Faulty DWD stations metadata when name contains comma

[joshuarrrrr]

Describe the bug When fetching the stations metadata for DWD observations, some stations are missing the state value and part of the name. This seems to affect stations with a comma in their name due to a prefix being moved to the end of the name, for example "Harzburg, Bad". This causes the station name to be cut off at the comma, so it would become "Harzburg", and the state name to be missing completely.

To Reproduce Using version 0.78.0 with the following code:

import polars as pl
from wetterdienst.provider.dwd.observation import DwdObservationRequest

request = DwdObservationRequest(
    parameter="kl",
    resolution="monthly",
    period="recent",
)
stations = request.all()
stations.df.filter(pl.col("state").is_null())

It results in 30 stations with a cut off name and null state value.

Expected behavior The full station name and state as present in the DWD descriptions txt file.

Desktop (please complete the following information):

• OS: Linux
• Python-Version 3.12

Additional context I loaded the stations using pandas for comparison with the following code:

stations.df.join(
            pl.from_pandas(
                pd.read_fwf(
                    "https://opendata.dwd.de/climate_environment/CDC/observations_germany/climate/monthly/kl/recent/KL_Monatswerte_Beschreibung_Stationen.txt",
                    encoding="latin-1",
                    skiprows=[0, 1],
                    names=stations.df.columns,
                    dtype={"station_id": "string"},
                )
            ).select(pl.col("station_id", "name", "state")),
            on="station_id",
        )
        .filter(pl.col("state").is_null())
        .select(["station_id", "name", "state", "name_right", "state_right"])

This shows which stations are affected by the bug and what the name and state values should be:	station_id	name	state	name_right	state_right
00314	Kubschütz	null	Kubschütz, Kr. Bautzen	Sachsen
00377	Bergzabern	null	Bergzabern, Bad	Rheinland-Pfalz
00379	Berka	null	Berka, Bad (Flugplatz)	Thüringen
00390	Berleburg	null	Berleburg, Bad-Stünzel	Nordrhein-Westfalen
00755	Buchen	null	Buchen, Kr. Neckar-Odenwald	Baden-Württemberg
01072	Dürkheim	null	Dürkheim, Bad	Rheinland-Pfalz
01207	Elster	null	Elster, Bad-Sohl	Sachsen
01332	Falkenberg	null	Falkenberg,Kr.Rottal-Inn	Bayern
02039	Harzburg	null	Harzburg, Bad	Niedersachsen
02171	Hersfeld	null	Hersfeld, Bad	Hessen
02323	Bevern	null	Bevern, Kr. Holzminden	Niedersachsen
02597	Kissingen	null	Kissingen, Bad	Bayern
02680	Königshofen	null	Königshofen, Bad	Bayern
02708	Kohlgrub	null	Kohlgrub, Bad (Rosshof)	Bayern
02878	Lauchstädt	null	Lauchstädt, Bad	Sachsen-Anhalt
03028	Lippspringe	null	Lippspringe, Bad	Nordrhein-Westfalen
03034	Lobenstein	null	Lobenstein, Bad	Thüringen
03164	Cölbe	null	Cölbe, Kr. Marburg-Biedenkopf	Hessen
03167	Marienberg	null	Marienberg, Bad	Rheinland-Pfalz
03257	Mergentheim	null	Mergentheim, Bad	Baden-Württemberg
03426	Muskau	null	Muskau, Bad	Sachsen
03442	Nauheim	null	Nauheim, Bad	Hessen
03490	Neuenahr	null	Neuenahr, Bad-Ahrweiler	Rheinland-Pfalz
04094	Weingarten	null	Weingarten, Kr. Ravensburg	Baden-Württemberg
04189	Altheim	null	Altheim, Kreis Biberach	Baden-Württemberg
04301	Kreuznach	null	Kreuznach, Bad	Rheinland-Pfalz
04371	Salzuflen	null	Salzuflen, Bad	Nordrhein-Westfalen
04813	Staffelstein	null	Staffelstein, Bad-Stublang	Bayern
04841	Steinau	null	Steinau, Kr. Cuxhaven	Niedersachsen
06272	Salzungen	null	Salzungen, Bad-Gräfen-Nitzend	Thüringen

[gutzbenj]

Dear @joshuarrrrr ,

thanks for reporting this issue! You're absolutely right! I mistakenly used read_csv here in combination with column specs not thinking of actual commata in the data... With the latest commit I changed that and the names should now be correct.

[joshuarrrrr]

Thanks for the swift fix! I can confirm it's now working perfectly as expected in v0.79.0.