sbg / sevenbridges-r Goto Github PK

Add a brief style guide to unify R code style.

This could lead to a more detailed contributing guide, eventually.

e.g.
https://github.com/tidyverse/ggplot2/blob/master/CONTRIBUTING.md

Update content with new look.

Maybe via http://hadley.github.io/pkgdown/

Now we have 5 tutorial, we still need to improve function manual

1. For reference class add help

Check if the API interface supported in core.R are synced with API v2.

Add, delete, or modify necessary parts to ensure consistency.

right now, users have to get files object by id first, support this will allow internal conversion and easier usage.

no depends on pandoc

Still in advance access but should support in the near future:
http://docs.sevenbridges.com/reference#folders-overview

1. inheritFromMetadata

should autocheck "#"

1. Add example in app.Rmd

Such as

or

-- while may need to remove one previous badge to save horizontal space.

Though it's advanced access feature, still need to be better integrated with public API and user flow when executing a task and use a query as task input.

This task is aimed for

1. removing all trailing spaces and unnecessary blank lines in all relevant files, especially R files.
2. updating the project configuration file .Rproj to make this the default setting if people contribute with RStudio.

This option is hidden in RStudio:

Single project-wise:
Build - Configure build tools - Code editing - Strip trailing horizonal whitespaces when saving

Global-wise:
Tools - Global options - Code - Saving - Strip trailing horizonal whitespaces when saving

API support task batching in R, but you can not add batch input node for Flow directly in R now.

We use different setup, should support both so user don't have to provide two different file. But on the other hand, current config setting allow users to maintain multiple users and platforms information.

what is a manifest file
http://docs.sevenbridges.com/docs/format-of-a-manifest-file

Supported in GUI and command line uploader, so in R API client we want to do

• interface allow you to specify which meta to keep and to ignore, by default, upload all meta with files
  p$upload(manifest_file = , manifest_metadata = TRUE, meta.keep = NULL, meta.ignore = NULL, verbal = FALSE) # with metadata = TRUE by default
• Validation: check file exists or not and stop if not all satisfied, assumption here is that user provide a manifest, it has to work and all files need to exists.
• Uploading files: print files info in verbal mode otherwise just progress bar

• Move badges
• Add logo image
• Clearer & cleaner hierarchy

e.g.
https://github.com/tidyverse/ggplot2/blob/master/README.md

Can't have 'require(sevenbridges)' as part of a shinyapp. It will not load.

e.g.

tidyverse/ggplot2@a0c9ba9

http://docs.sevenbridges.com/docs/creating-a-volume

Add AppVeyor CI for building & checking on Windows:

https://github.com/krlmlr/r-appveyor

which I think is essential now. New badge expected.

more conventient

Not sure about Bioc current status but CRAN and GitHub will render NEWS.md to HTML.

I test multiple revisions of a tool on CGC/SBG platform. Every time the revision is pushed the "Test" tab is blanked out and I have to re-input to test the command line expression.

Would be great if I could add the "Test" information via R script.

Write an interactive RStudio addin for generating tools. https://rstudio.github.io/rstudioaddins/

Hi,

When installing from github:

source("http://bioconductor.org/biocLite.R")
useDevel()
biocLite("BiocUpgrade")
library(devtools)
install_github("tengfei/sevenbridges", build_vignettes=TRUE, 
  repos=BiocInstaller::biocinstallRepos(),
  dependencies=TRUE, type = "source")

I get an error:

Quitting from lines 155-183 (bioc-workflow.Rmd) 
Error: processing vignette 'bioc-workflow.Rmd' failed with diagnostics:
invalid assignment for reference class field 'token', should be from class "character" or a subclass (was class "NULL")
Execution halted

It is a tiny issue - pull request that fixes it is #1

p.s. great work btw.

To solve libssl and related build issues.

Hi TengFei,
I followed your tutorials and tried to connect to SB-CGC from RStudio, but got the following error:

a <- Auth(token = "(my token here)", url = "https://cgc-api.sbgenomics.com/v2/")
Error: could not find function "Auth"

tx,
Durga

When I follow the tutorial for the R API in sbg (adapted for Cavatica, http://docs.sevenbridges.com/v1.0/reference#api-r-library, I get the following error Error in envRefSetField(.Object, field, classDef, selfEnv, elements[[field]]) : when creating a task (see screenshot).

This prevents the tsk object from being created. The task is still running, but I get this error message each time I handle the tsk object and it’s not possible to download it.

Any idea of what is happening there?

Need to sync misc_get_uploader(), misc_upload_cli() with the latest command line uploader:

http://docs.sevenbridges.com/docs/upload-via-the-command-line

Public API related validation question

• when file array specified, should I pass a single file a list of a single file
• when file[not array] specified, can I pass a single file or a list of a single file

need to test and need clear documentation, or make the R API extremely simple so user won't think about it or worry about it at all.

recommended by Dusan, very good point, also see examples from Dave as well, need good example for spin.

This is great! It would be super nice if CWL draft-3 was supported.

This is caused by reference class, hard to debug, but let me try to fix it by April 20.

Should be publicly searchable, with easy get/post access.

Maybe via Google Groups.

Possible solution:

• Change default to: drop undefined field, and print warning messages

Question:

• Do I need to keep the raw list or new field in Tool/Flow object

Currently t1 %>>% t2 only support simple flow simple matching, but we need flexible extra parameter in Flow function, that allow user to connect by id directly, so they can create non-linear or complex flow from R script easily.

CSS, logo, fonts.

search by name feature is convenient, but also limited by default limit/offset

user need to get more clear message about current limit/offset

from R markdown with headers into

1. command line script
2. Dockerfile
3. cwl json file
4. combined with API client, when token provided, push to repos you specified. should I support any public repos config?

How we make it better integrated with bunny? Can we run local test from bunny for our CWL Tool or Flow object? This enhancement is going to be easier for developers to test their app with test data (maybe in the IDE container?) need more discussion and brainstorm on this.

splits the file name by character into several parts after push to platform

not working on tengfei/sevenbridges docker images on docker hub

It might look better in vignettes:
docker pull sevenbridges/sevenbridges.

Maybe we need a new image name instead
of sevenbridges? cc @tengfei