poretools icon indicating copy to clipboard operation
poretools copied to clipboard

Published 20 hours ago • verified publisher icon arq5x

Missing test data

Open mdshw5 opened this issue 11 years ago • 14 comments

I noticed that there is not test data containing a full Minion run. Could you please fix this?

mdshw5 avatar Jun 26 '14 01:06 mdshw5

Will do so once I have some data that I am allowed to share.

arq5x avatar Jun 26 '14 13:06 arq5x

+1 to this one. I may be able to add some shotgun metagenomic data in the near future, but not how sure that will be for instructional purposes.

stephenturner avatar Jul 31 '14 19:07 stephenturner

Just in case it wasn't obvious, this was only partially a joke issue.

mdshw5 avatar Jul 31 '14 19:07 mdshw5

must be dense - what's the joke? that minion/nanopore sequencing would be vaporware? not sure what MAP agreements look like, but as soon as i've confirmed i can release data i'll put some here.

stephenturner avatar Jul 31 '14 19:07 stephenturner

I will put some up, probably a full run will need to be hosted outside this repo though, as the files are quite large.

nickloman avatar Jul 31 '14 20:07 nickloman

"Full run" and "fix this" urgency were in jest. Seriously though, just part of a run would be fine for tool test data.

mdshw5 avatar Jul 31 '14 20:07 mdshw5

+1 would love to see what it looks like

alexbw avatar Sep 11 '14 12:09 alexbw

Looks like @nickloman released some yesterday so maybe it could be linked from here?

mdshw5 avatar Sep 11 '14 13:09 mdshw5

We could take a subset of these reads to serve for the basis of our test suite.

nickloman avatar Sep 11 '14 13:09 nickloman

I think that is a solid plan, but we want to either keep it very lightweight so that cloning the repo is easy, or we could have a test_data command that just downloads a more informaticve subset from Amazon S3. Or both.

arq5x avatar Sep 11 '14 14:09 arq5x

It looks like there is no testing framework in poretools currently, so maybe once there is it could be split:

  1. Unit tests that are for internal consistency and do not require external data
  2. Tests that require external data, and this data can be downloaded (and cached)

mdshw5 avatar Sep 11 '14 14:09 mdshw5

Yep @mdshw5 - that is the way to go. We'll get there. Nick and I are both in crunch meeting and grant deadline time, but after the dust settles, we will take care of it.

arq5x avatar Sep 12 '14 13:09 arq5x

+1 to this! Need some test data for a homebrew-science formula I've been working on anyway, so will try kill 2 birds with one stone :smile:

See https://github.com/Homebrew/homebrew-science/pull/2300.

UPDATE: I've added some test data to the homebrew poretools package. Will write some tests for this in due course.

gawbul avatar May 22 '15 22:05 gawbul

Matt Loose has some data from his Read Until paper: https://github.com/mattloose/RUscripts . Seems pretty small and all the scripts seem to work on it.

jeffhsu3 avatar Feb 05 '16 05:02 jeffhsu3