Convert peilbeheerst notebooks to script (#171)

This runs `pixi run jupyter nbconvert --to script
path/to/notebook.ipynb` for all notebooks in `src\peilbeheerst_model\`.
I find it much easier to refactor code when it is all in Python. But
there are also some VSCode notebook bugs plaguing me, for instance
scrambling imports on autoformat when I save them.

I did some minor edits afterwards, e.g. removing `# In[]` comments and
some ipython magic that is no longer needed.

src/peilbeheerst_model/01b_ad_krw_to_peilgebieden.py

@@ -0,0 +1,44 @@
+import pathlib
+
+import geopandas as gpd
+import pandas as pd
+from IPython.core.display import HTML
+
+from peilbeheerst_model import ParseCrossings, waterschap_data
+
+print_df = {}
+for waterschap, waterschap_struct in waterschap_data.items():
+    for funcname, func_args in waterschap_struct.items():
+        if funcname not in print_df:
+            print_df[funcname] = []
+        print_df[funcname].append(pd.Series(func_args, name=waterschap))
+
+for funcname, df in print_df.items():
+    print(HTML(f"<h2>Function {funcname}:</h2>"))
+    print(pd.DataFrame(df))
+
+
+for waterschap, waterschap_struct in waterschap_data.items():
+    print(f"\n{waterschap}...")
+
+    init_settings, crossing_settings = waterschap_struct.values()
+    gpkg = pathlib.Path(init_settings["output_path"])
+    if not gpkg.exists():
+        raise ValueError(gpkg)
+
+    df_peilgebieden = gpd.read_file(gpkg, layer="peilgebied")
+    org_shape = df_peilgebieden.shape
+    df_peilgebieden = ParseCrossings._make_valid_2dgeom(df_peilgebieden)
+
+    df_peilgebieden = ParseCrossings.add_krw_to_peilgebieden(
+        df_peilgebieden,
+        init_settings["krw_path"],
+        init_settings["krw_column_id"],
+        init_settings["krw_column_name"],
+        init_settings["krw_min_overlap"],
+        ",",
+    )
+
+    assert df_peilgebieden.shape[0] == org_shape[0]
+    df_peilgebieden.to_file(gpkg, layer="peilgebied")
+    print(gpkg)