Dockerize shiny_prep

runs/shiny_prep/README.md

@@ -8,14 +8,6 @@ cd /Users/dominikbrilhaus/gitlab_dataplant/samplearc_rnaseq/runs/kallisto_sleuth
 ## Let it flow
 
 ```bash
-### store arc root (two levels up from here) as variable
-arc_root=$(echo ${PWD%/*/*})
-
-### replace arc root line in yml (specific to the machine from where this is run)
-### not sure, if this works on linux... 
-sed -i '' "s|^arc_root:.*|arc_root: $arc_root|g" kallisto_sleuth.yml
-
 ### run with cwltool
-cwltool ../../workflows/kallisto_sleuth.cwl kallisto_sleuth.yml
-
+cwltool --enable-dev run.cwl shiny_prep.yml
 ```

runs/shiny_prep/run.cwl

+#!/usr/bin/env cwl-runner
+cwlVersion: v1.2.0-dev1
+class: Workflow
+inputs:
+  out_folder:
+    type: string
+  in_kallisto_df:
+    type: File
+outputs:
+  out_dir:
+    type:
+      type: array
+      items: Directory
+    outputSource: shiny_prep/outdir
+steps:
+  shiny_prep:
+    run: ../../workflows/shiny_prep/workflow.cwl
+    in:
+      in_kallisto_df: in_kallisto_df
+      out_folder: out_folder
+    out: [outdir]

runs/shiny_prep/shiny_prep.yml

 cores: 1
-r_script:
+in_kallisto_df:
   class: File
-  path: ../../workflows/shiny_prep.R
-in_kallisto_df: runs/kallisto_collect/kallisto_df.csv
-out_folder: runs/shiny_prep
-arc_root: /Users/dominikbrilhaus/gitlab_dataplant/samplearc_rnaseq
+  path: ../kallisto_collect/kallisto_df.csv
+out_folder: out

workflows/install_dependencies/install_cwl.md

-
-# CWL Installation (MacOS)
-
-Docs: https://www.commonwl.org/
-GitHub: https://github.com/common-workflow-language/cwltool
-
-## Install CWL via conda
-
-> requires conda installation, which is a whole different story. Good luck.
-
-```bash
-conda install -c conda-forge cwltool
-```

workflows/install_dependencies/install_kallisto.md

-
-# Kallisto Installation
-
-- Docs: http://pachterlab.github.io/kallisto/manual.html
-- GitHub: https://github.com/pachterlab/kallisto
-
-## MacOS Installation via conda
-
-see: https://anaconda.org/bioconda/kallisto
-
-```bash
-conda install -c bioconda kallisto
-```

workflows/install_dependencies/install_r_packages.R

-#!/usr/bin/env Rscript
-
-################################################
-### Installation of R dependencies #############
-################################################
-
-### Default (CRAN)
-
-if(!"tidyverse" %in% row.names(installed.packages())){install.packages("tidyverse")}
-if(!"jsonlite" %in% row.names(installed.packages())){install.packages("jsonlite")}
-if(!"openxlsx" %in% row.names(installed.packages())){install.packages("openxlsx")}
-if(!"kableExtra" %in% row.names(installed.packages())){install.packages("kableExtra")}
-if(!"shiny" %in% row.names(installed.packages())){install.packages("shiny")}
-if(!"RColorBrewer" %in% row.names(installed.packages())){install.packages("RColorBrewer")}
-
-### From BiocManager
-
-if (!requireNamespace("BiocManager", quietly = TRUE)){install.packages("BiocManager")}
-if(!"BiocManager" %in% row.names(installed.packages())){BiocManager::install()}
-if(!"devtools" %in% row.names(installed.packages())){BiocManager::install("devtools")}
-if(!"sleuth" %in% row.names(installed.packages())){BiocManager::install("pachterlab/sleuth")}
-
-### Test load packages
-
-library(tidyverse)
-library(jsonlite)
-library(openxlsx)
-library(sleuth)
-library(kableExtra)
-library(shiny)
-library(RColorBrewer)

workflows/shiny_prep/shiny_prep.R

@@ -20,22 +20,21 @@
 
 args <- commandArgs(trailingOnly = T)
 
-arc_root <- args[1]
-out_folder <- args[2]
-in_kallisto_df <- args[3]
+out_folder <- args[1]
+in_kallisto_df <- args[2]
 
 ################################################
 #### If it does not exist, create out dir
 ################################################
 
-dir.create(paste(arc_root, out_folder, sep = "/"), recursive = T, showWarnings = F)
+dir.create(out_folder, recursive = T, showWarnings = F)
 
 ################################################
 #### Prep data for shiny app
 ################################################
 
-expression_data <- read.csv(file = paste(arc_root, in_kallisto_df, sep = "/"))
+expression_data <- read.csv(file = in_kallisto_df)
 
 available_genes <- unique(expression_data$target_id)
 
-save(expression_data, available_genes, file = paste(arc_root, out_folder, 'shiny_prep.RData', sep = "/"))
+save(expression_data, available_genes, file = paste(out_folder, 'shiny_prep.RData', sep = "/"))

workflows/shiny_prep/workflow.cwl

 #!/usr/bin/env cwl-runner
 
-cwlVersion: v1.2
+cwlVersion: v1.2.0-dev1
 class: CommandLineTool
-
+hints:
+  DockerRequirement:
+    dockerPull: zimmera95/rnaseq:latest
+requirements:
+  - class: InitialWorkDirRequirement
+    listing:
+      - class: File
+        location: shiny_prep.R
+arguments:
+  - position: 0
+    valueFrom: shiny_prep.R
 inputs:
-- id: r_script
-  type: File
-  inputBinding:
-    position: 0
-- id: arc_root
-  type: string
-  inputBinding:
-    position: 1
 - id: out_folder
   type: string
   inputBinding:
-    position: 2
+    position: 1
 - id: in_kallisto_df
-  type: string
+  type: File
   inputBinding:
-    position: 3
+    position: 2
 
 outputs:
 - id: outdir