hyperspy/hyperspy

Multidimensional data analysis

UrsulaLu

@UrsulaLu

If it is helpful, here a small code snippet, leading to it

Magnus Nord

@magnunor

Just to check: are you able click and drag the "span" the region where you want to fit the background? Do you also get the "GUI-box" directly in the notebook as well?

Jonas Lähnemann

@jlaehne

As far as I remember the closing windows are an issue that is trailing around for longer already and not specific to hyperspy, but comes somewhere from the gui-backends. I also experience a closing of the qt-windows, but not completely reproducibly, often only after a while. My personal workaround at the moment is to have both the traitsui and the ipywidgets gui packages installed. If the qt-window closes, I still have the inline-controls for background removal (or other functions), while the external plot window still remains active and I can drag and span the background region as described by @magnunor .

Does anyone know if we ever figured out the mystery of traits windows randomly closing during interaction (like during remove background or cropping)?

trailing around for a really long time ;-)

UrsulaLu

@UrsulaLu

@magnunor the thing is, there appear no additional plots on top of the initial spectrum, no region (light red) and neither the green nor the blue curve. And just see what you mean with the GUI box, no also that doesn't show up. Same with: %matplotlib notebook

Jonas Lähnemann

@jlaehne

@UrsulaLu the shaded region and additional curves appear only once you click and drag/span with the mouse on the figure to define the background region. It works for me, even though the initial values are nan in the dialog.

UrsulaLu

@UrsulaLu

@jlaehne I can for example still change the spectrum navigator, and the spectrum changes, but the shaded box doesn't appear - no matter where I click. Curious, but why is it showing nans?

Jonas Lähnemann

@jlaehne

Well, the background region is not initiated with any values, so the start and end values do not exist. In the traitsui (qt) window it shows nans, in the ipywidgets inline display it just shows an empty field.

For me an alternative way to make it work is to use %matplotlib widget together with the hyperspy-gui-ipywidgets package. The plot is then displayed inline and I can also add the background region by clicking and dragging on the spectrum window.

Minimum working example:

%matplotlib widget
import hyperspy.api as hs
import numpy as np
s = hs.signals.Signal1D(np.arange(1024))
s.remove_background()

UrsulaLu

@UrsulaLu

It is really weird, tried it, I get also now the inline GUI window, but still no chance for get
ting the shaded box

UrsulaLu

@UrsulaLu

Restarted Kernel and now throws the error. The ipywidgets are installed.

UrsulaLu

@UrsulaLu

@jlaehne After installing ipympl 0.7.0 the background subtraction now finally works. Thanks!

Naren

@narendraraj

we can use python based script the collect 4DSTEM data series with EDX data. We save in h5 format not able to structure it properly include metadata and make it supportable for hyperspy. Please someone guide me how to write out data on hdf5 formate that can support hypersy or /and py4dstem or/and pyxem. If there is information already on how write data please share the link.

7 replies

MurilooMoreira

@MurilooMoreira

Hello guys, I searched a lot in the github and here in the gitter for solutions about how to open jeol eds data and I saw people are trying to do that since few years, but I don't know if someone figured out how to do it in hyperspy. Is it possible? I am with data from jeol eds (.pts, .map, .img, .asw) and I am with difficulties to open it. Someone could help me please?

MurilooMoreira

@MurilooMoreira

I found here that the function to open this type of data is already implemented and it works hyperspy/hyperspy#2488

7 replies

But I don't know how to access the tool to use in my jupyter notebook for example.

wstolp

@wstolp

Hi, rebinning a ComplexSignal2D gave me an error, so I made a small change to the source code, which seems to work. It now rebins the real and imaginary parts seperately.
Anything I need to know before I send a pull request? The developer guide gives me a 404 :(

also I'm very new with github

Thomas Aarholt

@thomasaarholt

Hi @wstolp! Taking a look now about the dev guide!

I can't reproduce the 404 - where did you find it?

I went to https://hyperspy.org/ and pressed "Documentation" under both Stable and Development (on the right hand side)

And then look under "Developer Guide"

We're very happy to help new developers like yourself, so just have a go and we'll help with what you need :)

wstolp

@wstolp

Hi Thomas. Yes later I saw there are multiple developer guide type pages. The 404 I saw here https://github.com/hyperspy/hyperspy, scroll down to "contributing guidelines", then "developer guide".

purewendy

@purewendy

Hello, I would like to do EELS fitting on my data. The hs.preferences.gui() links to GOS directory to C:\Program Files\Gatan. This is the directory of my installed Digital Micrograph. But it's show show errors. Any idea to fix it? Thanks very much!

Weixin Song

@winston-song

Hi All, I use hs.plot.plot_images(atom_percent, scalebar =[0],scalebar_color='white'), is there any way to customize the scalebar length and the fontsize of the scalebar?

Weixin Song

@winston-song

Hi All, I use hs.plot.plot_images(atom_percent, scalebar =[0],scalebar_color='white'), is there any way to customize the scalebar length and the fontsize of the scalebar?

atom_percent is a list

Thomas Aarholt

@thomasaarholt

@purewendy what is your error?

Jędrzej Morzy

@JMorzy

Hi, I have a ragged (yeah, I know :/) data array (1 spectral dimension, 2 'navigation' dimensions that are actually scan number and time). I'd like to do some PCA on it. At the moment, the empty spots in the ragged array are filled with NaNs. Is there any way of making it work despite the NaNs? I was thinking it should be possible to reshape the array into a 1D array (flattening the scan and scan number away) and then re-reshaping it back and looking at the loadings afterwards, but that sounds like a total botch. Any more elegant ideas?

Sorry, 2D. One very long navigation dimension and one normal signal dimension of course

Thomas Aarholt

@thomasaarholt

Hi @JMorzy, which part of your data is ragged? Are you missing entries in the signal or navigation dimension?

I normally think of ragged arrays as ala:

[
    [1,2,3],
    [1,2,3,4],
]

How does this compare with your data?

Jędrzej Morzy

@JMorzy

Exactly, ragged in the navigation dimension. The scans are different lenghts (scan number is my y, time is my x)

Then to make it non-ragged to put it into numpy array/hs signal I filled the gaps with NaNs

2 replies

Corentin Le Guillou

@CorentinLG

Hi all, I am using the non_uniform_axes brnach. Since last update (apparently), the m.set_signal_range() function does not accept explicit axis values, but only axis indices. Ant idea why ? And a suggestion to go around that problem ? thanks

3 replies

Mingquan Xu

@Mingquan_Xu_twitter

Hi, all, is there any packages that can do local low rank denoising for EEL spectrum image?

Thomas Aarholt

@thomasaarholt

What is low rank denoising? I use PCA and ICA a lot for eels in hyperspy.

https://www.sciencedirect.com/science/article/abs/pii/S030439911730325X

I see. I hadn't heard about it before.

Mingquan Xu

@Mingquan_Xu_twitter

Hi, @thomasaarholt , yes, I also know this from this article, but do not know which software can do this process.

I have tried PCA and NMF in hyperspy on my data (SI), the results are not so good.

Thomas Aarholt

@thomasaarholt

Havd you thought about what might be causing your data's results to be "not so good"? What sort of eels is it?

rtangpy

@rtangpy

Hi all, I am doing model fitting, with 2D navigation and 1D signal. I have two questions: 1. since I need to fit more than 1 million pixels, using multifit will take me 50 mins. Is there any method that can speed up the code; 2. before I learned how to use hyperspy, I use multiprocessing with 8 cores to perform optimize.minimize to speed up code. Interestingly, using hyperspy multifit which only uses one core is even slighly faster than my code with multiprocessing (8 cores). I am curious about how hyperspy can reach such high speed.

13 replies

Where communities thrive

People

Repo info

Activity