Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
climada.hazard.trop_cyclone.TropCytlone.from_tracks: option for prede…
…fined dist_coast in centroids
- Loading branch information
1 parent
e59eb18
commit 0ad0804
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -268,8 +268,12 @@ def from_tracks( | |
| [idx_centr_filter] = (np.abs(centroids.lat) <= max_latitude).nonzero() | ||
| else: | ||
| # Select centroids which are inside max_dist_inland_km and lat <= max_latitude | ||
| if 'dist_coast' not in centroids.gdf.columns: | ||
| dist_coast = centroids.get_dist_coast() | ||
| else: | ||
| dist_coast = centroids.gdf.dist_coast | ||
This comment has been minimized.
Sorry, something went wrong.
This comment has been minimized.
Sorry, something went wrong. 
tovogt
Collaborator
|
||
| [idx_centr_filter] = ( | ||
| (dist_coast <= max_dist_inland_km * 1000) | ||
| & (np.abs(centroids.lat) <= max_latitude) | ||
| ).nonzero() | ||
|
|
||
|
|
||
@emanuel-schmid What do you think about attribute-style access to DataFrame columns in general? After years of working with pandas, I feel like this is very error-prone and should be avoided outside of private 10-line scripts that are written for a single purpose on your own machine. Reasons are: Improved syntax highlighting, more consistency (since in many cases you cannot use attribute-style access, so you are forced to fall back to item-style access), avoid mixing up attribute and column names. Actually, I have been wondering for a long time why pandas does not drop this anti-feature. As a maintainer, I would enforce item-style access in my code base. What my users do privately is up to them. But it's much easier to maintain the code if you have item-style access to DataFrame columns.
In this case, that would mean
centroids.gdf["dist_coast"], but there are cases in CLIMADA where the benefits would be really obvious. For example, in Exposures you have a column called"value"and it is so easy to mix this up with the attributevalues. Also, as a maintainer, you never know which built-in attributes will be added by pandas in the future that might eventually override your project-specific column names.If you agree, I would propose to add a user story in Taiga that enforces an item-style access policy throughout CLIMADA and goes through the existing code base to clean up the current mess. By the way, the same applies to variables and attributes of xarray Datasets and DataArrays.