Non-reliable saving

[liadJB]

Hi,

I am attaching a script we are using. We are having problems whereby all the files are not consistently saved. Always on the first run of the protocol not everything is saved. The chances of saving are higher on the second run, but still sometimes not.
The thing that's not always saving is the arudino input:

Any idea why it might happen?

Gratings.zip

Thanks

[bruno-f-cruz]

Can you provide a little more details? Is it the case that data is missing? The full file is missing? part of the last row is not showing? Do you have a sense how many samples are missing? How did you arrive to this conclusion?

Additionally, can you provide a minimal example that reproduces the bug without having to run the full workflow? Thanks

[liadJB]

Thanks. I will try to see if I can reproduce it without the rest. In the meantime. The file is created with 0kb. So it seema like there is some sort of race condition that sometimes prevents it being written.

…

On Thu, 25 Apr 2024, 14:47 brunocruz, ***@***.***> wrote: Can you provide a little more details? Is it the case that data is missing? The full file is missing? part of the last row is not showing? Do you have a sense how many samples are missing? How did you arrive to this conclusion? Additionally, can you provide a minimal example that reproduces the bug without having to run the full workflow? Thanks — Reply to this email directly, view it on GitHub <#1758 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALPHYVOQI3RWMUHX4PRQMEDY7ECNPAVCNFSM6AAAAABGYTQL7SVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4TEMRVGU4DS> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[bruno-f-cruz]

Are you sure you have events taking place downstream from your condition node? The CSV will be open at subscription time, NOT when the first sample arrives.

[liadJB]

Yep. Nothing changes between the firat run and the second. Stimuli are oresented and I can view everything.

…

On Thu, 25 Apr 2024, 15:19 brunocruz, ***@***.***> wrote: Are you sure you have events taking place downstream from your condition node? The CSV will be open at subscription time, NOT when the first sample arrives. — Reply to this email directly, view it on GitHub <#1758 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALPHYVK3KQOYXDJ44H6SZOTY7EGHPAVCNFSM6AAAAABGYTQL7SVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4TEMRVHE4TM> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[bruno-f-cruz]

So if you inspect the output of the condition node while bonsai is running you see values in both "runs"? (What do you mean by runs btw?)

[liadJB]

Yes. One can see values in all nodes. By run I mean pressing the start button and stopping. If one then clises Bonsai one needs to 'run' a script first again.

…

On Thu, 25 Apr 2024, 15:49 brunocruz, ***@***.***> wrote: So if you inspect the output of the condition node while bonsai is running you see values in both "runs"? (What do you mean by runs btw?) — Reply to this email directly, view it on GitHub <#1758 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALPHYVIVIZJASEBCCQZFI43Y7EJYVAVCNFSM6AAAAABGYTQL7SVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4TEMRWGM2DK> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[Awlala]

Thank you for raising this question. I have the same problem. I have not yet been able to solve it and look forward to get some hints here. I was also wondering if you had any error messages raised in the log. It seems that it may provide more information about the problem and when my .csv files save properly, the error message does not show up.

[bruno-f-cruz]

I am not sure I follow. What I sent shouldn't throw an exception. Retry should keep trying to subscribe to the serial connection. Can you share the workflow you are using?

[liadJB]

Here is the example.
In the cases when the timeout is triggered there's a timeout exception, otherwise it's working as usual (in the good times, when the connection is actually established)

Example.zip

[glopesdev]

@liadJB there is a chance this could be a bug in the operator. If the error appears in the status bar at the bottom of the Bonsai editor, can you right-click the text and copy the full stack trace here?

[liadJB]

I also tried to implement it as Bruno suggested.

If I take the include the circled parts in a Defer group, there is an error saying the Arduino subject does not exist in the context of the part circled in blue.

Would there be any way to run a quick loop at the start of the run to verify that the serial port is working and somehow then continue ?

[liadJB]

Hi,

I am re-flagging it, since we updated the bonsai script in another of our rigs (without changing anything else - so the hardware and Arduino remained the same), and we are now having the same problem. So we can be fairly certain that the problem lies in the updated operators.
Since it is asynchronous it's hard to say what is happening, but maybe a comparison of what change between version could help understand what changed between the versions?

[liadJB]

Hi, There is just an error box saying that a Timeout exception was thrown.

…

On Wed, 22 May 2024 at 09:40, glopesdev ***@***.***> wrote: @liadJB <https://github.com/liadJB> there is a chance this could be a bug in the operator. If the error appears in the status bar at the bottom of the Bonsai editor, can you right-click the text and copy the full stack trace here? — Reply to this email directly, view it on GitHub <#1758 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALPHYVMJKJXVHPSSLRHU3ODZDRKZPAVCNFSM6AAAAABGYTQL7SVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4TKMJYGQ3DK> . You are receiving this because you were mentioned.Message ID: ***@***.***>