fprintf invalid format in rezToPhy2

[DradeAW]

Hi,

I'm trying out Kilosort 3, and I managed to make it run, but it crashes at the very end (it saves everything to phy except for the params.py file)

The error is:

Error using fprintf
Invalid format.

Error in rezToPhy2 (line 197)
        fprintf(fid,['dat_path = ''', strrep(rez.ops.fproc, '\', '/') '''\n']);

Error in main_kilosort3 (line 52)
rezToPhy2(rez, rootZ);

[marius10p]

Can you stop it in debug mode to see what fprintf is upset about? The problem seems to be specific to your data paths and maybe file system format. What OS?

[DradeAW]

I'm re-running it with a breakpoint on the fprintf, but I'm not sure how to debug it (this is the very first time I'm using Matlab).

I'm on Ubuntu 20.04, with Matlab R2019b.

[marius10p]

You could try sprintf instead of fprintf, it might be more informative. Also, just disp the relevant variables (ops.fproc) and post the results here. Is this a new line in the actual code or just in the error: ```fprintf(fid,['dat_path = ''', strrep(rez.ops.fproc, '\', '/') '''\n']);``` There shouldn't be one there.

…

On Fri, Feb 12, 2021, 6:55 AM Dradeliomecus ***@***.***> wrote: I'm re-running it with a breakpoint on the fprintf, but I'm not sure how to debug it (this is the very first time I'm using Matlab). I'm on Ubuntu 20.04, with Matlab R2019b. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#336 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA6AYDUP26P2SKGMQTXXJ5LS6UJKPANCNFSM4XQRQPCQ> .

[DradeAW]

sprintf did not change anything.

The value of rez.ops.fproc is /users/nsr/wyngaard/Documents/tmp/temp_wh.dat (shouldn't it point to the original dat file?).

The error is the line fprintf(fid,['dat_path = ''', strrep(rez.ops.fproc, '\', '/') '''\n']); itself.

[marius10p]

It's supposed to point to the drift-corrected file created by Kilosort, otherwise the clusters won't make sense if you have a lot of drift. I don't really know why that line is failing for you. The params.py file should just look like this:

dat_path = 'G:/Spikes/Sample/temp_wh.dat'
n_channels_dat = 383
dtype = 'int16'
offset = 0
sample_rate = 30000.
hp_filtered = True

[DradeAW]

I tried the following in the console, does it help?

>> aze = "/users/nsr/wyngaard/Documents/tmp/temp_wh.dat"

aze = 

    "/users/nsr/wyngaard/Documents/tmp/temp_wh.dat"

>> ['dat_path =''', strrep(aze, '\', '/') '''\n']

ans = 

  1×3 string array

    "dat_path ='"    "/users/nsr/wyngaard/Documents/tmp/temp_wh.dat"    "'\n"

>> fid = fopen("/users/nsr/wyngaard/Documents/cells_tracking/0153/session_1/rec/test.py", 'w');
>> fid

fid =

     5

>> fprintf(fid, ['dat_path = ''', strrep(aze, '\', '/') '''\n'])
Error using fprintf
Invalid format.

[marius10p]

Can you please try it without the strrep, so just

fprintf(fid, ['dat_path = ''', aze '''\n'])

[DradeAW]

I tried it, and it still crashes:

>> fid = fopen("/users/nsr/wyngaard/Documents/cells_tracking/0153/session_1/rec/test.py", 'w');
>> fprintf(fid, ['dat_path = ''', aze '''\n'])
Error using fprintf
Invalid format.

However, I managed to change the line so that it does work (using concatenation instead of an array):

>> fid = fopen("/users/nsr/wyngaard/Documents/cells_tracking/0153/session_1/rec/test.py", 'w');
>> fprintf(fid, 'dat_path = ''' + aze + '''\n')

ans =

    59

[marius10p]

Does params.py look ok, and does it open in Phy? I'll try concatenation on my end too.

[DradeAW]

Yes params.py looks as it should, and it opens with Phy!