Merge pull request #43 from menchelab/v1.1.1

v1.1.1

.github/workflows/CI.yml

@@ -28,7 +28,7 @@ jobs:
     steps:
       - uses: actions/checkout@v2
       - uses: actions/checkout@v2
-      - uses: r-lib/actions/setup-r@v1
+      - uses: r-lib/actions/setup-r@v2
       - run: echo "LD_LIBRARY_PATH=$(R RHOME)/lib" >> $GITHUB_ENV
       - run: sudo R -e "install.packages('robustbase', repos='http://cran.rstudio.com/')"
       - uses: julia-actions/setup-julia@latest

.github/workflows/NightlyCI.yml

@@ -27,7 +27,7 @@ jobs:
     steps:
       - uses: actions/checkout@v2
       - uses: actions/checkout@v2
-      - uses: r-lib/actions/setup-r@v1
+      - uses: r-lib/actions/setup-r@v2
       - run: echo "LD_LIBRARY_PATH=$(R RHOME)/lib" >> $GITHUB_ENV
       - run: sudo R -e "install.packages('robustbase', repos='http://cran.rstudio.com/')"
       - uses: julia-actions/setup-julia@latest

.gitignore

@@ -2,3 +2,4 @@
 /Manifest.toml
 /dev/
 docs/build/
+.julia

CHANGELOG.md

@@ -4,6 +4,15 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/)
 and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.html).
 
+## v1.1.1 - 2022-11-30
+### Changed
+- Change R repository for increased compatibility
+- Update links in documentation
+
+### Added
+- Documentation on codespaces
+- Tests for TIF-related methods
+
 ## v1.1.0 - 2021-09-23
 ### Added
 - Option to remove R seed from distance computation

CONTRIBUTING.md

@@ -32,7 +32,8 @@ All tests should pass for the supported versions of Julia and the package's depe
 
 ## Development environment
 
-BioProfiling.jl is intended to be compatible with a large set of environments as detailed in [Project.toml](Project.toml). If you need to setup a development environment, an easy solution is to start from the [latest Docker image designed to run the example notebooks](https://hub.docker.com/r/koalive/bioprofilingnotebooks), which comes with all dependencies installed. To avoid inteferences between the pre-installed version of BioProfiling.jl and your local version of the source code, you can remove the version installed in the container with the following lines:
+BioProfiling.jl is intended to be compatible with a large set of environments as detailed in [Project.toml](Project.toml). If you need to setup a development environment, an easy solution is to **edit and run code directly in your browser thanks to GitHub codespaces**. Some instructions are available in [our documentation](https://menchelab.github.io/BioProfiling.jl/dev/).  
+Another option is to start from the [latest **Docker image** designed to run the example notebooks](https://hub.docker.com/r/koalive/bioprofilingnotebooks), which comes with all dependencies installed. To avoid inteferences between the pre-installed version of BioProfiling.jl and your local version of the source code, you can remove the version installed in the container with the following lines:
 
 ```julia
 import Pkg

Project.toml

@@ -1,7 +1,7 @@
 name = "BioProfiling"
 uuid = "cef322dc-4d82-11ea-04a7-113231db804d"
 authors = ["Loan Vulliard <lvulliard@cemm.oeaw.ac.at>"]
-version = "1.1.0"
+version = "1.1.1"
 
 [deps]
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"

README.md

@@ -57,7 +57,7 @@ using BioProfiling
 * Have a look at the [documentation](https://menchelab.github.io/BioProfiling.jl/dev/)
 * Run our [example analysis notebooks](https://github.com/menchelab/BioProfilingNotebooks)
 * Check out the [JuliaCon 2021 poster](https://drive.google.com/file/d/1sjRONQ8dRJDGAiR-wBhC_rEBKiMIs5Rh/preview) presenting BioProfiling.jl 
-* Read our [preprint on BioRxiv](https://www.biorxiv.org/content/10.1101/2021.06.18.448961v1)
+* Read our [article published in Bioinformatics](https://doi.org/10.1093/bioinformatics/btab853)
 
 ## Contribute
 
@@ -67,10 +67,10 @@ We welcome all sorts of contributions to this project! See our [contribution gui
 
 This package was created by [Loan Vulliard](http://vulliard.loan) @ [Menche lab](https://menchelab.com/).  
 BioProfiling.jl relies on several amazing open-source Julia packages, listed in the requirement file (see [*Project.toml*](Project.toml)).  
-If you use this tool in your research work, please cite the [preprint](https://www.biorxiv.org/content/10.1101/2021.06.18.448961v1) in which we detail how this tool is implemented and can be used:
+If you use this tool in your research work, please cite [the article](https://doi.org/10.1093/bioinformatics/btab853) in which we detail how this tool is implemented and can be used:
 
-    BioProfiling.jl: Profiling biological perturbations with high-content imaging in single cells and heterogeneous populations
-    Loan Vulliard, Joel Hancock, Anton Kamnev, Christopher W. Fell, Joana Ferreira da Silva, Joanna Loizou, Vanja Nagy, Loïc Dupré, Jörg Menche
-    bioRxiv 2021.06.18.448961; doi: https://doi.org/10.1101/2021.06.18.448961 
+    BioProfiling.jl: profiling biological perturbations with high-content imaging in single cells and heterogeneous populations
+    Loan Vulliard, Joel Hancock, Anton Kamnev, Christopher W Fell, Joana Ferreira da Silva, Joanna I Loizou, Vanja Nagy, Loïc Dupré, Jörg Menche
+    Bioinformatics 2022; 38: 1692–9. DOI: https://doi.org/10.1093/bioinformatics/btab853
 
-We also thank the reviewers of this manuscript whose suggestions contributed to improve the [example analyses](https://github.com/menchelab/BioProfilingNotebooks), their biological interpretation as well as the package in itself.
+We also thank the reviewers whose suggestions contributed to improve the [example analyses](https://github.com/menchelab/BioProfilingNotebooks), their biological interpretation as well as the package in itself.

docs/src/codespaces.md

@@ -0,0 +1,28 @@
+# BioProfiling.jl examples
+
+This section aim to describe how to develop and test the BioProfiling.jl package using a virtual machine hosted by GitHub and VS directly in your browser thanks to GitHub codespaces.  
+Please note that no paid account is typically needed, although you might have limited access to codespaces depending on which type of GitHub account you have (see [Pricing information](https://docs.github.com/en/billing/managing-billing-for-github-codespaces/about-billing-for-github-codespaces)). 
+
+## Starting a codespace and setting up the environment
+
+Go to the [repository](https://github.com/menchelab/BioProfiling.jl), on the branch you would like to edit. Click on "Code <>" > "Codespaces" > "+". This will open VS, where you can edit the files, run code, and contribute to the repository using all usual GitHub features.  
+First, you will need to start a Docker container to be able to run Julia code. In the terminal, run `docker run -it -v `pwd`:/home/jovyan jupyter/datascience-notebook:latest julia`. The environment should include both R and Julia, although you might want to use [another image](https://hub.docker.com/r/jupyter/datascience-notebook/tags) if you want a version in particular.
+
+## Setting up Julia for package development
+
+The terminal should display a Julia session. Now, set up the package as follows:
+```julia
+import Pkg
+Pkg.activate(".") # Set up requirements for environment
+Pkg.status() # This should display the dependencies of the package
+
+Pkg.instantiate() # Actually prepare all packages
+# You can now modify what you and test your new features!
+
+# In the end, do not forget to test your code to make sure you did not break anything
+Pkg.test() 
+```
+
+## Opening a pull request
+
+Once you are satisfied with your changes and made sure that are not breaking existing features, go to the "Source Control" tab to commit your changes. Typically you will want to create a pull request with your code. Please follow the [contribution guidelines](https://github.com/menchelab/BioProfiling.jl/blob/master/CONTRIBUTING.md) and [code of conduct](https://github.com/menchelab/BioProfiling.jl/blob/master/CODE_OF_CONDUCT.md). Thanks for taking the time to contribute to BioProfiling.jl!

src/rmpv.jl

@@ -113,7 +113,7 @@ function distance_robust_mahalanobis_median(data, indpert, indref; r_seed = true
 
     R"""
     if (!require("robustbase")) install.packages("robustbase", 
-                                repos = "https://cloud.r-project.org")
+                                repos = "http://cloud.r-project.org")
     library(robustbase)
 
     if (r_seed){set.seed(777)}
@@ -161,7 +161,7 @@ function shuffled_distance_robust_mahalanobis_median(data, indpert, indref; nb_r
 
         R"""
         if (!require("robustbase")) install.packages("robustbase", 
-                                    repos = "https://cloud.r-project.org")
+                                    repos = "http://cloud.r-project.org")
         library(robustbase)
 
         if (r_seed){set.seed(3895)}
@@ -202,7 +202,7 @@ function distance_robust_hellinger(data, indpert, indref; r_seed = true)
 
     R"""
     if (!require("robustbase")) install.packages("robustbase", 
-                                repos = "https://cloud.r-project.org")
+                                repos = "http://cloud.r-project.org")
     library(robustbase)
 
     if (r_seed){set.seed(777)}
@@ -261,7 +261,7 @@ function shuffled_distance_robust_hellinger(data, indpert, indref; nb_rep = 250,
 
         R"""
         if (!require("robustbase")) install.packages("robustbase", 
-                                    repos = "https://cloud.r-project.org")
+                                    repos = "http://cloud.r-project.org")
         library(robustbase)
 
         if (r_seed){set.seed(777)}

test/runtests.jl

@@ -8,6 +8,8 @@ using Random
 using RCall
 using Distributed
 using ParallelDataTransfer
+using Images
+
 
 @testset "freqtable" begin
 	d = DataFrame(Any[0.05131 0.32830 "Exp1"; 0.83296 0.97647 "Exp1"; 0.66463 0.66939 "Exp2"; 
@@ -278,12 +280,38 @@ end
 	@test diagnostic_images(e1, cf1, :Experiment, rgx = [r".*" => s"example.png"], saveimages = false)
 	# Additional checks that could be performed:
 	# Centers of diagnostic_path
-	# getColorImage [internal]
-	# colimgifrgb [internal]
 	# rgb parameters of diagnostic_images
 	# output of diagnostic_images
 end
 
+@testset "tiff" begin
+	# Test TIF images at 8, 16 or 32 bits, RGB or single channel
+    i8 = load("example8b.tiff")
+    i16 = load("example16b.tiff")
+    i32 = load("example32b.tiff")
+	pathbw = "example32b_bw.tiff"
+    i32bw = load(pathbw)
+
+	# Images are not empty
+	@test all([sum(channelview(x)) > 0 for x in [i8, i16, i32, i32bw] ])
+	# Black-and-white image should correspond to the second channel of i32
+	@test channelview(i32bw) == channelview(i32)[2,:,:]
+
+	i32_fakergb = BioProfiling.getColorImage(pathbw, pathbw, pathbw)
+	@test channelview(i32_fakergb)[1,:,:] == channelview(i32_fakergb)[2,:,:] == channelview(i32_fakergb)[3,:,:]
+	# All pixel intensities should be normalized by the same amount
+	ratio_intensity = float(channelview(i32_fakergb)[3,6,6]) ./ channelview(i32bw)[6,6]
+	@test all([isapprox(channelview(i32bw)[x,y] * ratio_intensity, 
+				  channelview(i32_fakergb)[3,x,y], 
+				  atol = 2/(2^32))
+		 for x in 1:32 for y in 1:96])
+
+	@test_throws ArgumentError BioProfiling.colimgifrgb("not_a_path", nothing)
+	# Should load images as in getColorImage and support regex for RGB images
+	@test BioProfiling.colimgifrgb(pathbw, nothing) == i32bw
+	@test BioProfiling.colimgifrgb(pathbw, ["", "", ""]) == i32_fakergb
+end
+
 @testset "interpret" begin
 	# Define example dataset
 	d = DataFrame([[0,2,4,6,8,21],[0,1,2,3,4,0],[0,3,6,9,12,3],'A':'F'])
@@ -605,7 +633,7 @@ end
 	# Robustbase must be installed
 	R"""
 	if (!require("robustbase")) install.packages("robustbase", 
-												  repos = "https://cloud.r-project.org")
+												  repos = "http://cloud.r-project.org")
 	library(robustbase)
 	"""
 	d = rand(100,5)
@@ -764,16 +792,23 @@ end
 end
 
 @testset "parallel_rmpv" begin
+	"""
+	Warning: this is resource intensive!
+	Running it on the default 2- or 4-core machines in Github 
+	codespaces lead to an error (a KILL signal is sent). 
+	This should work well with 8-core machines.
+	You might want to skip it in your routine checks.
+	It is best including it in the actual checks in CI anyway.
+	"""
 	d = DataFrame(rand(100,5))
 	d.Condition = sample('A':'D', 100);
 
 	e = Experiment(d)
 
-    addprocs(4)
+    addprocs(4, exeflags="--project=$(Base.active_project())")
     pool = CachingPool(workers())
     @everywhere using BioProfiling
 
-
 	slt = NameSelector(x -> x != "Condition")
 	select_features!(e, slt)
 	f = Filter('C', :Condition)
@@ -785,4 +820,4 @@ end
 
 	# 4 conditions, 3 columns
 	@test size(rmpv) == (4,3)
-end
+end