Create raster layer from file

[wang-boyu]

Partly addresses #91.

The support for compressed files requires more effort, so I'm thinking to implement that in the next release. That probably won't be a breaking change, i.e., from

RasterLayer.from_file(raster_file, cell_cls=Cell, attr_name=None)

to

RasterLayer.from_file(raster_file, cell_cls=Cell, attr_name=None, compression="infer")

[codecov]

Codecov Report

Merging #92 (94cf2d0) into main (93e8d60) will increase coverage by 0.70%.
The diff coverage is 94.11%.

@@            Coverage Diff             @@
##             main      #92      +/-   ##
==========================================
+ Coverage   78.57%   79.28%   +0.70%     
==========================================
  Files           7        7              
  Lines         574      584      +10     
  Branches       97       98       +1     
==========================================
+ Hits          451      463      +12     
+ Misses        107      105       -2     
  Partials       16       16              

Impacted Files	Coverage Δ
mesa_geo/raster_layers.py	72.41% <94.11%> (+2.14%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 93e8d60...94cf2d0. Read the comment docs.

[rht]

We should be consistently use verb for function/method name. E.g. pd.read_csv.

[wang-boyu]

To me read_file is better when it's from the package level, e.g., pd.read_csv, gpd.read_file, whereas from_file sounds more natural for class method constructors, e.g., pd.DataFrame.from_dict, gpd.GeoDataFrame.from_file.

But this is just my personal preference.

[wang-boyu]

Shall I rename to RasterLayer.read_file then?

[rht]

For the pandas naming, it is not conclusive that they use read_* convention because it is for package level functions. But given that geopandas already uses from_file, then we have to follow suite, according to the principle of least surprise.