Recreation model crashes on server with NumPy header error even with compute_regression = False

Software Support
1

Hi,

I’m running the InVEST Recreation model (version 3.16.1) for a study area in İzmir, Türkiye. I tested both with and without regression enabled, but the model fails in both cases at the server side.

Parameters used:

  • compute_regression = False
  • start_year = 2012
  • end_year = 2017
  • grid_aoi = True
  • cell_size = 1800

What happens:

The model successfully connects to the server, retrieves Flickr/Twitter points, and begins calculating PUD/TUD. However, during quadtree flushing and zipping, the model crashes on the server with this error:

AttributeError: module ‘numpy.lib.format’ has no attribute ‘_read_array_header’

This happens regardless of the regression flag. The full log confirms that the error occurs deep within the server’s numpy handling during user-day-calculation.

Could you please check the NumPy version on the recreation server and consider updating the related code? Alternatively, is there any known workaround?

Thanks in advance for your support!

InVEST-recreation-log-2025-08-07–11_46_09.txt (18.9 KB)

2

@ebruska , thanks for reporting this. We are not sure yet why the verison of numpy running on the server appears to have changed, but we will address the issue soon.

3

@ebruska We updated the server so please try again with your same study area. Thanks again for reporting the problem.

1 Like
4

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.