empyre issueshttps://iffgit.fz-juelich.de/empyre/empyre/-/issues2021-09-03T15:01:37Zhttps://iffgit.fz-juelich.de/empyre/empyre/-/issues/58Loading vtk data as 2D data set2021-09-03T15:01:37ZTeresa WeßelsLoading vtk data as 2D data set### Summary
I am trying to load an unstructured vtk data set. I would like to have a 2D array as an output. However, field setter complains because it gets a 3D scale for the 2D output.
### Steps to reproduce
Use emp.ip.load_field() a...### Summary
I am trying to load an unstructured vtk data set. I would like to have a 2D array as an output. However, field setter complains because it gets a 3D scale for the 2D output.
### Steps to reproduce
Use emp.ip.load_field() and set the scale in such a way that an output grid of (1, z, x, 3) is generated.
### What is the current *bug* behavior?
```AssertionError: Each of the 2 dimensions needs a scale, but (20, 20, 20) was given!```
### Relevant logs and/or screenshots
![loading_2d_from_vtk](/uploads/b38b752a5c78f8bf78dead7527b2f4d1/loading_2d_from_vtk.png)
~bug ~io ~fieldshttps://iffgit.fz-juelich.de/empyre/empyre/-/issues/57Loading Pyramid dataset with Empyre2021-08-04T09:32:18ZTeresa WeßelsLoading Pyramid dataset with Empyre## Summary
When loading Pyramid datasets with Empyre, the component axis needs to be shifted. This can be done by setting the parameter comp_pos which works for hdf5 files. However, it is not working for npy files.
### Steps to reprodu...## Summary
When loading Pyramid datasets with Empyre, the component axis needs to be shifted. This can be done by setting the parameter comp_pos which works for hdf5 files. However, it is not working for npy files.
### Steps to reproduce
Load any dataset save in numpy with "emp.io.load_field('dataset.npy',comp_pos=3)" while setting the parameter comp_pos
### What is the current *bug* behavior?
Numpy.load() gets the parameter comp_pos which it does not know and throws an error.
### Relevant logs and/or screenshots
```
TypeError Traceback (most recent call last)
<ipython-input-10-5fcc66790585> in <module>
----> 1 field_sum = emp.io.load_field('Hdem-45deg_10100100_step3.npy',comp_pos=3)
~/.conda/envs/empyre_1/lib/python3.8/site-packages/empyre/io/io_field.py in load_field(filename, scale, vector, **kwargs)
67 for plugin in plugin_list: # Iterate over all plugins:
68 if extension in plugin.file_extensions: # Check if extension is recognised:
---> 69 return plugin.reader(filename, scale=scale, vector=vector, **kwargs)
70 # If nothing was found, try HyperSpy
71 _log.debug('Using HyperSpy')
~/.conda/envs/empyre_1/lib/python3.8/site-packages/empyre/io/field_plugins/numpy.py in reader(filename, scale, vector, **kwargs)
24 if scale is None:
25 scale = 1.0
---> 26 return Field(np.load(filename, **kwargs), scale, vector)
27
28
TypeError: load() got an unexpected keyword argument 'comp_pos'
```
### Possible fixes
Exclude comp_pos from kwargs for numpy.
Workaround: only use hdf5 format to save files
~bug ~iohttps://iffgit.fz-juelich.de/empyre/empyre/-/issues/563d plot only partly shown using png backend2021-01-12T09:33:48ZTeresa Weßels3d plot only partly shown using png backend### Summary
When plotting a 3d image using the png backend, only a part of the whole plot is displayed.
### Steps to reproduce
```python
import empyre as emp
from mayavi import mlab
import numpy as np
mlab.init_notebook('png')
test = ...### Summary
When plotting a 3d image using the png backend, only a part of the whole plot is displayed.
### Steps to reproduce
```python
import empyre as emp
from mayavi import mlab
import numpy as np
mlab.init_notebook('png')
test = emp.fields.Field(np.ones((20,20,20)))
emp.vis.mask3d(test)
```
### What is the current *bug* behavior?
Only the lower left quarter of the image is shown, the rest stays black.
### What is the expected *correct* behavior?
The full image should be visible.
### Relevant logs and/or screenshots
![issue_png_back](/uploads/02d9e7e1756b9933cb10fe5eba2ef63f/issue_png_back.png)
~bug ~vishttps://iffgit.fz-juelich.de/empyre/empyre/-/issues/54Too small `coords` for some reason2020-10-16T11:11:40ZAlexander Clausena.clausen@fz-juelich.deToo small `coords` for some reasonEnvironment: Windows
Looks something like this:
![image](/uploads/328acd4762f9c989b110337bde39e4f1/image.png)Environment: Windows
Looks something like this:
![image](/uploads/328acd4762f9c989b110337bde39e4f1/image.png)https://iffgit.fz-juelich.de/empyre/empyre/-/issues/38pip install .[io] not working due to hyperspy depending on traits (no wheels ...2020-03-17T23:27:39ZJan Caronpip install .[io] not working due to hyperspy depending on traits (no wheels on PyPI)Currently, `pip install .[io]` is not working, due to hyperspy depending on traits, which does not have precompiled wheels on PyPI. The `test_install` job does not work without a wheel, or without getting Microsoft Visual C++ 14 for its ...Currently, `pip install .[io]` is not working, due to hyperspy depending on traits, which does not have precompiled wheels on PyPI. The `test_install` job does not work without a wheel, or without getting Microsoft Visual C++ 14 for its gcc (which is a large download and would complicate `gitlab-ci.yml`).
The [issue](https://github.com/hyperspy/hyperspy/issues/2315) is known to the HyperSpy guys and traits also has a [request to build wheels for PyPI](https://github.com/enthought/traits/issues/357), but it does not look like this will be fixed in the near future...
For now, hyperspy is removed from `[all]`, if people want to use hyperspy for io, they need to install it with conda for now (or
get a wheel for Windows from [gohlke](https://www.lfd.uci.edu/~gohlke/pythonlibs/).
~bug
UPDATE: `mayavi` has the same problem and `tvtk` is also not easily installable via pip. As it stands, the `[all]` is quite empty...