Pyrosm likely incompatible with newer Pandas (3.0+)

[pickpj]

Pyrosm uses a chained assignment in the OSM() function. This is deprecated in current 3.0+ versions of pandas. Here is the following warning from older pandas:

FutureWarning: ChainedAssignmentError: behaviour will change in pandas 3.0!  
You are setting values through chained assignment. Currently this works in certain cases, but when using Copy-on-Write (which will become the default behaviour in pandas 3.0) this will never work to update the original DataFrame or Series, because the intermediate object on which we are setting values will behave as a copy.  
A typical example is when you are setting values in a column of a DataFrame, like:  

df["col"][row_indexer] = value  

Use `df.loc[row_indexer, "col"] = values` instead, to perform the assignment in a single step and ensure this keeps updating the original `df`.

[pickpj]

It might be better to just roads = gpd.read_file(fileroad, layer="lines")

[pickpj]

Further testing results:

roads = OSM(fileroad) 
roads = roads.get_network("driving+service")

Produced the same size dataframes in pandas 3.0+ and <3.0
roads.highway.value_counts():

highway
service           410406
residential       123506
tertiary           18398
secondary          17783
primary            13441
trunk               6387
unclassified        5023
motorway_link       4388
motorway            2581
primary_link        2010
secondary_link      1706
trunk_link          1389
tertiary_link       1004
living_street        118
track                 54
cycleway              33
path                  17
footway               17
raceway               16
road                  14
services              11
pedestrian             8
construction           7
disused                3
busway                 2
rest_area              1
bridleway              1
ramp                   1
Name: count, dtype: int64

roads = gpd.read_file(fileroad, layer="lines")
Produces a larger dataframe that includes footpaths and other ways
roads.highway.value_counts():

highway
service           401909
residential       121843
footway           115707
tertiary           18397
secondary          17783
primary            13441
path                6454
trunk               6387
unclassified        5004
motorway_link       4388
cycleway            3148
motorway            2581
primary_link        2010
steps               1819
secondary_link      1706
track               1697
trunk_link          1389
tertiary_link       1004
corridor             407
pedestrian           337
construction         245
bridleway            144
living_street        113
raceway               68
proposed              42
road                  14
services              10
disused                3
busway                 2
rest_area              1
ramp                   1
Name: count, dtype: int64

There are some discrepancies in the value_counts on the same file, but largely seem to track. I had less trouble with read_file() in larger datasets, so I will be changing it to that. Closing, with the conclusion that the Pandas version does not seem to affect the road data use case here.