search

alan-turing-institute / DetectorChecker

Project to develop software to assess developing detector screen damage (Web App based on the original DetectorChecker package)
https://detectorchecker.azurewebsites.net
MIT License
0 stars 1 forks source link

Travis / Vignette memory usage #62

Closed OscartGiles closed 5 years ago

OscartGiles commented 5 years ago

At the moment travis builds fail because there is not enough memory to build the vignettes.

The primary issue seems to be the GLM fits as the returned object is ~3.1 Gb.

ejulia17 commented 5 years ago

I'm a bit puzzled why the returned object is to big... I'm not sure I understand which object you are referring to. Could you say which function name, which object name exactly? Then I'll check.

OscartGiles commented 5 years ago

I'll have a chat with Tomas about this. It's not really a problem for the Vignette, more testing on Linux as the virtual machine it uses doesn't have enough memory.

If you run the Vignettes on the development branch lines 202 run this code (can also just run in R):

`detector_pilatus <- create_module("Pilatus")

file_path <- "../inst/extdata/Pilatus/badpixel_mask.tif"

detector_pilatus <- load_pix_matrix(detector = detector_pilatus, file_path = file_path)

glm_fit <- glm_pixel_ctr_eucl(detector = detector_pilatus) `

The object is returned by the glm function.

OscartGiles commented 5 years ago

Currently removed GLM model fitting from vignette. Closing for now, but could be reopened if required later.