search

4DModeller / fdmr

4DModeller, Bayesian Spatio-temporal modeling in R
https://4dmodeller.github.io/fdmr/
GNU General Public License v3.0
15 stars 10 forks source link

fdmr::model_builder crashes when building a Gaussian model #314

Open servinar opened 3 months ago

servinar commented 3 months ago

fdmr::model_builder crashes when building a Gaussian model

fdmr::model_builder(spatial_data = nepal_data, measurement_data = nepal_data, mesh = mesh, time_variable = "cyear") Warning: Cannot read CRS from mesh or spatial_data, using default CRS = +proj=longlat +datum=WGS84We will write log files to /var/folders/89/fmfv62b901711mv5c3lwbdqr0000gn/T/fdmr/logs Listening on http://127.0.0.1:5967 Loading required package: INLA Loading required package: Matrix Loading required package: sp This is INLA_24.03.24 built 2024-03-24 18:15:14 UTC.

  • See www.r-inla.org/contact-us for how to get help.
  • List available models/likelihoods/etc with inla.list.models()
  • Use inla.doc() to access documentation Warning: Error in : C stack usage 7955544 is too close to the limitWarning: no non-missing arguments to max; returning -InfError in rep_len(TRUE, length(callnames) - toRemove) : invalid 'length.out' value

For reference, the data we are using has the following details:

nepal_data class : SpatialPointsDataFrame features : 306000 extent : 79.62891, 88.06641, 26.36719, 30.35156 (xmin, xmax, ymin, ymax) crs : +proj=longlat +datum=WGS84 +no_defs variables : 9 names : year, locid, model, realization, prcp, topo, weights, cyear, optional min values : 3287, 1, GloSea5, 1, 2.3797812461853, 3.04741930961609, 1, 1, 1 max values : 18262, 450, MSWEP, 24, 999.987182617188, 6757, 1, 42, 1

head(nepal_data) year locid model realization prcp topo weights cyear 1 1979-01-01 1 MSWEP 1 35.66994 138 1 1 2 1980-01-01 1 MSWEP 1 77.51134 138 1 2 3 1981-01-01 1 MSWEP 1 95.28750 138 1 3 4 1982-01-01 1 MSWEP 1 44.88603 138 1 4 5 1983-01-01 1 MSWEP 1 89.32889 138 1 5 6 1984-01-01 1 MSWEP 1 45.30796 138 1 6

mnky9800n commented 3 months ago

Can you try with a smaller subset of the data

servinar commented 3 months ago

100 rows caused the same error. We are able to fit the model directly in R however. The shiny app will not process the models

mnky9800n commented 3 months ago

That's good news.