Only saving header.npy, not writing image data to jpg

[mpm896]

Hi Alex,

I'm trying this program out and it seems like it's only writing the header data to a header.npy file and not producing the tomogram.jpg file. I thought maybe it was saving to some unknown location, so I moved up a few directories and ran find . -name "*jpg*" but there are no files returned. I checked out the source code and I don't really know why it wouldn't be writing the data since pack_file calls mrc_to_jpeg_stack which calls save_image and writes the data. All the appropriate output messages are printing.

Additionally, it seems that if I set the output location to a different path than the input data, it fails to find and read the .npy header.

An example command I'm running (from the jpeg_tomogram directory):
./jpeg_tomogram.py pack ../mm2024-07-25-1/Position_1_rec.mrc

I've tried this on both MacOS and Ubuntu 20.04, with the same results for both

[alexjnoble]

Hi Matthew,

The default output path is the same as the input path. If you would like to change it, use the --output or -o flag like this:
./jpeg_tomogram.py pack ../mm2024-07-25-1/Position_1_rec.mrc -o .

Let me know if that solves the problem.
Best,
-Alex

[mpm896]

Hi Alex,

Thanks! The issue I'm having I think is a bit more nuanced. For the first question where I couldn't find the saved jpg file, I found the bug in line 176:
jpeg_stack_filename = jpeg_stack_filename.rsplit('.', 1)[0] + f'.jpgs'

Since I was entering a path to the file as ../mm2024-07-25-1/Position_1_rec.mrc, this was resulting in jpeg_stack_filename being ..jpgs. I think something like jpeg_stack_filename = f'{Path(jpeg_stack_filename).as_posix()}.jpgs' should avoid this.

However now that I have the output jpeg file, I can't seem to open it with 3dmod or FIJI. Any format I choose for 3dmod (signed/unsigned bytes, etc) opens a file with 1 slice that looks like a bunch of noise. When I try opening with FIJI nothing happens. I wonder if there's something I might be missing for this? Thanks!

[mpm896]

Ok ignore my last question - I thought the resulting jpg file was viewable by 3dmod. I just realized I had to unpack it back to an MRC file, sorry about that!

[alexjnoble]

Thanks Matthew. I've updated the script here, which hopefully addresses your issue:

cf71918

Let me know if that fixes things.

Best,
-Alex