Unused arguments in crosstalk::filter_slider

[sigven]

Unused arguments in crosstalk::filter_slider

[sigven]

Ah thanks a lot for reporting! In order for me to reproduce the bug, would it be possible for you to share the variant set (VCF) with me somehow? And build (grch37/grch38) and config file?

best,
Sigve

[ridwan-shaikh]

Hi Sigven,

This was on your example vcf in the CPSR-0.5.0 build

[sigven]

Got it. Will test later! Thanks again.

best,
Sigve

[ridwan-shaikh]

Just looking at the help function for filter_slider, it looks like the argument 'min', 'max' does not exist. Maybe dragRange would work:

crosstalk::filter_slider("CLINVAR_REVIEW_STATUS_STARS", "ClinVar review status stars", variants_class5_1, ~CLINVAR_REVIEW_STATUS_STARS, ticks = T, dragRange = range(CLINVAR_REVIEW_STARS)) #not tested

regards,
Ridwan

[sigven]

Hm. A bit strange. Ran the example now, and it works (Dockerized). If you look at the commit comment for the man directory at https://github.com/rstudio/crosstalk, it seems it has been added to the help text also? But that seem to be committed quite long time a go..

Will make a push to dev for you to test on your local installation.

best,
Sigve

[sigven]

It might be that the install_no_docker version (i.e. conda) installs the crosstalk package from CRAN (1.0.0) and not the updated version on GitHub (1.0.1). Could that be the reason, @vladsaveliev?

[ridwan-shaikh]

Yes that looks to be the reason. I just had a look at the packages suggested in a fresh install of the conda environment it say: conda-forge/noarch::r-crosstalk-1.0.0-r35h6115d3f_1002. Curiously, I had an error of the following packages not existing when attempting to run for the first time.

• BSgenome.Hsapiens.UCSC.hg38
• BSgenome.Hsapiens.UCSC.hg19

Sorry I didn't take a screenshot at the time of the error as both packages a clearly listed in the yaml but for some reason are not downloaded/installed in the condo environment.

[vladsavelyev]

Huh, thanks for the forensics! I will modify the conda recipe to use the GitHub version of crosstalk (or just dump the conda version)

Regarding the BSGenome packages, I remember having some issues with them too. I will just install them directly in the recipe from bioconductor instead of using conda versions, that should solve the problem.

[vladsavelyev]

Hmm, it looks like though GitHub has 1.0.0 as well?

If PCGR needs some unreleased features, one way would be to build our own package for crosstalk under the pcgr channel. Happy to do that (I think I already did it once for the condarized version, before rstudio/crosstalk#56 was merged)

[vladsavelyev]

Alright, updated the recipes and pushed pcgr==0.8.3.1 and cpsr==0.5.1.2 into anaconda. This versions use Crosstalk from GitHub which fixes this issue. Also I'm not having the BSGenome errors, but let me know if it happens to you again.

To update:

conda install -c pcgr "pcgr>=0.8.3.1" "cpsr>=0.5.1.2"

[ridwan-shaikh]

Hi Vlad,

when attempting to update using:

conda install -c pcgr "pcgr>=0.8.3.1" "cpsr>=0.5.1.2"

$ conda install -c pcgr "pcgr>=0.8.3.1" "cpsr>=0.5.1.2"
Collecting package metadata: done
Solving environment: failed

PackagesNotFoundError: The following packages are not available from current channels:

  - cpsr[version='>=0.5.1.2'] -> r-ggpubr

Current channels:

  - https://conda.anaconda.org/pcgr/linux-64
  - https://conda.anaconda.org/pcgr/noarch
  - https://conda.anaconda.org/bioconda/linux-64
  - https://conda.anaconda.org/bioconda/noarch
  - https://repo.anaconda.com/pkgs/main/linux-64
  - https://repo.anaconda.com/pkgs/main/noarch
  - https://repo.anaconda.com/pkgs/free/linux-64
  - https://repo.anaconda.com/pkgs/free/noarch
  - https://repo.anaconda.com/pkgs/r/linux-64
  - https://repo.anaconda.com/pkgs/r/noarch

Does this indicate that conda-forge channel does not exist as a list of channels?

So I installed ggpubr:

conda install -c conda-forge r-ggpubr
Collecting package metadata: done
Solving environment: done

and then tried updating cpsr

$ conda install -c pcgr "cpsr>=0.5.1.2"
Collecting package metadata: done
Solving environment: failed


UnsatisfiableError: The following specifications were found to be in conflict:
  - cpsr[version='>=0.5.1.2']
Use "conda search <package> --info" to see the dependencies for each package.

$ conda search cpsr --info
Loading channels: done
No match found for: cpsr. Search: *cpsr*
r-icpsrdata 0.4.0 r36h6115d3f_0
-------------------------------
file name   : r-icpsrdata-0.4.0-r36h6115d3f_0.tar.bz2
name        : r-icpsrdata
version     : 0.4.0
build       : r36h6115d3f_0
build number: 0
size        : 428 KB
license     : MIT
subdir      : noarch
url         : https://repo.anaconda.com/pkgs/r/noarch/r-icpsrdata-0.4.0-r36h6115d3f_0.tar.bz2
md5         : 0470d469c698fef23fb71d11e9e0e1d5
timestamp   : 2019-09-12 21:29:34 UTC
dependencies: 
  - r-base >=3.6,<3.7.0a0
  - r-httr
  - r-purrr
  - r-rvest
  - _r-mutex 1.* anacondar_1

but cpsr is listed when running conda list

[vladsavelyev]

Ah, sorry @ridwan-shaikh, I already have conda-forge and bioconda in my .condrc and that's why didn't bother adding them into the command line. The correct one should be:

conda install -c pcgr -c conda-forge -c bioconda "pcgr>=0.8.3.1" "cpsr>=0.5.1.2"

Worked for me, can you try? If it errors with UnsatisfiableError again, can you try installing into a clean env please? (conda create -n pcgr -c pcgr -c conda-forge -c bioconda "pcgr>=0.8.3.1" "cpsr>=0.5.1.2") Just for a clean experiment.

UPD: Though hold on a sec, I have a problem with this package version still calling the older crosstalk, looking into it now

[ridwan-shaikh]

Hi Vlad,

yes adding the correct channels helps to update to the latest cpsr version.

For testing purposes i tried a fresh install too, to see what crosstalk is added and you're right, cross talk remains at v1.0.0.

[vladsavelyev]

Sorted it out. Conda details. Turns out I need to update crosstalk in post-link.sh script instead of build.sh, to make sure it's executed on a host machine. Anyway. Try updating? :)

conda install -c pcgr -c conda-forge -c bioconda "pcgr==0.8.3.2" "cpsr==0.5.1.3"

[ridwan-shaikh]

Hi Vlad,

Running the above command results in:

added / updated specs:
    - cpsr==0.5.1.3
    - pcgr==0.8.3.2


The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    cpsr-0.5.1.3               |                0          10 KB  pcgr
    ------------------------------------------------------------
                                           Total:          10 KB

The following NEW packages will be INSTALLED:

  cpsr               pcgr/linux-64::cpsr-0.5.1.3-0
  r-ggpubr           conda-forge/noarch::r-ggpubr-0.2.3-r35h6115d3f_0
  r-ggrepel          conda-forge/linux-64::r-ggrepel-0.8.1-r35h0357c0b_1
  r-ggsci            conda-forge/noarch::r-ggsci-2.9-r35h6115d3f_1002
  r-ggsignif         conda-forge/noarch::r-ggsignif-0.6.0-r35h6115d3f_0
  r-polynom          conda-forge/noarch::r-polynom-1.4_0-r35h6115d3f_1
  r-upsetr           conda-forge/noarch::r-upsetr-1.4.0-r35h6115d3f_1

and conda list results in:

r-crosstalk               1.0.0           r35h6115d3f_1002    conda-forge

Does not look like it has updated crosstalk. To get around it, I have just been installing the github version of crosstalk through R

devtools::install_github("https://github.com/rstudio/crosstalk.git")

[vladsavelyev]

Ridwan,

conda list will show 1.0.0 because 1.0.1 does not exist on conda (and neither on cran - it wasn't released). My change to the pcgr package was to actually run devtools::install_github("https://github.com/rstudio/crosstalk.git") automatically in the end of the installation. So 1.0.1 should be reflected if you run R -e "library(crosstalk); sessionInfo();" (regardless of the conda version). Though if it didn't work for you, I might have screwed something up and need to investigate further. Wondering if you could try installing the package a new env and checking sessionInfo?

Also I'm wondering why did it only updated cpsr, but not the pcgr package as well. Did you update the packages in 2 steps by any chance?