Numpy array raises error when used in create_junctions function specifically in geodata attribute

[oakca]

Describe the bug
When a numpy array of (x, y) tuples are used as geodata in create_junctions() function, an error is raised. When converted the same array to a list then it works.

here is the nodes file: https://congasnet-my.sharepoint.com/:u:/g/personal/okan_akca_congas_net/ET8b6UY-W9pIurtbPBMkT94BLqnS27GoyyHx15GbflWqsQ?e=qTBPBt

To Reproduce

import geopandas as gpd
import pandapipes as pp

n_ = gpd.read_file('nodes.geojson', driver='GeoJSON').set_index('id')

net = pp.create_empty_network(fluid='hydrogen')
nodes = pp.create_junctions(
        net,
        nr_junctions=len(n_),
        pn_bar=50,
        tfluid_k=283.15,
        height_m=n_['height'].values,
        name=n_['name'].values,
        index=n_.index,
        geodata=n_['geometry'].map(lambda p: (p.x, p.y)).values)

Error message

Traceback (most recent call last):
  File "C:\.venv\Lib\site-packages\IPython\core\interactiveshell.py", line 3577, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-18-cb2062836528>", line 3, in <module>
    nodes = pp.create_junctions(
            ^^^^^^^^^^^^^^^^^^^^
  File "C:\.venv\Lib\site-packages\pandapipes\create.py", line 1135, in create_junctions
    net.junction_geodata.loc[index, ["x", "y"]] = geodata
    ~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^
  File "C:\.venv\Lib\site-packages\pandas\core\indexing.py", line 885, in __setitem__
    iloc._setitem_with_indexer(indexer, value, self.name)
  File "C:\.venv\Lib\site-packages\pandas\core\indexing.py", line 1893, in _setitem_with_indexer
    self._setitem_with_indexer_split_path(indexer, value, name)
  File "C:\.venv\Lib\site-packages\pandas\core\indexing.py", line 1978, in _setitem_with_indexer_split_path
    raise ValueError(
ValueError: Must have equal len keys and value when setting with an iterable

Expected behavior
it works when numpy array changed to list via n_['geometry'].map(lambda p: (p.x, p.y)).values.tolist(), it should also work with numpy array since the other attributes raise no errors.

Python environment (please complete the following information):

• OS: Windows 11
• pandapipes 0.11.0
• geopandas

[dad616610]

I believe this is not a bug with pandapipes, but rather a shape mismatch in the input data.. Your data is not in the form of

(x, y) tuples

- its shape is (48, ), whereas pandapipes expects (48, 2)

>>> n_["geometry"].shape
(48,)

Your solution with .tolist() works because it explicitly converts each Point into a tuple of two values, resulting in a (48, 2) shape:

>>> np.array(n_["geometry"].map(lambda p: (p.x, p.y)).values.tolist()).shape
(48, 2)

However, a more direct approach is to use .get_coordinates(), which is built into geopandas:

>>> n_["geometry"].get_coordinates().shape
(48, 2)

See the GeoSeries.get_coordinates() documentation for more details.