search

alihaydaroglu / suite3d

Fast, accurate, volumetric cell detection. Developed for Light Beads Microscopy, usable for other volumetric 2P. In development
6 stars 0 forks source link

Support for non lbm data loading #41

Closed landoskape closed 5 months ago

landoskape commented 1 year ago

Main changes

Added a new method in lbmio called load_and_stitch_tifs_notLBM that loads tiff files from standard scanimage outputs.

The main load_and_stitch_tiffs method has a few extra input arguments, and if it sees lbm=False, it'll return the output of the new method.

The default_parameters now include this lbm, num_colors, and functional_color_channel.

Smaller updates

I had to add these extra inputs a few places so that it uses the right loading method. Including:

But I ignored adding those parameters here because it seems like this is never used

landoskape commented 1 year ago

Should I update the load_and_stitch_tiffs method within the suite2p folder too? I don't know when this one is used...

alihaydaroglu commented 1 year ago

Should I update the load_and_stitch_tiffs method within the suite2p folder too? I don't know when this one is used...

* suite2p / io / lbm.py

No, this is legacy code that will be removed soon.

Thanks for the update! I am not going to merge this right now since there are probably a few more changes required in other parts. Once you are able to run an example tiff file through all the steps I'll merge

landoskape commented 1 year ago

Sounds good, I'll let you know when I've tested it on my data.

landoskape commented 1 year ago

@alihaydaroglu, in init_pass there is a code block that I think is exclusively related to stitching ROIs together in LBM.

This is it, found here in the code:

job.log("Estimating fusing shifts")
__, xs = lbmio.load_and_stitch_full_tif_mp(init_tifs[0], channels=n.arange(1), get_roi_start_pix=True)
fuse_shifts, fuse_ccs = utils.get_fusing_shifts(im3d_raw, xs)
fuse_shift = int(n.round(fuse_shifts.mean()))
if job.params.get('fuse_shift_override', None) is not None:
    fuse_shift = int(job.params['fuse_shift_override'])
    job.log("Overriding", 2)
job.log("Using best fuse shift of %d" % fuse_shift)
job.log("Building ops file")
# return

Then fuse_shifts is used again here: 

init_mov, xpad, ypad, new_xs, og_xs = utils.pad_and_fuse(init_mov, plane_shifts=tvecs, fuse_shift=fuse_shift, xs=xs)

I'm not sure how of these sections I need for non-LBM data. Can you advise how I should edit this section?

alihaydaroglu commented 1 year ago

I think I tried to fix this in the latest commit in my dev branch: https://github.com/alihaydaroglu/s2p-lbm/commit/0edac702d3176f0c1830c65506059e70dcd907fe

If you set the paremeter fuse_strips to False it should skip that part

landoskape commented 1 year ago

Yep, made it through that section now. Thanks!

I suppose the params['subtract_crosstalk'] section is also unnecessary for non-LBM data?

alihaydaroglu commented 1 year ago

Yep, that should be false

alihaydaroglu commented 5 months ago

@landoskape is this a branch you use? how do you load your data in?

landoskape commented 5 months ago

I have been working from a branch on my fork called landoskape/standard-2p-pipeline. I'm still getting things working there (refactors had to be addressed after my previous work). I can move this to a branch on your main fork if you'd prefer, let me know

alihaydaroglu commented 5 months ago

Ah ok, totally up to you where you keep the fork, I'd appreciate if eventually when things work you can create a PR to merge those back into main

* no rush

From: Andrew Landau @.> Sent: Tuesday, June 25, 2024 4:20 PM To: alihaydaroglu/suite3d @.> Cc: Haydaroglu, Ali @.>; State change @.> Subject: Re: [alihaydaroglu/suite3d] Support for non lbm data loading (PR #41)

⚠ Caution: External sender

I have been working from a branch on my fork called landoskape/standard-2p-pipeline. I'm still getting things working there (refactors had to be addressed after my previous work). I can move this to a branch on your main fork if you'd prefer, let me know

— Reply to this email directly, view it on GitHubhttps://github.com/alihaydaroglu/suite3d/pull/41#issuecomment-2189243774, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ADSYD3X636B72XLVFSTCLZLZJGDCXAVCNFSM6AAAAABJYX25IOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCOBZGI2DGNZXGQ. You are receiving this because you modified the open/close state.Message ID: @.***>