Merge pull request #39 from bedrock-engineer/dev

Dev

Author: JoostGevaert

RELEASE.md

@@ -11,7 +11,7 @@ Follow [Semantic Versioning](https://semver.org/) (e.g., `1.0.0`, `1.1.0`, `1.1.
 Update the version number in:
 
 - [`pyproject.toml`](pyproject.toml)
-- [`/src/bedrock/__init__.py`](/src/bedrock/__init__.py)
+- [`/src/bedrock/__init__.py`](/src/bedrock_ge/__init__.py)
 - Inline script dependencies of marimo notebooks in [`examples`](/examples/)
 
 ## 2. Update the Changelog

examples/hk_kaitak_ags3/kaitak_gi.gpkg

@@ -1,6 +1,6 @@
 [project]
 name = "bedrock-ge"
-version = "0.2.3"
+version = "0.2.4"
 description = "Bedrock's Python library for geotechnical engineering."
 authors = [
     {name = "Bedrock", email = "info@bedrock.engineer"}

pyproject.toml

@@ -1,3 +1,3 @@
 """Bedrock, the open source foundation for ground engineering."""
 
-__version__ = "0.2.3"
+__version__ = "0.2.4"

src/bedrock_ge/__init__.py

@@ -1,12 +1,25 @@
 import os
 import shutil
+import sqlite3
 import subprocess
-import sys
 from pathlib import Path
 from tempfile import TemporaryDirectory
 
+import geopandas as gpd
+import pandas as pd
+
 
 def test_kaitak_ags3_notebook_runs_and_creates_gpkg(examples_dir):
+    """Tests the Kai Tak, Hong Kong AGS 3 example marimo notebook.
+
+    Tests that the `hk_kaitak_ags3_to_brgi_geodb.py` marimo notebook:
+    - Runs successfully as a script using `uvx uv run` with the Python version and
+      dependencies specified in the PEP 723 inline script metadata.
+    - Creates a valid GeoPackage
+    - That the GeoPackage contains the expected tables
+    - That the Project, Location, Sample, InSitu_GEOL, InSitu_ISPT and InSitu_WETH
+      tables have the expected number of rows.
+    """
     notebook_dir = examples_dir / "hk_kaitak_ags3"
     notebook_path = notebook_dir / "hk_kaitak_ags3_to_brgi_geodb.py"
     gpkg_output_path = notebook_dir / "kaitak_gi.gpkg"
@@ -19,14 +32,15 @@ def test_kaitak_ags3_notebook_runs_and_creates_gpkg(examples_dir):
     # to the one created when executing the notebook.
     # And to put back to the original state at the end of the test.
     with TemporaryDirectory() as temp_dir:
-        temp_gpkg_path = Path(temp_dir) / "temp_kaitak_gi.gpkg"
-        shutil.move(gpkg_output_path, temp_gpkg_path)
+        temp_original_gpkg_path = Path(temp_dir) / "temp_kaitak_gi.gpkg"
+        shutil.move(gpkg_output_path, temp_original_gpkg_path)
 
-        print(f"Running: `python {notebook_path}`\n")
         # Run the notebook as a script
         # TODO: implement logging
         # NOTE: The env (environment variables) and encoding are required for running
-        # the notebook as a script from both Windows and Linux. Wihtout: UnicodeDecodeError
+        # the notebook as a script from both Windows and Linux. Without => UnicodeDecodeError
+        # NOTE: `uvx uv run` runs the marimo notebook as a script in a temporary environment,
+        # with the Python version and dependencies specified in the PEP 723 inline script metadata.
         env = os.environ.copy()
         env["PYTHONIOENCODING"] = "utf-8"
         result = subprocess.run(
@@ -40,7 +54,7 @@ def test_kaitak_ags3_notebook_runs_and_creates_gpkg(examples_dir):
 
         # Check that the script ran successfully
         assert result.returncode == 0, (
-            f"\n📛 Running `uvx run marimo notebook.py` failed with code {result.returncode}\n"
+            f"📛 Running `uvx run marimo notebook.py` failed with code {result.returncode}\n"
             f"📄 STDOUT:\n{result.stdout}\n"
             f"⚠️ STDERR:\n{result.stderr}"
         )
@@ -50,10 +64,76 @@ def test_kaitak_ags3_notebook_runs_and_creates_gpkg(examples_dir):
             f"The expected GeoPackage {gpkg_output_path} was not created."
         )
 
-        # TODO: write some logic to compare the original and new GeoPackages.
+        # Compare the original and new GeoPackages and check the number of rows
+        # in the important tables.
+        conn_original = sqlite3.connect(temp_original_gpkg_path)
+        conn_output = sqlite3.connect(gpkg_output_path)
+
+        tables_original = conn_original.execute(
+            "SELECT name FROM sqlite_master WHERE type='table';"
+        ).fetchall()
+        conn_original.close()
+        tables_output = conn_output.execute(
+            "SELECT name FROM sqlite_master WHERE type='table';"
+        ).fetchall()
+        conn_output.close()
+
+        assert tables_original == tables_output, (
+            f"The original GeoPackage {temp_original_gpkg_path.name} and the output "
+            f"GeoPackage {gpkg_output_path.name} have different tables:\n"
+            f"Original: {tables_original}\n"
+            f"Output: {tables_output}"
+        )
+
+        important_tables = [
+            {
+                "table_name": "Project",
+                "no_rows": 88,
+            },
+            {
+                "table_name": "Location",
+                "no_rows": 754,
+            },
+            {
+                "table_name": "Sample",
+                "no_rows": 17_774,
+            },
+            {
+                "table_name": "InSitu_GEOL",
+                "no_rows": 7_764,
+            },
+            {
+                "table_name": "InSitu_ISPT",
+                "no_rows": 3_986,
+            },
+            {
+                "table_name": "InSitu_WETH",
+                "no_rows": 3_928,
+            },
+        ]
+        for table in important_tables:
+            gdf_output = gpd.read_file(gpkg_output_path, layer=table["table_name"])
+            assert len(gdf_output) == table["no_rows"], (
+                f"The output GeoPackage {gpkg_output_path.name} table {table['table_name']} "
+                f"has {len(gdf_output)} rows instead of {table['no_rows']}."
+            )
+            gdf_original = gpd.read_file(
+                temp_original_gpkg_path, layer=table["table_name"]
+            )
+            pd.testing.assert_frame_equal(
+                gdf_original, gdf_output, check_exact=False, rtol=1e-5
+            )
+            # It's also possible to assert that GIS geometries are not exactly equal.
+            # However, when testing the equality of GeoDataFrames with pandas, the GIS
+            # geometry are compared precisely, because the geometry is converted to a
+            # WKT string and compared as strings. Therefore, if a less precise comparison
+            # of GIS geometries is necessary, the assertion above needs changing too.
+            # gpd.testing.assert_geoseries_equal(
+            #     gdf_original, gdf_output, check_less_precise=False
+            # )
 
         # Remove the newly generated kaitak_gi.gpkg
         os.remove(gpkg_output_path)
         # Place back the original kaitak_gi.gpkg from the temporary directory
         # to its original location.
-        shutil.move(temp_gpkg_path, gpkg_output_path)
+        shutil.move(temp_original_gpkg_path, gpkg_output_path)