Adding the function to add dataframe as pts with nesting in geohy5.py

.github/workflows/tester.yml

@@ -7,38 +7,49 @@ on:
     paths:
       - '**.py'
       - .github/workflows/tester.yml
-
   pull_request:
     branches:
       - master
     paths:
       - '**.py'
       - .github/workflows/tester.yml
   workflow_dispatch:
+
 jobs:
   continuous-integration:
     name: Continuous integration ${{ matrix.os }} python ${{ matrix.python-version }}
     runs-on: ${{ matrix.os }}
     strategy:
       fail-fast: false
       matrix:
-        os: ${{ fromJSON(vars.BUILD_OS)}}
-        python-version: ${{ fromJSON(vars.PYTHON_VERSIONS)}}
+        os: ${{ fromJSON(vars.BUILD_OS) }}
+        python-version: ${{ fromJSON(vars.PYTHON_VERSIONS) }}
+
     steps:
       - uses: actions/checkout@v4
-      - uses: conda-incubator/setup-miniconda@v3
 
+      - name: Set up uv
+        uses: astral-sh/setup-uv@v3
         with:
-          python-version: ${{ matrix.python }}
-      - name: Installing dependencies
-        shell: bash -l {0}
-        run: |
-          conda install -c conda-forge numpy scipy scikit-image scikit-learn pyvista pandas pytest networkx osqp matplotlib -y
-      - name: Building and install
-        shell: bash -l {0}
+          version: "latest"
+
+      - name: Set up Python ${{ matrix.matrix.python-version }}
+        run: uv python install ${{ matrix.python-version }}
+
+      - name: Install dependencies and library
         run: |
-          pip install . --user
+          uv pip install --system \
+            numpy \
+            scipy \
+            scikit-image \
+            scikit-learn \
+            pytest \
+            networkx \
+            osqp \
+            matplotlib \
+            geoh5py \
+            -e .[tests]
+
       - name: pytest
-        shell: bash -l {0}
         run: |
           pytest

LoopStructural/datatypes/_point.py

@@ -14,7 +14,18 @@ class ValuePoints:
     values: np.ndarray = field(default_factory=lambda: np.array([0]))
     name: str = "unnamed"
     properties: Optional[dict] = None
-
+    def __post_init__(self):
+
+        self.values = np.asarray(self.values)
+        self.locations = np.asarray(self.locations)
+        if self.locations.shape[1] != 3:
+            raise ValueError('locations must be of shape (n, 3)')
+        if len(self.values) != len(self.locations):
+            raise ValueError('values must be the same length as locations')
+        for k, v in (self.properties or {}).items():
+            if len(v) != len(self.locations):
+                raise ValueError(f'Property {k} must be the same length as locations')
+            self.properties[k] = np.asarray(v)
     def to_dict(self):
         return {
             "locations": self.locations,
@@ -112,7 +123,17 @@ class VectorPoints:
     vectors: np.ndarray = field(default_factory=lambda: np.array([[0, 0, 0]]))
     name: str = "unnamed"
     properties: Optional[dict] = None
-
+    def __post_init__(self):
+        self.vectors = np.asarray(self.vectors)
+        self.locations = np.asarray(self.locations)
+        if self.locations.shape[1] != 3:
+            raise ValueError('locations must be of shape (n, 3)')
+        if len(self.vectors) != len(self.locations):
+            raise ValueError('vectors must be the same length as locations')
+        for k, v in (self.properties or {}).items():
+            if len(v) != len(self.locations):
+                raise ValueError(f'Property {k} must be the same length as locations')
+            self.properties[k] = np.asarray(v)
     def to_dict(self):
         return {
             "locations": self.locations,

LoopStructural/export/geoh5.py

@@ -1,16 +1,33 @@
 import geoh5py
 import geoh5py.workspace
 import numpy as np
-from LoopStructural.datatypes import ValuePoints, VectorPoints
+import pandas as pd
 
+from LoopStructural.datatypes import ValuePoints, VectorPoints
 
-def add_surface_to_geoh5(filename, surface, overwrite=True, groupname="Loop"):
+def add_group_to_geoh5(filename, groupname="Loop", parent=None, overwrite=True):
     with geoh5py.workspace.Workspace(filename) as workspace:
+
         group = workspace.get_entity(groupname)[0]
+        if group and overwrite:
+            group.allow_delete = True
+            workspace.remove_entity(group)
         if not group:
             group = geoh5py.groups.ContainerGroup.create(
                 workspace, name=groupname, allow_delete=True
             )
+        if parent is not None:
+            parent = workspace.get_entity(parent)[0]    
+        if parent:
+            parent.add_children(group)
+        return group.uid
+def add_surface_to_geoh5(filename, surface, overwrite=True, group="Loop"):
+    with geoh5py.workspace.Workspace(filename) as workspace:
+        group = workspace.get_entity(group)[0]
+        if not group:
+            group = geoh5py.groups.ContainerGroup.create(
+                workspace, name=group, allow_delete=True
+            )
         if surface.name in workspace.list_entities_name.values():
             existing_surf = workspace.get_entity(surface.name)
             existing_surf[0].allow_delete = True
@@ -32,6 +49,7 @@ def add_surface_to_geoh5(filename, surface, overwrite=True, groupname="Loop"):
 
 def add_points_to_geoh5(filename, point, overwrite=True, groupname="Loop"):
     with geoh5py.workspace.Workspace(filename) as workspace:
+
         group = workspace.get_entity(groupname)[0]
         if not group:
             group = geoh5py.groups.ContainerGroup.create(
@@ -52,15 +70,73 @@ def add_points_to_geoh5(filename, point, overwrite=True, groupname="Loop"):
             data['vz'] = {'association': "VERTEX", "values": point.vectors[:, 2]}
 
         if isinstance(point, ValuePoints):
-            data['val'] = {'association': "VERTEX", "values": point.values}
+            data['values'] = {'association': "VERTEX", "values": point.values}
         point = geoh5py.objects.Points.create(
             workspace,
             name=point.name,
             vertices=point.locations,
             parent=group,
         )
         point.add_data(data)
+
+def overwrite_object(workspace, name, overwrite):
+    if name in workspace.list_entities_name.values():
+        existing_entity = workspace.get_entity(name)
+        existing_entity[0].allow_delete = True
+        if overwrite:
+            workspace.remove_entity(existing_entity[0])
 
+def add_points_from_df(filename, df, name='pointset', overwrite=True,                        columns=None, groupname="Loop", x_col='X', y_col='Y', z_col='Z'):
+        """
+        Add points to a geoh5 file from a pandas DataFrame. The DataFrame must have columns 'name', 'X', 'Y', 'Z' for the point locations.
+        Additional columns can be added as data associated with the points.
+        Parameters
+        ----------
+        filename: str
+            Path to the geoh5 file.
+        df: pandas.DataFrame
+            DataFrame containing point data. Must have columns 'name', 'X', 'Y', 'Z'. Additional columns will be added as data.
+        overwrite: bool, optional
+            Whether to overwrite existing points with the same name. Default is True.
+        columns: list of str, optional
+            List of columns in the DataFrame to add as data. If None, all columns except 'name', 'X', 'Y', 'Z' will be added. Default is None.
+
+        """
+        if columns is None:
+            columns = df.columns.tolist()
+        if x_col not in columns or y_col not in columns or z_col not in columns:
+            raise ValueError("DataFrame must contain 'name', 'X', 'Y', 'Z' columns. " \
+            "Specify the column names using x_col, y_col, z_col parameters if they are different.")
+        with geoh5py.workspace.Workspace(filename) as workspace:
+            if groupname:
+                group = workspace.get_entity(groupname)
+                group = group[0] if group else None
+                if not group:
+                    group = geoh5py.groups.ContainerGroup.create(
+                        workspace, name=groupname, allow_delete=True,
+                    )
+
+            location = np.array(df[[x_col, y_col, z_col]].values)  # shape (n,3)
+
+            overwrite_object(workspace, name, overwrite)
+
+
+            pts = geoh5py.objects.Points.create(
+                workspace,
+                name=name,
+                vertices=location,
+                parent=group,
+            )
+            data = {}
+            for col in columns:
+                if col in ['name', x_col, y_col, z_col]:
+                    continue
+                data[col] = {"association": "VERTEX", "values": np.array(df[col]).flatten()}
+
+
+            if data:
+                pts.add_data(data)
+
 
 def add_structured_grid_to_geoh5(filename, structured_grid, overwrite=True, groupname="Loop"):
     with geoh5py.workspace.Workspace(filename) as workspace:

LoopStructural/utils/__init__.py

@@ -26,6 +26,7 @@
     plungeazimuth2vector,
     azimuthplunge2vector,
     normal_vector_to_strike_and_dip,
+    normal_vector_to_dip_and_dip_direction,
     rotate,
 )
 from .helper import create_surface, create_box

pyproject.toml

@@ -46,6 +46,7 @@ visualisation = ["matplotlib", "pyvista", "loopstructuralvisualisation>=0.1.14"]
 export = ["geoh5py", "pyevtk", "dill"]
 jupyter = ["pyvista[all]"]
 inequalities = ["loopsolver"]
+tests = ['pytest','all']
 docs = [
     "pyvista[all]",
     "pydata-sphinx-theme",

tests/unit/io/test_geoh5.py

@@ -0,0 +1,90 @@
+from LoopStructural.export.geoh5 import add_group_to_geoh5, add_points_to_geoh5, add_points_from_df
+import geoh5py
+import pytest
+from pathlib import Path
+from LoopStructural.datatypes import ValuePoints, VectorPoints
+import numpy as np
+@pytest.fixture
+def tmp_path():
+    import tempfile
+    with tempfile.TemporaryDirectory() as tmpdir:
+        yield Path(tmpdir)
+@pytest.fixture
+def test_setup(tmp_path):
+    filename = tmp_path / "test.geoh5"
+    with geoh5py.workspace.Workspace.create(filename) as workspace:
+        yield filename
+        workspace.close()
+
+def test_add_group_to_geoh5(test_setup):
+    filename = test_setup
+    group_uid = add_group_to_geoh5(filename, groupname="TestGroup")
+
+    with geoh5py.workspace.Workspace(filename) as workspace:
+        assert workspace.get_entity(group_uid)[0].name == "TestGroup"
+
+def test_add_points_to_geoh5(test_setup):
+    filename = test_setup
+    group_uid = add_group_to_geoh5(filename, groupname="TestGroup")
+    points = ValuePoints(
+        name="TestPoints",
+        locations=[[0, 0, 0], [1, 1, 1], [2, 2, 2]],
+        values=[10., 20, 30],
+    )
+    add_points_to_geoh5(filename, points, groupname=group_uid)
+    with geoh5py.workspace.Workspace(filename) as workspace:
+        point_entity = workspace.get_entity("TestPoints")[0]
+        assert point_entity.name == "TestPoints"
+        assert point_entity.vertices.tolist() == [[0, 0, 0], [1, 1, 1], [2, 2, 2]]
+        assert np.sum(point_entity.get_data("values")[0].values-np.array([10., 20., 30.])) == 0
+
+def test_add_vector_points_to_geoh5(test_setup):
+    filename = test_setup
+    group_uid = add_group_to_geoh5(filename, groupname="TestGroup")
+    points = VectorPoints(
+        name="TestVectorPoints",
+        locations=[[0, 0, 0], [1, 1, 1], [2, 2, 2]],
+        vectors=[[1, 0, 0], [0, 1, 0], [0, 0, 1]],
+    )
+    add_points_to_geoh5(filename, points, groupname=group_uid)
+    with geoh5py.workspace.Workspace(filename) as workspace:
+        point_entity = workspace.get_entity("TestVectorPoints")[0]
+        assert point_entity.name == "TestVectorPoints"
+        assert point_entity.vertices.tolist() == [[0, 0, 0], [1, 1, 1], [2, 2, 2]]
+        assert np.sum(point_entity.get_data("vx")[0].values-np.array([1., 0., 0.])) == 0
+        assert np.sum(point_entity.get_data("vy")[0].values-np.array([0., 1., 0.])) == 0
+        assert np.sum(point_entity.get_data("vz")[0].values-np.array([0., 0., 1.])) == 0
+
+def test_add_df_to_geoh5(test_setup):
+    import pandas as pd
+    filename = test_setup
+    group_uid = add_group_to_geoh5(filename, groupname="TestGroup")
+    df = pd.DataFrame({
+        'X': [0, 1, 2],
+        'Y': [0, 1, 2],
+        'Z': [0, 1, 2],
+        'value': [10., 20., 30.],
+    })
+    add_points_from_df(filename, df, name='df_points', groupname=group_uid)
+    with geoh5py.workspace.Workspace(filename) as workspace:
+        point_entity = workspace.get_entity("TestGroup")[0].children[0]
+        assert point_entity.name == "df_points"
+        assert point_entity.vertices.tolist() == [[0, 0, 0], [1, 1, 1], [2, 2, 2]]
+        assert np.sum(point_entity.get_data("value")[0].values-np.array([10., 20., 30.])) == 0
+
+def test_add_df_with_alternate_xyz_to_geoh5(test_setup):
+    import pandas as pd
+    filename = test_setup
+    group_uid = add_group_to_geoh5(filename, groupname="TestGroup")
+    df = pd.DataFrame({
+        'EAST': [0, 1, 2],
+        'NORTH': [0, 1, 2],
+        'RL': [0, 1, 2],
+        'value': [10., 20., 30.],
+    })
+    add_points_from_df(filename, df, name='df_points',groupname=group_uid, x_col='EAST', y_col='NORTH', z_col='RL')
+    with geoh5py.workspace.Workspace(filename) as workspace:
+        point_entity = workspace.get_entity("TestGroup")[0].children[0]
+        assert point_entity.name == "df_points"
+        assert point_entity.vertices.tolist() == [[0, 0, 0], [1, 1, 1], [2, 2, 2]]
+        assert np.sum(point_entity.get_data("value")[0].values-np.array([10., 20., 30.])) == 0