AttributeError: '_GeneratorContextManager' object has no attribute 'read_directory'

[gihanpanapitiya]

I am getting the following error when trying to run tags = dm4data.read_directory()

AttributeError                            Traceback (most recent call last)
Cell In[20], line 1
----> 1 tags = dm4data.read_directory()

AttributeError: '_GeneratorContextManager' object has no attribute 'read_directory'

[jamesra]

I'll take a look and see if I can recreate this problem.

[jamesra]

The unit tests are passing. This may be a usage or documentation issue. Can you share the code causing the error?

The test_dm4.py file is how I'm testing. If I'm missing a test case you need please let me know.

[gihanpanapitiya]

I am using the example given in the README.

import dm4

dm4data = dm4.DM4File.open("sample.dm4")

tags = dm4data.read_directory()

image_data_tag = tags.named_subdirs['ImageList'].unnamed_subdirs[1].named_subdirs['ImageData']
image_tag = image_data_tag.named_tags['Data']

XDim = dm4data.read_tag_data(image_data_tag.named_subdirs['Dimensions'].unnamed_tags[0])
YDim = dm4data.read_tag_data(image_data_tag.named_subdirs['Dimensions'].unnamed_tags[1])

np_array = np.array(dm4data.read_tag_data(image_tag), dtype=np.uint16)
np_array = np.reshape(np_array, (YDim, XDim))

output_fullpath = "sample.tif"
image = PIL.Image.fromarray(np_array, 'I;16')
image.save(output_fullpath)

The .dm4 files I use are from here https://figshare.com/articles/dataset/Data_Set_from_Forecasting_of_In_Situ_Electron_Energy_Loss_Spectroscopy/20288730/1?file=36266061

[gihanpanapitiya]

By going through test_dm4.py, I was able to come up with the
following code to convert a .dm4 to .tif.

dm4_input_fullpath="Image.dm4"

output_dirname='./'

with dm4.dm4file.DM4File.open(dm4_input_fullpath) as dm4file:
    tags = dm4file.read_directory()

    data_tag = tags.named_subdirs['ImageList'].unnamed_subdirs[1].named_subdirs['ImageData'].named_tags['Data']
    
    file_basename = os.path.basename(dm4_input_fullpath)
    output_filename = os.path.basename(file_basename + '.tif')
    

    output_fullpath = os.path.join(output_dirname, output_filename)
    image_dimensions_tag = tags.named_subdirs['ImageList'].unnamed_subdirs[1].named_subdirs['ImageData'].named_subdirs[
                'Dimensions']

    np_array = np.array(dm4file.read_tag_data(data_tag), dtype=np.uint16)

    xdim = dm4file.read_tag_data(image_dimensions_tag.unnamed_tags[0])
    ydim = dm4file.read_tag_data(image_dimensions_tag.unnamed_tags[1])

    np_array = np.reshape(np_array, (xdim, ydim))


image = Image.fromarray(np_array, 'I;16')
image.save(output_fullpath)

Thank you for creating this repository!!

[jamesra]

I'm glad you got something working. I've dug into this and think I have a solution. I did not notice this issue was on my personal repository and not the official repository. The official repository had the same issue that I forgot to merge a fix for, sorry. My personal repository (this one) was several commits behind the official version but is caught up now.

I've updated the documentation, added the example code to the unit tests, and published a new version to PyPI. I suggest using pip install dm4 if you are not already. Definitely switch to the code in the official repository if you want to run this package from the source code. The official version renamed the package to dm4 from dm4reader once it was published to PyPI. The update I made today also allows you to list a dm4 files directory tree by invoking dm4 as a package python -m dm4 your_dm4_file.dm4 which should help if your dm4 files have a slightly different structure than our platform.

Let me know if you have any more trouble with the example code.