Fix #49 error of caused by Windows GCC compiler (#50)

* Remove igraph dependency

* Fix windows installation of Rust library

* Remove unicode symbols not compatible with LaTeX

* remove vendor folder

* force linker for windows to be rust-lld

* improved rust arch detection

* remove forced linker

* revert change of rust arch detection script

* revert change of rust arch detection script

* add empty libgcc_eh.a

* fix(win): fix libgcc_mock folder not being created

* remove documentation page

DESCRIPTION

@@ -20,7 +20,7 @@ Depends: R (>= 4.0)
 Imports: methods, dplyr, doRNG, readr, parallel (>= 3.3.2),
         doParallel (>= 1.0.10), foreach (>= 1.4.0),
         jsonlite, httr, rlang, svglite,
-        igraph, whisker, apcluster, Rcpp, cluster, poolr
+        whisker, apcluster, Rcpp, cluster, poolr
 NeedsCompilation: yes
 LinkingTo: Rcpp
 RoxygenNote: 7.3.1

NAMESPACE

@@ -52,9 +52,6 @@ importFrom(httr,GET)
 importFrom(httr,POST)
 importFrom(httr,content)
 importFrom(httr,modify_url)
-importFrom(igraph,V)
-importFrom(igraph,get.adjacency)
-importFrom(igraph,graph.edgelist)
 importFrom(jsonlite,fromJSON)
 importFrom(jsonlite,toJSON)
 importFrom(parallel,makeCluster)

R/extendr-wrappers.R

@@ -18,13 +18,13 @@ NULL
 #'
 #' ## Diagram
 #' ```shell
-#'         Gene Sets
-#'       ┌───────────┐  First column named 'gene' containing gene name
-#'       │A0100110100│  1 = in set
-#' Genes │B0100101000│  0 = not in set
-#'       │C1011101001│  Due to limitiations with extendr-api v 0.6.0,
-#'       └───────────┘  function returns a list, and the R package will
-#'                      add the first 'gene' column
+#'                    Gene Sets
+#'                    First column named 'gene' containing gene name
+#'       A0100110100  1 = in set
+#' Genes B0100101000  0 = not in set
+#'       C1011101001  Due to limitiations with extendr-api v 0.6.0,
+#'                    function returns a list, and the R package will
+#'                    add the first 'gene' column
 #' ```
 #' @param gmt A Data Frame with geneSet and gene columns from the GMT file
 #' @param genes A vector of genes

R/randomWalkEnrichment.R

@@ -1,7 +1,6 @@
 #' @importFrom httr content modify_url
 #' @importFrom readr read_tsv
 #' @importFrom dplyr filter arrange %>% desc
-#' @importFrom igraph graph.edgelist V
 randomWalkEnrichment <- function(organism, network, method, inputSeed, topRank, highlightSeedNum, sigMethod, fdrThr, topThr, projectDir, projectName, cache, hostName) {
     fileName <- paste(projectName, network, method, sep = ".")
     if (startsWith(hostName, "file://")) {
@@ -20,8 +19,6 @@ randomWalkEnrichment <- function(organism, network, method, inputSeed, topRank,
         gmtUrl <- modify_url(geneSetUrl, query = list(organism = organism, database = "geneontology_Biological_Process", standardId = "genesymbol", fileType = "gmt"))
         goAnn <- readGmt(gmtUrl, cache = cache)
     }
-    # netGraph <- graph.edgelist(net, directed=FALSE)
-    # netNode <- V(netGraph)$name
     unique_nodes <- unlist(unique(c(net[, 1], net[, 2])))
 
 
@@ -44,10 +41,7 @@ randomWalkEnrichment <- function(organism, network, method, inputSeed, topRank,
     for (i in 1:nrow(net)) {
         organize_net[[i]] <- c(net[i, 1], net[i, 2])
     }
-    # pt1 <- .netwalker(seeds, netGraph, r=0.5)
-    # print(head(seeds))
     random_walk_res <- nta_rust(organize_net, seeds)
-    # print(head(random_walk_res))
     gS <- data.frame(name = random_walk_res$nodes, score = random_walk_res$scores, per = 1, stringsAsFactors = F)
 
     if (method == "Network_Expansion") {
@@ -110,27 +104,6 @@ randomWalkEnrichment <- function(organism, network, method, inputSeed, topRank,
     return(termInfo)
 }
 
-#' @importFrom igraph get.adjacency
-.netwalker <- function(seed, network, r = 0.5) {
-    adjMatrix <- get.adjacency(network, sparse = FALSE)
-    de <- apply(adjMatrix, 2, sum)
-    w <- t(t(adjMatrix) / de)
-
-    p0 <- array(0, dim = c(nrow(w), 1))
-    rownames(p0) <- rownames(w)
-    p0[seed, 1] <- 1 / length(seed)
-
-    pt <- p0
-    pt1 <- (1 - r) * (w %*% pt) + r * p0
-
-    while (sum(abs(pt1 - pt)) > 1e-6) {
-        pt <- pt1
-        pt1 <- (1 - r) * (w %*% pt) + r * p0
-    }
-
-    return(pt1)
-}
-
 #' @importFrom dplyr select filter arrange left_join mutate %>%
 #' @importFrom httr POST content
 #' @importFrom readr read_tsv

src/Makevars.win.in

@@ -1,60 +1,23 @@
-TARGET = $(subst 64,x86_64,$(subst 32,i686,$(WIN)))-pc-windows-gnu
-
-TARGET = x86_64-pc-windows-gnu
-
-# Rtools42 doesn't have the linker in the location that cargo expects, so we
-# need to overwrite it via configuration.
-CARGO_LINKER = x86_64-w64-mingw32.static.posix-gcc.exe
-
-VENDORING = @VENDORING@
-OFFLINE_OPTION = @OFFLINE_OPTION@
-
-
+TARGET = $(shell "$(R_HOME)/bin$(R_ARCH_BIN)/Rscript.exe" "../tools/rustarch.R")
+LIBDIR = rust/target/$(TARGET)/release
+STATLIB = $(LIBDIR)/WebGestaltR.a
+PKG_LIBS = -L$(LIBDIR) -lWebGestaltR -lws2_32 -ladvapi32 -lgdi32 -lbcrypt -lcrypt32 -luserenv -lntdll
 TARGET_DIR = rust/target
-LIBDIR = $(TARGET_DIR)/$(TARGET)/release
-STATLIB = $(LIBDIR)/libWebGestaltR.a
-PKG_LIBS = -L$(LIBDIR) -lWebGestaltR -lws2_32 -ladvapi32 -luserenv -lbcrypt -lntdll
-
-all: C_clean
+all: clean rustup
 
 $(SHLIB): $(STATLIB)
 
-CARGOTMP = $(CURDIR)/.cargo
-
 $(STATLIB):
+	if [ -f rust/vendor.tar.xz ]; then tar xf rust/vendor.tar.xz && mkdir -p .cargo && cp windows-config.toml .cargo/config.toml; fi
 	mkdir -p $(TARGET_DIR)/libgcc_mock
-	# `rustc` adds `-lgcc_eh` flags to the compiler, but Rtools' GCC doesn't have
-	# `libgcc_eh` due to the compilation settings. So, in order to please the
-	# compiler, we need to add empty `libgcc_eh` to the library search paths.
-	#
-	# For more details, please refer to
-	# https://github.com/r-windows/rtools-packages/blob/2407b23f1e0925bbb20a4162c963600105236318/mingw-w64-gcc/PKGBUILD#L313-L316
 	touch $(TARGET_DIR)/libgcc_mock/libgcc_eh.a
 	mkdir -p $(LIBDIR)/libgcc_mock && touch $(LIBDIR)/libgcc_mock/libgcc_eh.a
-
-	# CARGO_LINKER is provided in Makevars.ucrt for R >= 4.2
-	# vendoring (Note: to avoid NOTE of "Found the following hidden files and
-	# directories", .cargo needs to be created here)
-	if [ "$(VENDORING)" = "yes" ]; then \
-	  $(TAR) --extract --xz -f ./rust/vendor.tar.xz -C ./rust && \
-	    mkdir -p ./.cargo && \
-	    cp ./config.toml ./.cargo/config.toml; \
-	fi
-
-	if [ "$(NOT_CRAN)" != "true" ]; then \
-		export CARGO_HOME=$(CARGOTMP); \
-	fi && \
-		@BEFORE_CARGO_BUILD@  export CARGO_TARGET_X86_64_PC_WINDOWS_GNU_LINKER="$(CARGO_LINKER)" && \
-		export LIBRARY_PATH="$${LIBRARY_PATH};$(CURDIR)/$(TARGET_DIR)/libgcc_mock" && \
-		cargo build --target=$(TARGET) --jobs 2 --lib --release --manifest-path=./rust/Cargo.toml --target-dir $(TARGET_DIR) $(OFFLINE_OPTION)
-	if [ "$(NOT_CRAN)" != "true" ]; then \
-		rm -Rf $(CARGOTMP) && \
-		rm -Rf $(LIBDIR)/build; \
-	fi
-	@AFTER_CARGO_BUILD@
-
-C_clean:
-	rm -Rf $(SHLIB) $(STATLIB) $(OBJECTS)
+	export LIBRARY_PATH="$${LIBRARY_PATH};$(CURDIR)/$(TARGET_DIR)/libgcc_mock"
+	PATH="$(USERPROFILE)\.cargo\bin:$(PATH)" cargo build --target=$(TARGET) --release --manifest-path=rust/Cargo.toml
 
 clean:
-	rm -Rf $(SHLIB) $(STATLIB) $(OBJECTS) $(TARGET_DIR)
+	rm -Rf $(SHLIB) $(STATLIB) $(OBJECTS) rust/target
+	rm -Rf vendor
+
+rustup:
+	rustup target add $(TARGET) || true

src/config.toml

@@ -10,4 +10,4 @@ branch = "master"
 replace-with = "vendored-sources"
 
 [source.vendored-sources]
-directory = "rust/vendor"
+directory = "rust/vendor"

src/rust/src/lib.rs

@@ -23,13 +23,13 @@ use webgestalt_lib::{
 ///
 /// ## Diagram
 /// ```shell
-///         Gene Sets
-///       ┌───────────┐  First column named 'gene' containing gene name
-///       │A0100110100│  1 = in set
-/// Genes │B0100101000│  0 = not in set
-///       │C1011101001│  Due to limitiations with extendr-api v 0.6.0,
-///       └───────────┘  function returns a list, and the R package will
-///                      add the first 'gene' column
+///                    Gene Sets
+///                    First column named 'gene' containing gene name
+///       A0100110100  1 = in set
+/// Genes B0100101000  0 = not in set
+///       C1011101001  Due to limitiations with extendr-api v 0.6.0,
+///                    function returns a list, and the R package will
+///                    add the first 'gene' column
 /// ```
 /// @param gmt A Data Frame with geneSet and gene columns from the GMT file
 /// @param genes A vector of genes

src/windows-config.toml

@@ -0,0 +1,13 @@
+[net]
+git-fetch-with-cli = true
+
+[source.crates-io]
+replace-with = "vendored-sources"
+
+[source."git+https://github.com/bzhanglab/webgestalt_rust.git?branch=master"]
+git = "https://github.com/bzhanglab/webgestalt_rust.git"
+branch = "master"
+replace-with = "vendored-sources"
+
+[source.vendored-sources]
+directory = "vendor"

tools/rustarch.R

@@ -0,0 +1,12 @@
+# See notes in FAQ about ARM64 support:
+# https://github.com/r-rust/faq#does-rust-support-windows-on-arm64-aarch64
+arch <- if (grepl("aarch", R.version$platform)) {
+    "aarch64-pc-windows-gnullvm"
+} else if (grepl("clang", Sys.getenv("R_COMPILED_BY"))) {
+    "x86_64-pc-windows-gnullvm"
+} else if (grepl("i386", R.version$platform)) {
+    "i686-pc-windows-gnu"
+} else {
+    "x86_64-pc-windows-gnu"
+}
+cat(arch)