fix reading raster tiles with temporal dimensions

[goergen95]

No description provided.

[goergen95]

Hi, I this fixes the issue, I think. Would you mind to check? I also tried to make the code more understandable. Maybe let's have a discussion about the assumptions used here?

[codecov]

Codecov Report

Attention: 1 lines in your changes are missing coverage. Please review.

Comparison is base (70157bb) 73.31% compared to head (4fd992c) 74.40%.
Report is 1 commits behind head on main.

Files	Patch %	Lines
R/calc_indicators.R	95.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #211      +/-   ##
==========================================
+ Coverage   73.31%   74.40%   +1.09%     
==========================================
  Files          43       43              
  Lines        1825     1817       -8     
==========================================
+ Hits         1338     1352      +14     
+ Misses        487      465      -22     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[karpfen]

Sorry, I saw your request too late. I'm testing on my data right now that caused the problem in the first place.

[karpfen]

@goergen95 I think the original problem is solved, but it's still failing for me because the calc_landcover function returns zero-length tibbles instead of NAs. I'm currently going through my data to identify a reproducible example

[goergen95]

But shouldn't that be covered for with 87b99ba?

[karpfen]

I don't have the original error anymore, because my session crashed. But I think I received the new error message from calc_indicators:

mapme.biodiversity/R/calc_indicators.R

Lines 212 to 215 in 87b99ba

	if (any(n_rows == 0)) {
	stop(paste("0-length tibbles returned for some assets.\n",
	"Make sure the indicator function returns NA if it cannot be calculated for an asset."))
	}

This would mean that the landcover indicator returned a zero-length tibble. That's my guess at least. I'm still looking for the offending polygon.

[goergen95]

Oh, I see what you mean now. So there is a bug in the landcover function that is not directly related to this PR.

[karpfen]

You're right, I'll add it as an issue, I managed to find a reprex now, too.