correcting import of rectangular image array

[jhellerstedt]

Hi,

I've been dealing with a rectangular image file, 512x128 pixels. It seems that the convention for reshaping the raw bytes by the scan_pixels tuple from the header is mixed up? When I switched the nx and ny, it works perfectly on my image.

[underchemist]

hey just replying to say I'm looking at this finally.

I'm just going to write some test cases for differently sized grid inputs. It's very possible i mixed up the nx ny order, almost all my data is square!

[underchemist]

Just to clarify, how did you figure out that the nx ny order was mixed up? Are you plotting it with matplotlib's imshow?

The reason I'm asking is because imshow has an 'origin' parameter which defaults to 'upper' which flips all array images (i.e. scan data). The proper orientation that corresponds to what I view in the nanonis scan inspector is to use origin='lower' in my imshow call.

See https://stackoverflow.com/questions/14320159/matplotlib-imshow-data-rotated

If the nx ny issue is different can you explain it a bit more? or have a sample file that I can play with to see exactly how it's being loaded in with the read functions?

Thanks

[jhellerstedt]

I'm dealing with some 512x192 images, here's hopefully an sxm file that will reproduce the error:
https://www.dropbox.com/s/cn22o126ln9x6au/Ag111_TbPc2_830K_5min_072.sxm?dl=0
When I used nanonispy as-is, I'd get an image like this:

Then after trying a lot of things (the origin thing is a protip, thanks), after switching your nx, ny convention I would get the "correct" answer:

[jhellerstedt]

PS the same solution (swapping nx/ ny) fixed the same problem with reading in a rectangular grid spectroscopy.

[angelo-peronio]

Dear Yann-Sebastien,

thank you for nanonispy! It lets me enjoy more my research.

I support the merging of this pull request, in particular a6439b6 fixes #8, an issue I am facing in every .dat spectroscopy file I load.