changed the README and small improvement to writing xlsx files.

The de xlsx writer will no longer save plots to the rda file,
_significantly_ reducing their size from absurd Gb scale to small Mb.

Author: abelew

R/de_xlsx.R

@@ -272,9 +272,9 @@ combine_de_tables <- function(apr, extra_annot = NULL, keepers = "all", excludes
 
   ## Finished!  Dump the important stuff into a return list.
   ret <- list(
+    "input" = apr,
     "data" = extracted[["data"]],
     "table_names" = extracted[["table_names"]],
-    "input" = apr,
     "plots" = extracted[["plots"]],
     "comp_plot" = comp,
     "keepers" = keepers,
@@ -286,14 +286,18 @@ combine_de_tables <- function(apr, extra_annot = NULL, keepers = "all", excludes
 
   if (!is.null(rda)) {
     varname <- gsub(x = basename(rda), pattern = "\\.rda$", replacement = "")
-    assigned <- assign(varname, ret)
-    ## When saving a rda of the combined tables, it is less likely that one wants a copy of the
-    ## entire differential expression analysis produced by all_pairwise().
+    tmp <- ret
     if (!isTRUE(rda_input)) {
-      varname[["input"]] <- NULL
+      message("Removing the input and plots from the rda result.")
+      tmp[["plots"]] <- NULL
+      tmp[["input"]] <- NULL
     }
+    assigned <- assign(varname, tmp)
+    removed <- rm(list = "tmp")
+    ## When saving a rda of the combined tables, it is less likely that one wants a copy of the
+    ## entire differential expression analysis produced by all_pairwise().
     message("Saving de result as ", varname, " to ", rda, ".")
-    saved <- save(list = varname, file = rda)
+    saved <- save(list = varname, file = rda, compress = "xz")
     removed <- rm(varname)
   }
   ## Cleanup the saved image files.

R/zzz_print.R

@@ -548,8 +548,8 @@ print.snps_genes <- function(x) {
   most_idx <- x[["count_by_gene"]] == most_num
   most_name <- names(x[["count_by_gene"]])[most_idx]
   summary_string <- glue("When the variants observed were cross referenced against annotated genes,
-{gt_zero} genes were observed with at least 1 variant.  {most_name} had the most variants, with
-{most_num}.")
+{gt_zero} genes were observed with at least 1 variant.
+{most_name} had the most variants, with {most_num}.")
   message(summary_string)
   return(invisible(x))
 }

README.md

@@ -22,17 +22,13 @@ computer with nothing pre-installed.
 ## Compiled Prerequisites
 
 Quite a few packages used by hpgltools have compiled code, here is an incomplete list of
-requisites and the likely Debian-based (circa 201811) solution.
+requisites and the likely Debian-based.
 
 * The basics: gcc,g++,gfortran,make,boost,build-essential,automake,autoconf
 * curl
-* graphviz (This has problems, I ended up using the Rgraphviz from github and fiddling around in
-    the src/ directory to get it to compile.)
-* java
 * libgvc
 * libnlopt
 * libxml2
-* mysqlclient (mariadb)
 * pandoc
 * pkg-config
 * texlive
@@ -43,29 +39,21 @@ requisites and the likely Debian-based (circa 201811) solution.
 
 ## For the basics:
 apt-get install gcc g++ cc1plus gfortran gawk sed bison libboost-all-dev build-essential \
-  curl libcurl4-gnutls-dev:amd64 libcurl4:amd64 default-jdk automake autoconf autotools-dev \
+  curl libcurl4-gnutls-dev:amd64 libcurl4:amd64 automake autoconf autotools-dev \
   pkg-config
 
-## For a large set of potentially relevant development tools:
-apt-get install kdesdk
-
 ## For the prerequisite tools:
 apt-get install udunits2-bin libudunit2-0 libudunits2-dev libnlopt0 libnlopt-dev \
   libxml2-dev pandoc texlive-latex-recommended texlive-science texlive-extra-utils \
-  texlive-fonts-extra texlive-latex-extra libgvc6 libgraphviz-dev libmariadbclient-dev
+  texlive-fonts-extra texlive-latex-extra
 ```
 
-### Caveats
-
-1.  Rgraphviz compilation fails with 'too few arguments to agedge' and similar problems.
-
 ## The actual R installation
 
 Download the package via 'git pull' or from the github download link, go
 into the hpgltools/ directory and:
 
 ```bash
-R CMD javareconf -e
 make prereq  ## install knitr, devtools, and friends.
 make deps  ## Pick up dependencies from the DESCRIPTION file.
 make install  ## Perform the installation
@@ -85,32 +73,10 @@ make build
 make check
 ## Cleanup this tree
 make clean
-## Clean after making vignettes
-make clean_vignette
-## Install dependencies
-make dep
 ## Rerun roxygen, rebuild the vignettes, and the reference manual
 make document
 ## Install this via R CMD INSTALL
 make install
-## Install _ALL_ of bioconductor!
-make install_bioconductor
-## Install prerequisite packages which are not explicitly dependencies, but are useful
-make prereq
-## push to github
-make push
-## remake the roxygen docs
-make roxygen
-## Remake the reference manual
-make reference
-## Install suggested packages
-make suggests
-## Run the test suite
-make test
-## Update R packages via bioconductor
-make update
-## Update bioconductor
-make update_bioc
 ## Rebuild the vignettes
 make vignette
 ```