excessive clipper run time due to long timeout setting?

[Sol Katzman]

In the past, I noticed excessive run times with clipper.

Investigating the code I found that there is a timeout set to 1000 seconds in peakfinder.py

That is nearly 17 minutes!

If you get 100 timeouts it is 28 hours!

I changed the code to use 100sec (probably 10sec would be sufficient for most data)

was: results.append(job.get(timeout=1000))
 is: results.append(job.get(timeout=100))

To see if this is a problem for you, look for these messages:

ERROR:root:transcript timed out

This default value should probably be changed in the source code.

/Sol.

[Brittany Cain]

Hi Sol,

What version of CLIPper are you using? I believe the version on the git hub master branch has a timeout option for this purpose.

[Sol Katzman]

Dear Brittany,

I was using clipper 0.2.0

That's what I got when I did pip install back in 2018. I don't think the program is actively maintained.

What version do you have? Do you see a timeout option?

/Sol.

[Brittany Cain]

Hi Sol,

I am using CLIPper 1.2.1, which can be cloned from the github master branch. This version introduces a timeout option as well as removes the bonferroni tag and algorithm options. This version was released in 2017, while the version on pip has not been updated since 2014.

If you ever need CLIPper again, I would use this updated version. It calls five times as many peaks and was easier to get running. It does take a lot longer to run than CLIPper 0.2.0 though.

Brittany