Length vs. Q and channels epb plots empty
Opened this issue · 13 comments
I ran the script on a recent dataset and mostly worked fine except that the two plots length_vs_q.png and flowcell_channels_epb.png don't have anything plotted on them.
What could be the issue?
Here's also the Albacore summary file if you want to give it a try:
https://owncloud.tuebingen.mpg.de/index.php/s/VvbpGIYuCqHMfv8
Thx!
Sorry for the radio silence - I've been away. Taking a look now.
Not sure what the problem was, but I downloaded your file and it seems to work with the latest update. Note that you need a couple more dependencies - details at the top of the readme.
I reran it, but these two plots still are empty for me. In addition, the new corresponding plots *_per_hour.png also don't look how they should (I suppose).
When I was running the script now, this also popped up:
Loading required package: viridisLite
Attaching package: ‘data.table’
The following objects are masked from ‘package:reshape2’:
dcast, melt
What version of each package are you running? Also what R version?
R-3.3.1
ggplot2 2.2.1
viridis 0.4.0
reshape2 1.4.1
plyr 1.8.4
ggjoy 0.2.0
purrr 0.2.2.2
I've just done a clean local Linux installation of R 3.4.1 from source and all needed packages, but the aforementioned plots still look the same or, in this case, empty.
Interestingly, I also have R 3.3.1 installed (from a .deb package) on this machine and when I used it the script worked fine.
I just tried it on a system wide installation of R 3.3.1 on a cluster and the the problematic plots are empty again.
Either the script is using a recent version of a package, which I just happen to have on my local installation, or the installation of R from source and the base package is missing something which the other R distributions have...
I'm not sure how to do this in R, but perhaps your script can log the packages used, their paths and versions? Could help to track which package causes the difference.
@roblanf Can you list the versions of your packages I mentioned above?
@wdecoster I added sessionInfo() to the script (which some argue might not be the most accurate way to do this, but whatever).
"Good"
R version 3.3.1 (2016-06-21)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 16.04.2 LTS
locale:
[1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
[3] LC_TIME=de_DE.UTF-8 LC_COLLATE=en_US.UTF-8
[5] LC_MONETARY=de_DE.UTF-8 LC_MESSAGES=en_US.UTF-8
[7] LC_PAPER=de_DE.UTF-8 LC_NAME=C
[9] LC_ADDRESS=C LC_TELEPHONE=C
[11] LC_MEASUREMENT=de_DE.UTF-8 LC_IDENTIFICATION=C
attached base packages:
[1] stats graphics grDevices utils datasets base
other attached packages:
[1] **data.table_1.10.4** scales_0.4.1 **ggjoy_0.3.0** yaml_2.1.14
[5] **reshape2_1.4** plyr_1.8.4 viridis_0.4.0 viridisLite_0.2.0
[9] ggplot2_2.2.1
loaded via a namespace (and not attached):
[1] **Rcpp_0.12.10** **digest_0.6.9** grid_3.3.1 **gtable_0.1.2**
[5] magrittr_1.5 **stringi_1.0-1** lazyeval_0.2.0 labeling_0.3
[9] tools_3.3.1 stringr_1.0.0 purrr_0.2.2.2 **munsell_0.4.2**
[13] **colorspace_1.2-4** methods_3.3.1 gridExtra_2.2.1 **tibble_1.3.0**
"Bad"
R version 3.3.1 (2016-06-21)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 16.04.2 LTS
locale:
[1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
[3] LC_TIME=en_US.UTF-8 LC_COLLATE=en_US.UTF-8
[5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
[7] LC_PAPER=en_US.UTF-8 LC_NAME=C
[9] LC_ADDRESS=C LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
attached base packages:
[1] stats graphics grDevices utils datasets base
other attached packages:
[1] **data.table_1.9.6** scales_0.4.1 **ggjoy_0.2.0** yaml_2.1.14
[5] **reshape2_1.4.1** plyr_1.8.4 viridis_0.4.0 viridisLite_0.2.0
[9] ggplot2_2.2.1
loaded via a namespace (and not attached):
[1] **Rcpp_0.12.6** **digest_0.6.10** **assertthat_0.1** **chron_2.3-47**
[5] grid_3.3.1 **gtable_0.2.0** magrittr_1.5 **stringi_1.1.1**
[9] lazyeval_0.2.0 labeling_0.3 tools_3.3.1 stringr_1.0.0
[13] purrr_0.2.2.2 **munsell_0.4.3** **colorspace_1.2-6** methods_3.3.1
[17] gridExtra_2.2.1 **tibble_1.1**
There is a version difference in quite a few packages (even some extra ones loaded in the "bad" R) though with some of the "bad" R packages being even more recent than the "good" R but I guess it's about pinning the right one down.
Thanks all for the tips and ideas. Here's my session info from my desktop mac (works fine):
R version 3.3.2 (2016-10-31)
Platform: x86_64-apple-darwin13.4.0 (64-bit)
Running under: OS X El Capitan 10.11.6
locale:
[1] en_AU.UTF-8/en_AU.UTF-8/en_AU.UTF-8/C/en_AU.UTF-8/en_AU.UTF-8
attached base packages:
[1] stats graphics grDevices utils datasets base
other attached packages:
[1] data.table_1.10.4 scales_0.4.1 ggjoy_0.2.0 yaml_2.1.14
[5] reshape2_1.4.2 plyr_1.8.4 viridis_0.4.0 viridisLite_0.2.0
[9] ggplot2_2.2.1
loaded via a namespace (and not attached):
[1] Rcpp_0.12.12 grid_3.3.2 gtable_0.2.0 magrittr_1.5
[5] rlang_0.1.1 stringi_1.1.5 lazyeval_0.2.0 tools_3.3.2
[9] stringr_1.2.0 munsell_0.4.3 colorspace_1.3-2 methods_3.3.2
[13] gridExtra_2.2.1 tibble_1.3.3
My bet so far - data.table. @DNAsaurus you have 1.9.6 in your 'bad' version, and in your good version and my good version we have 1.10.4
And on my server, the plots work fine. Note though that I just reinstalled the latest R and all the latest packages.
At least to me (obviously not done until @DNAsaurus can confirm empirically on their own system), this starts to suggest that getting all the latest packages might address this issue. Of course, it would be nice to know which package is the issue.
@DNAsaurus, maybe you could update your packages one-by-one on the 'bad' system, and see which update fixes the issue (if indeed this approach fixes it at all).
R version 3.4.1 (2017-06-30)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 16.04.3 LTS
Matrix products: default
BLAS: /usr/lib/libblas/libblas.so.3.6.0
LAPACK: /usr/lib/lapack/liblapack.so.3.6.0
locale:
[1] LC_CTYPE=en_AU.UTF-8 LC_NUMERIC=C
[3] LC_TIME=en_AU.UTF-8 LC_COLLATE=en_AU.UTF-8
[5] LC_MONETARY=en_AU.UTF-8 LC_MESSAGES=en_AU.UTF-8
[7] LC_PAPER=en_AU.UTF-8 LC_NAME=C
[9] LC_ADDRESS=C LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_AU.UTF-8 LC_IDENTIFICATION=C
attached base packages:
[1] stats graphics grDevices utils datasets base
other attached packages:
[1] data.table_1.10.4 scales_0.4.1 ggjoy_0.3.0 yaml_2.1.14
[5] reshape2_1.4.2 plyr_1.8.4 viridis_0.4.0 viridisLite_0.2.0
[9] ggplot2_2.2.1
loaded via a namespace (and not attached):
[1] Rcpp_0.12.12 grid_3.4.1 gtable_0.2.0 magrittr_1.5
[5] rlang_0.1.1 stringi_1.1.5 lazyeval_0.2.0 tools_3.4.1
[9] stringr_1.2.0 purrr_0.2.2.2 munsell_0.4.3 compiler_3.4.1
[13] colorspace_1.3-2 methods_3.4.1 gridExtra_2.2.1 tibble_1.3.3
I updated the packages one by one, at least to the version which works in "good" R where the plots are made correctly, or higher but there's no difference. The plots are still wrong with "bad" R...
The only other reason I can think of is that it might have to do something with those extra packages being called in "bad" R for whatever reason - assertthat_0.1 chron_2.3-47
I don't know what these exactly do or how they could be interferring.
If it's not this then there's something missing in the base packages because it doesn't make sense otherwise - the "bad" R now has all the packages which are at least the same or higher version (as yours). This would lead me to believe that it's not a package issue.
Hey @DNAsaurus, did you ever figure out more about what was going on here? If you have time, I'd be interested to know how you go with the latest version.