savente93
commented
3 months ago how would this work if someone enters a name like "松阿察河" (river in china)? In general doing these kinds of conversions can be dangerous if you don't know how your input space is structured. For example, if we would just Latinise these we'd get into trouble with things like: 问 = wèn = to ask 吻 = wěn = to kiss
Now we could require that input is ASCII encoded, but I think this will be quite an impediment to the adoption of FA outside of Anglicised regions.
I'm assuming that the text in the sfincs.obs file is mostly either decorative (labels) or used as some kind of unique identifier. Would it be at all feasible to ask SFINCS to allow utf-8 text in the obs file? I feel like that's the better solution here, as it would benifit the adoption of both sfincs and FA
FloodAdapt version checks
Reproducible Example
Including ¨ (umlauts) in the name of an obsrvation station makes floodadapt break at the SFINCS postprosessing step.
Current behaviour
Using special characters in the observation station name in the site.toml, breaks the workflow, since hydromt-sfincs cannot read the SFINCS model anymore (more specifically the issue is reading in the sfincs.obs file).
Desired behaviour
Add validators in the pydantic model for the site object to make sure that special characters are not included in the names of the observations stations. This way even if the user has manually change something, the validator would be run during startup and give an informative message.
This could be an issue for other parameters!
Task list
Additional context
No response