates, values=source_values, target=target_coordinates)"
+    "approximator = Q4.approximator()\n",
+    "approximator(source=source_coordinates, values=source_values, target=target_coordinates)"
    ]
   },
   {
@@ -178,8 +178,8 @@
     }
    ],
    "source": [
-    "interpolator = Q9.interpolator()\n",
-    "interpolator(source=source_coordinates, values=source_values, target=target_coordinates)"
+    "approximator = Q9.approximator()\n",
+    "approximator(source=source_coordinates, values=source_values, target=target_coordinates)"
    ]
   },
   {
@@ -213,8 +213,8 @@
     "source_values = [i + 1 for i in range(3)]\n",
     "target_coordinates = master_coordinates\n",
     "\n",
-    "interpolator = L3.interpolator()\n",
-    "interpolator(source=source_coordinates, values=source_values, target=target_coordinates)"
+    "approximator = L3.approximator()\n",
+    "approximator(source=source_coordinates, values=source_values, target=target_coordinates)"
    ]
   },
   {
@@ -228,7 +228,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "Set up an interpolator:"
+    "Set up an approximator:"
    ]
   },
   {
@@ -239,7 +239,7 @@
    "source": [
     "import numpy as np\n",
     "\n",
-    "interpolator = H8.interpolator()\n",
+    "approximator = H8.approximator()\n",
     "master_coordinates = H8.master_coordinates()\n",
     "\n",
     "source_coordinates = master_coordinates / 2\n",
@@ -271,7 +271,7 @@
    ],
    "source": [
     "source_values = np.random.rand(10, 2, 8)\n",
-    "interpolator(\n",
+    "approximator(\n",
     "    source=source_coordinates, \n",
     "    values=source_values, \n",
     "    target=target_coordinates[:3]\n",
@@ -303,7 +303,7 @@
    ],
    "source": [
     "source_values = np.random.rand(8, 2, 10)\n",
-    "interpolator(\n",
+    "approximator(\n",
     "    source=source_coordinates, \n",
     "    values=source_values, \n",
     "    target=target_coordinates[:3],\n",
@@ -317,6 +317,50 @@
    "source": [
     "The workaround here is to use `numpy.ascontiguousarray` and reordering the input data."
    ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Custom cells"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Getting an approximator for a custom cell goes the same way, after the cell have been properly set up."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "array([-0.5,  1.5,  3.5,  5.5])"
+      ]
+     },
+     "execution_count": 11,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "from sigmaepsilon.mesh.cells.base import PolyCell1d\n",
+    "\n",
+    "Custom1dCell: PolyCell1d = PolyCell1d.generate_class(NNODE=4)\n",
+    "\n",
+    "master_coordinates = Custom1dCell.master_coordinates()\n",
+    "source_coordinates = master_coordinates / 2\n",
+    "source_values = [i + 1 for i in range(Custom1dCell.NNODE)]\n",
+    "target_coordinates = master_coordinates\n",
+    "\n",
+    "approximator = Custom1dCell.approximator()\n",
+    "approximator(source=source_coordinates, values=source_values, target=target_coordinates)"
+   ]
   }
  ],
  "metadata": {

pyproject.toml

@@ -8,7 +8,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "sigmaepsilon.mesh"
-version = "1.0.0"
+version = "1.1.0"
 description = "A Python package to build, manipulate and analyze polygonal meshes."
 classifiers=[
         "Development Status :: 5 - Production/Stable",
@@ -43,7 +43,6 @@ where = ["src"]
 dependencies = { file = ["requirements.txt"] }
 optional-dependencies.test = { file = ["requirements-test.txt"] }
 optional-dependencies.dev = { file = ["requirements-dev.txt"] }
-optional-dependencies.devops = { file = ["requirements-devops.txt"] }
 optional-dependencies.docs = { file = ["docs/requirements.txt"] }
 
 [project.urls]

src/sigmaepsilon/mesh/celldata.py

@@ -1,4 +1,4 @@
-from typing import Union
+from typing import Union, Iterable
 from copy import copy, deepcopy
 from functools import partial
 
@@ -51,11 +51,7 @@ class CellData(CellDataBase):
         For every key and value pair where the value is a numpy array
         with a matching shape (has entries for all cells), the key
         is considered as a field and the value is added to the database.
-
-    See Also
-    --------
-    :class:`awkward.Array`
-    :class:`awkward.Record`
+        
     """
 
     _attr_map_ = {
@@ -290,7 +286,7 @@ def __getattr__(self, attr):
             name = self.__class__.__name__
             raise AttributeError(f"'{name}' object has no attribute called {attr}")
 
-    def set_nodal_distribution_factors(self, factors: ndarray, key: str = None):
+    def set_nodal_distribution_factors(self, factors: ndarray, key: str = None) -> None:
         """
         Sets nodal distribution factors.
 
@@ -350,7 +346,7 @@ def pull(
         return d
 
     @property
-    def fields(self):
+    def fields(self) -> Iterable[str]:
         """Returns the fields in the database object."""
         return self._wrapped.fields
 
@@ -400,7 +396,7 @@ def frames(self, value: ndarray):
         self._wrapped[self._dbkey_frames_] = frames
 
     @property
-    def t(self):
+    def t(self) -> ndarray:
         """Returns the thicknesses of the cells."""
         return self._wrapped[self._dbkey_thickness_].to_numpy()
 
@@ -412,7 +408,7 @@ def t(self, value: Union[float, int, ndarray]):
         self._wrapped[self._dbkey_thickness_] = value
 
     @property
-    def A(self):
+    def A(self) -> ndarray:
         """Returns the thicknesses of the cells."""
         return self._wrapped[self._dbkey_areas_].to_numpy()
 

src/sigmaepsilon/mesh/cells/__init__.py

@@ -19,7 +19,7 @@
 from .w6 import W6 as Wedge
 from .w18 import W18
 
-from .base.interpolator import LagrangianCellInterpolator
+from .base.approximator import LagrangianCellApproximator
 
 __all__ = [
     "L2",
@@ -41,5 +41,5 @@
     "W6",
     "Wedge",
     "W18",
-    "LagrangianCellInterpolator"
+    "LagrangianCellApproximator"
 ]

src/sigmaepsilon/mesh/cells/base/__init__.py

@@ -0,0 +1,9 @@
+from .cell1d import PolyCell1d
+from .cell2d import PolyCell2d
+from .cell3d import PolyCell3d
+
+__all__ = [
+    "PolyCell1d",
+    "PolyCell2d",
+    "PolyCell3d",
+]