start geopandas extension #728
Conversation
|
resolves #136 |
pyproject.toml
Outdated
| @@ -44,6 +44,7 @@ classifiers = [ | |||
| cli = ["click>=8.1.7", "pyogrio>=0.8", "shapely>=2"] | |||
| geopandas = ["geopandas>=0.13", "pandas>=2", "shapely>=2"] | |||
| movingpandas = ["movingpandas>=0.17"] | |||
| explore = ["mapclassify>=2.8.1"] | |||
There was a problem hiding this comment.
I think we can include this in the geopandas extra
There was a problem hiding this comment.
ok, wasnt sure how you wanted to handle that part. Geopandas itself only has a soft dependency on mapclassify so its only imported when scheme is used; wasn't sure if you wanted to keep it similarly isolated (even though I hadn't structured the import like that)
There was a problem hiding this comment.
Hmm, in that case, maybe it is better to keep it separate. Especially if we move the mapclassify import only within the else clause where it's used.
In that case, it's probably fine to not even define mapclassify in an extra.
lonboard/geopandas.py
Outdated
| def explore( | ||
| self, | ||
| column=None, | ||
| cmap=None, | ||
| scheme=None, | ||
| k=6, | ||
| categorical=False, | ||
| elevation=None, | ||
| extruded=False, | ||
| elevation_scale=1, | ||
| alpha=1, | ||
| layer_kwargs=None, | ||
| map_kwargs=None, | ||
| classify_kwargs=None, | ||
| nan_color=[255, 255, 255, 255], | ||
| color=None, | ||
| wireframe=False, | ||
| tiles="CartoDB Darkmatter", | ||
| m=None, | ||
| ): |
There was a problem hiding this comment.
Can you expand on what differences (if any) there are with the upstream geopandas.explore method? I think we'd want it to be as similar as possible.
That would be great to include in this docstring.
Additionally, ideally there would be a way to get valid type hinting here, but presumably that won't work if this is added at runtime to the GeoDataFrame instance.
There was a problem hiding this comment.
sure happy to add both. In general I tried to keep all possible arguments the same (hence the tiles that uses a str --> class--> str, which isn't ideal but I'm hard conditioned to type tiles='CartoDB Positron') any time I explore.
the one thing I know is different is classification_kwargs used here vs classification_kwds used in geopandas, but I'll switch it to martin's version for compat. Otherwise, any arguments available in explore have the same name. Those that don't map between the two are unique to the backengs, e.g. layer_kwargs here vs style_kwds in vanilla explore
also I generally use numpy docstrings but I think the rest of the packages uses another format, so I should probably update that too
|
Keep me in the loop. Happy to do a proper review of this from the geopandas perspective once @kylebarron is happy. |
lonboard/geopandas.py
Outdated
| from mapclassify.util import get_color_array | ||
| except ImportError as e: | ||
| raise ImportError( | ||
| "you must have the `mapclassify` package installed to use this function" |
There was a problem hiding this comment.
Maybe more clear?
"you must have the `mapclassify` package installed if you pass `classification_kwds`"
pyproject.toml
Outdated
| @@ -42,9 +42,8 @@ classifiers = [ | |||
|
|
|||
| [project.optional-dependencies] | |||
| cli = ["click>=8.1.7", "pyogrio>=0.8", "shapely>=2"] | |||
| geopandas = ["geopandas>=0.13", "pandas>=2", "shapely>=2"] | |||
| geopandas = ["geopandas>=0.13", "mapclassify>=2.8.1", "pandas>=2", "shapely>=2"] | |||
There was a problem hiding this comment.
I think we can take out the dependency altogether from here. If someone hits the error they'll be instructed to install it.
| highlight=False, | ||
| m=None, | ||
| ): | ||
| """explore a dataframe using lonboard and deckgl |
There was a problem hiding this comment.
If this docstring is exactly the same as the public-facing docstring, you can just write here to refer to the public docstring for parameter information.
| ): | ||
| """explore a dataframe using lonboard and deckgl | ||
|
|
||
| Parameters |
There was a problem hiding this comment.
Yeah we do use the other docstring format in lonboard (I can't remember which of numpy/google format is which), and it would be good to standardize on that format.
to begin review. The notebook at examples/explore should run through and demonstrate functionality. I need to take a look at how you prefer to test against Maps