.Rhistory

"Sample Size: ", extra_study_details$n_title, "   |  ",
"Map: ", study_details$map_type)
# if field cons_mv_estimate exists in extra_study_details_multi, add to bottom text
if ("mv_estimate" %in% names(extra_study_details)) {
if (!is.null(extra_study_details$mv_estimate)) {
summary_info$bottom_text <- paste0(summary_info$bottom_text,"\n",
"Multivariate effect size: ", round(extra_study_details$mv_estimate, 2), " [", round(extra_study_details$mv_ci[1], 2), ", ", round(extra_study_details$mv_ci[2], 2), "]")
# bottom_text <- paste0("Max conservative effect size: ", extra_study_details$max_cons_effect, "\n",
#                       "Percent not overlapping zero: ", round(extra_study_details$percent_not_zero * 100, 1), "%\n",
#                       "Multivariate effect size: ", round(extra_study_details$mv_estimate, 2), " [", round(extra_study_details$mv_ci[1], 2), ", ", round(extra_study_details$mv_ci[2], 2), "]")
}
}
} else {
summary_info$title_text <- paste0(summary_info$grouping_var_title, ": ", extra_study_details$group_level, "   |  ",
"Reference Space: ", extra_study_details$ref)
# if field cons_mv_estimate exists in extra_study_details_multi, add to bottom text # TODO: currently not defined when using group_data
if ("mv_estimate" %in% names(extra_study_details)) { # meta
if (!is.null(extra_study_details$mv_estimate)) {
summary_info$bottom_text <- paste0(summary_info$bottom_text,"\n",
"Multivariate effect size: ", round(extra_study_details$mv_estimate, 2), " [", round(extra_study_details$mv_ci[1], 2), ", ", round(extra_study_details$mv_ci[2], 2), "]")
# bottom_text <- paste0("Max conservative effect size: ", extra_study_details$max_cons_effect, "\n",
#                       "Percent not overlapping zero: ", round(extra_study_details$percent_not_zero * 100, 1), "%\n",
#                       "Multivariate effect size: ", round(extra_study_details$mv_estimate, 2), " [", round(extra_study_details$mv_ci[1], 2), ", ", round(extra_study_details$mv_ci[2], 2), "]")
}
} else if ("max_cons_mv_estimate" %in% names(extra_study_details)) { # overlapping
summary_info$bottom_text <- paste0(summary_info$bottom_text,"\n",
"Max conservative multivariate effect size: ", round(extra_study_details$max_cons_mv_estimate, 2))
# bottom_text <- paste0("Max conservative effect size: ", extra_study_details$max_cons_effect, "\n",
#                       "Percent not overlapping zero: ", round(extra_study_details$percent_not_zero * 100, 1), "%\n",
#                       "Max conservative multivariate effect size: ", round(extra_study_details$max_cons_mv_estimate, 2))
# } else {
# bottom_text <- paste0("Max conservative effect size: ", extra_study_details$max_cons_effect, "\n",
#                       "Percent not overlapping zero: ", round(extra_study_details$percent_not_zero * 100, 1), "%")
}
}
return(summary_info)
}
#########################################################################
# Combine summary info if multiple overlapping plots so there's one description per panel
combine_summary_info <- function(summary_info) {
summary_info2 <- list()
summary_info2$title_text <- summary_info[[1]]$title_text
# append num studies
summary_info2$title_text <- paste0(summary_info2$title_text, " (", length(summary_info), " studies)")
# for summary_info2$bottom_text, get summary_info[:]$bottom_text strings for each study and parse max cons effect size, perce not overlapping zero, and multivar effect size. Take average
max_cons_effect__overlapping <- max(as.numeric(sapply(summary_info, function(x) as.numeric(sub(".*Max conservative effect size: ([^\n]+).*", "\\1", x$bottom_text)))))
percent_not_zero__overlapping <- mean(as.numeric(sapply(summary_info, function(x) as.numeric(sub(".*Percent not overlapping zero: ([^%]+).*", "\\1", x$bottom_text)))))
mv_estimate__overlapping <- mean(as.numeric(sapply(summary_info, function(x) as.numeric(sub(".*Max conservative multivariate effect size: ([^\n]+).*", "\\1", x$bottom_text)))))
summary_info2$bottom_text <- paste0("Max conservative effect size: ", round(max_cons_effect__overlapping, 2), "\n",
"Percent not overlapping zero: ", round(percent_not_zero__overlapping, 1), "%\n",
"Multivariate effect size: ", round(mv_estimate__overlapping, 2))
summary_info <- summary_info2
}
#########################################################################
#' Add summary labels
#'
#' This function adds summary labels to a ggplot object.
#'
#' @param p A ggplot object to which the labels will be added.
#' @param study_details A list of original study details: orig_stat_type, test_component_1, test_component_2, dataset, map_type, group, and ref.
#' @param extra_study_details A list containing extra descriptive info: percent_not_zero, max_cons_effect, group_by_title, n_title, mv_estimate, and mv_ci.
#'
#' @return A ggplot object with summary labels.
#' @export
#'
#' @examples
#' # Example usage
#' # add_plot_description(p, pp, study_details, extra_study_details)
add_plot_description <- function(p, pp, summary_info, add_extra_text, do_minimal_title) {
# add description-specific plot params
pp$title_size <- 21
pp$caption_size <- 15
pp$title_hjust <- 0.5
pp$title_lmargin <- -6 # adjust it a bit left of the plot y-axis
pp$caption_hjust <- 0
# if overlapping plots, make single summary_info
# study_details__overlapping <- plot_data_list[[1]]$study_details
#   extra_study_details__overlapping <- plot_data_list[[1]]$extra_study_details
#
#   # get summary info across all studies in group
#   study_summary <- lapply(plot_data_list, function(x) {
#     list(
#       max_cons_estimate = x$extra_study_details$max_cons_estimate,
#       percent_not_zero = x$extra_study_details$percent_not_zero,
#       n_variables = length(x$data$cons_estimate),
#       cons_mv_estimate = x$extra_study_details$mv_ci[1]
#     )
#   })
#
#   study_summary <- do.call(rbind, lapply(study_summary, as.data.frame))
#
#   # take max of all max effects and average percent nonzeros
#   extra_study_details__overlapping$max_cons_effect <- study_summary$max_cons_estimate[which.max(abs(study_summary$max_cons_estimate))]
#   extra_study_details__overlapping$percent_not_zero <- sum(study_summary$percent_not_zero*study_summary$n_variables)/sum(study_summary$n_variables)
#   extra_study_details__overlapping$max_cons_mv_estimate <- max(study_summary$cons_mv_estimate)
#
#   p <- add_plot_description(p, pp, study_details__overlapping, extra_study_details__overlapping)
#
# for manuscript, remove labels from title for cleaner look
if (do_minimal_title) {
summary_info$title_text <- gsub("(Dataset: )|  ([|])[^:]*:", "\\2", summary_info$title_text)
summary_info$title_text <- gsub("([|][^|]*$)", "", summary_info$title_text)
}
p <- p +
ggtitle(summary_info$title_text) +
theme(plot.title = element_text(hjust = pp$title_hjust, margin = margin(l = pp$title_lmargin, unit = "pt"),
size = pp$title_size, face = "bold"),
plot.caption = element_text(hjust = pp$caption_hjust, size = pp$caption_size),
panel.background = element_blank(),
plot.background = element_blank()
)
if (add_extra_text) {
p <- p + labs(caption = summary_info$bottom_text)
}
return(p)
}
## generate_figures.R
####################################################################
#
# Run meta-analysis & create plots
#
####################################################################
## Load libraries
library(metafor)
library(ggpubr) # requires svglite
library(devtools) # for installing/loading utils
library(dplyr)
library(devtools)
library(png)
library(reshape2)
#library(BrainEffeX.utils)
library(neurobase)
library(colorspace)
## Load utils
# Option 1. simple add local scripts to path (for dev):
#utils_package_local <- "/Users/stephanienoble/Library/CloudStorage/GoogleDrive-s.noble@northeastern.edu/My Drive/Lab/xMore/Software/scripts/R/myscripts/effect_size/misc/BrainEffeX_utils/"
#load_all(utils_package_local)
# Option 2. full install of local package (for dev):
# utils_package_local <- "/Users/stephanienoble/Library/CloudStorage/GoogleDrive-s.noble@northeastern.edu/My Drive/Lab/xMore/Software/scripts/R/myscripts/effect_size/misc/BrainEffeX_utils/"
# install(utils_package_local)
# library(BrainEffeX.utils)
# Option 3. install package from github:
# utils_github_path <- "neuroprismlab/BrainEffeX_utils"
# install_github(utils_github_path)
# library(BrainEffeX.utils)
# set working directory to the BrainEffeX local cloned repo
repo_location = "/Users/shearer.h/Google Drive/My Drive/Github/BrainEffeX/"
setwd(repo_location)
testing = FALSE
## User-defined paths & parameters
meta_data_file_name <- 'meta_analysis_2025-05-01.RData'
combined_data_file_name <- 'combined_data_2025-02-24.RData'
data_dir <- "data/"
out_basename_basename <- "figures/"
# if testing, save the outputs to
if (testing) {
out_basename_basename <- paste0(repo_location, "testing/")
}
# plot params - USER-DEFINED
plot_output_style <- c('shiny') # c('shiny', 'manuscript') # 'shiny': save plots for one study at a time, simci-spatial type, add descrip; 'manuscript': save plots concatenated across all studies, add descrip
all_effect_size_types <- c('d','r_sq')              # c('d', 'r_sq', 'd.full_res')
all_motion <- c('none', 'regression', 'threshold')  # c('none', 'regression', 'threshold') # TODO: stat_control -> "...regression...$d", full_residualization -> "...regression...$d.full_res"
all_pooling <- c('none', 'net') #  # c('none','net')
do_multi <- c(FALSE)
all_plot_combination_styles <- c('meta')   # c('single','meta','overlapping') # note: overlapping can only be used for manuscript
all_grouping_var <- c('orig_stat_type', 'category')  # c('none', 'category', 'orig_stat_type') # used only for meta & overlap plots - TODO: separate out?
all_manuscript_plot_types <- c('simci', 'spatial')  # c('simci', 'spatial', 'density', 'power') # only used for plot_output_style = 'manuscript'
make_plots <- TRUE
save_plots <- TRUE
# save_logs <- FALSE
rearrange_by_stat_type <- TRUE # for single plots
# check if the data directory contains the data files
# if not, download them from OSF
if (!file.exists(paste0(data_dir, combined_data_file_name))) {
library(osfr)
print("Data is not downloaded. Downloading now from OSF.")
project <- osf_retrieve_node("https://osf.io/cwnjd")
files <- osf_ls_files(project)
file_idx <- files$name == combined_data_file_name
file <- files[file_idx,]
osf_download(file, path = file.path(data_dir))
if (combined_data_file_name %in% list.files(data_dir)) {
# file already exists
print("Main data file succesfully downloaded from OSF")
} else {
# file does not exist
print("Data could not be downloaded from OSF. Try manually downloading and moving the file to the data directory from https://osf.io/cwnjd.")
}
}
if (!file.exists(paste0(data_dir, meta_data_file_name))) {
library(osfr)
print("Data is not downloaded. Downloading now from OSF.")
project <- osf_retrieve_node("https://osf.io/cwnjd")
files <- osf_ls_files(project)
file_idx <- files$name == meta_data_file_name
file <- files[file_idx,]
osf_download(file, path = file.path(data_dir))
if (meta_data_file_name %in% list.files(data_dir)) {
# file already exists
print("Meta-analysis data file succesfully downloaded from OSF")
} else {
# file does not exist
print("Data could not be downloaded from OSF. Try manually downloading and moving the file to the data directory from https://osf.io/cwnjd.")
}
}
# special settings for shiny vs. manuscript
if (plot_output_style == 'shiny') {
all_plot_types <- c('simci-spatial')
if ('overlapping' %in% all_plot_combination_styles) {
warning("Overlapping plots not supported in shiny mode. Removing.")
all_plot_combination_styles <- all_plot_combination_styles[all_plot_combination_styles != 'overlapping']
}
# if (save_logs == TRUE) {
# warning("Logs not used in shiny. Setting save_logs to FALSE.")
save_logs <- FALSE
# }
add_plt_description <- TRUE # text at bottom of screen
use_minimal_title <- FALSE
} else if (plot_output_style == 'manuscript') {
all_plot_types <- all_manuscript_plot_types
all_plot_combination_styles <- all_plot_combination_styles
save_logs = TRUE
add_plt_description <- FALSE # text at bottom of screen
use_minimal_title <- TRUE # cleaner title for manuscript
}
# input and output directories
## Loop over plot types and styles
for (pooling in all_pooling) {
for (motion in all_motion) {
for (grouping_var in all_grouping_var) {
for (effect_size_type in all_effect_size_types) {
for (plot_combination_style in all_plot_combination_styles) {
for (plot_type in all_plot_types) {
print(paste0('Doing plot_combination_style: ', plot_combination_style, ' | plot_type: ', plot_type, ' | pooling: ', pooling, ' | motion: ', motion, ' | grouping_var: ', grouping_var))
## Set up strings
combo_name <- paste0('pooling.', pooling, '.motion.', motion, '.mv.none')
mv_combo_name <- paste0('pooling.', pooling, '.motion.', motion, '.mv.multi')
# Correct args
if (plot_combination_style == 'single' & grouping_var != 'none') {
grouping_var <- 'none'
cat("Warning: grouping_var set to 'none' for single plots\n")
}
## Load data
if (!exists("v")) {
v <- load_data(data_dir)
}
## Run meta-analysis, if specified
# TODO: move to combine_gl with other more intensive processing / stat estimates & run all relevantmeta beforehand
run_meta <- FALSE
try_meta_file <- FALSE
if (plot_combination_style == 'meta') {
# setup string and filename
meta_str <- paste0('meta_',grouping_var)
meta_fn_dir <- system.file("meta/", package = "BrainEffeX.utils") # TODO: set this somewhere else
meta_fn <- file.path(meta_fn_dir, "v.RData")
# load meta_analysis data file
#meta_str <- paste0('meta_', grouping_var)
meta_path <- paste0(data_dir, meta_data_file_name)
load(meta_path)
# this loads v that includes meta data
# check if this var contains the meta for this grouping var / combo
if (!(meta_str %in% names(v))) { # check if this grouping var already exists in data
try_meta_file <- TRUE
} else if (!(combo_name %in% names(v[[meta_str]]$data[[1]]))) { # check if this combo_name has been run
try_meta_file <- TRUE
}
# TODO: will also have to catch the case where d is defined but r_sq is not
# check if saved meta file contains the meta for this grouping var / combo
if (try_meta_file) {
if (file.exists(meta_fn)) { # try to read pre-saved meta-analysis
load(meta_fn)
if (!(meta_str %in% names(v))) { # check again for grouping var
run_meta <- TRUE
} else if (!(combo_name %in% names(v[[meta_str]]$data[[1]]))) { # check again for combo
run_meta <- TRUE
}
} else { # no file
run_meta <- TRUE
}
}
}
if (run_meta) {
v <- meta_analysis(v, v$brain_masks, combo_name, grouping_var = grouping_var)
save(v, file = meta_fn)
}
## Set up unique identifiers for each plot
plot_info__idx <- list() # each row = list of study(s) in data to include in each plot
# for single plots: each row = 1 entry per study to index into v$data
# for meta-analysis plots: each row = 1 entry per category to index v[[meta_str]]$data
# for overlapping plots: each row = list of indices per group (x map type) to index into v$data
plot_info__grouping_var <- list() # each row = grouping variable (same value repeated for each plot)
plot_info__group_level <- list() # each row = level within grouping variable
plot_info__ref <- list() # each row = ref(s) used for a study or grouping variable
if (plot_combination_style == 'single') {  # name by study
all_study_names <- names(v$data)
for (i in 1:length(v$data)) {
plot_info__idx[[all_study_names[[i]]]] <- i
plot_info__grouping_var[[all_study_names[[i]]]] <- "none"  # overwrite any other grouping var if doing single plots
plot_info__group_level[[all_study_names[[i]]]] <- NA
plot_info__ref[[all_study_names[[i]]]] <- v$study$ref[i]
}
# sort all rows of plot_info__idx by orig_stat_type, with studies sharing same stat type next to each other
if (rearrange_by_stat_type) {
orig_stat_type_order <- c(which(v$study$orig_stat_type == 'r'), which(v$study$orig_stat_type == 't2'), which(v$study$orig_stat_type == 't'))
plot_info__idx <- plot_info__idx[orig_stat_type_order]
plot_info__grouping_var <- plot_info__grouping_var[orig_stat_type_order]
plot_info__group_level <- plot_info__group_level[orig_stat_type_order]
plot_info__ref <- plot_info__ref[orig_stat_type_order]
}
} else if (plot_combination_style == 'meta') { # name by average of grouping var
for (i in 1:length(v[[meta_str]]$data)) {
plot_info__idx[[names(v[[meta_str]]$data)[[i]]]] <- i
plot_info__grouping_var[[names(v[[meta_str]]$data)[[i]]]] <- grouping_var
plot_info__group_level[[names(v[[meta_str]]$data)[[i]]]] <- v[[meta_str]]$study$group_level[i]
plot_info__ref[[names(v[[meta_str]]$data)[[i]]]] <- v[[meta_str]]$study$ref[i]
}
} else if (plot_combination_style == 'overlapping') { # overlapping individual plots
if (grouping_var == 'category') {
study_group_name <- v$study$category
} else if (grouping_var == 'orig_stat_type') {
study_group_name <- v$study$orig_stat_type
}
all_group_names <- unique(study_group_name)
all_map_types <- unique(v$study$map_type)
for (this_map_type in all_map_types) {
for (this_group_name in all_group_names) {
idx <- which(study_group_name == this_group_name & v$study$map_type == this_map_type)
plot_info__idx[[paste0(this_group_name, '.', this_map_type)]] <- idx
plot_info__grouping_var[[paste0(this_group_name, '.', this_map_type)]] <- grouping_var
plot_info__group_level[[paste0(this_group_name, '.', this_map_type)]] <- this_group_name
plot_info__ref[[paste0(this_group_name, '.', this_map_type)]] <- unique(v$study$ref[idx])
}
}
}
plot_info <- data.frame(
idx = I(plot_info__idx),
grouping_var = unlist(plot_info__grouping_var),
group_level = unlist(plot_info__group_level),
ref = I(plot_info__ref),
row.names = names(plot_info__idx),
stringsAsFactors = FALSE
)
rm(plot_info__idx, plot_info__grouping_var, plot_info__group_level, plot_info__ref)
## Make Plots
panel_list <- list() # list of panels
panel_list_2 <- list() # list of panels
panel_list <-list()
log_list <- list() # list of logs
for (i in 1:length(plot_info$idx)) { # loop over panels - this_study_or_group is the name of the group or study
this_study_or_group <- rownames(plot_info)[i]
this_plot_info <- plot_info[this_study_or_group,]
pd_list <- list() # list of plot info for single panel
pd_list_2 <- list() # add'l list of plot info for plot type to add to pd_list
ld_list <- list() # list of log info for single panel
n_studies_in_pd_list <- 1
# 1. Prep
for (j in plot_info$idx[[i]]) {
print(plot_info$idx[[i]])
# change metadata based on whether using meta-analysis
if (plot_combination_style == 'meta') {
data <- v[[meta_str]]$data[[j]]
study_details <- list()
brain_masks <- v[[meta_str]]$brain_masks[[j]]$pooling.none.motion.none.mv.none # TODO: this is because we explicitly set this for meta but not for single studies - assuming motion type shouldn't affect the mask and always using an external mask for pooling
} else {
data <- v$data[[j]]
study_details <- v$study[j, ]
# if (pooling == "net" && study_details$map_type == "act") {
#   next
# } #testing
brain_masks <- v$brain_masks[[j]]
}
if (combo_name %in% names(data)) { # if combo_name exists in data (e.g., not all studies have net)
if (any(!is.na(data[[combo_name]][[effect_size_type]])) > 0) {  # data is not just NA
# prep
if (plot_type == 'simci-spatial') { # shiny
pd_list[[n_studies_in_pd_list]] <- prep_data_for_plot(data = data, study_details = study_details, combo_name = combo_name, mv_combo_name = mv_combo_name, estimate = effect_size_type, plot_info = this_plot_info)
pd_list_2[[n_studies_in_pd_list]] <- prep_data_for_spatial_plot(data = data, brain_masks = brain_masks, study_details = study_details, combo_name = combo_name, mv_combo_name = mv_combo_name, estimate = effect_size_type, plot_info = this_plot_info)
} else if (plot_type == 'spatial') {
# TODO: we probably don't even need a dedicated function for the spatial plots, just pass the relevant info
pd_list[[n_studies_in_pd_list]] <- prep_data_for_spatial_plot(data = data, brain_masks = brain_masks, study_details = study_details, combo_name = combo_name, mv_combo_name = mv_combo_name, estimate = effect_size_type, plot_info = this_plot_info)
} else {
pd_list[[n_studies_in_pd_list]] <- prep_data_for_plot(data = data, study_details = study_details, combo_name = combo_name, mv_combo_name = mv_combo_name, estimate = effect_size_type, plot_info = this_plot_info)
}
ld_list[[n_studies_in_pd_list]] <- get_summary_info(pd_list[[n_studies_in_pd_list]]$study_details, pd_list[[n_studies_in_pd_list]]$extra_study_details)
n_studies_in_pd_list <- n_studies_in_pd_list + 1
}
}
}
# 2. Plot & Log
if (make_plots) {
if (length(pd_list) > 0) { # plot only if pd_list isn't empty
# set up plot
if (length(ld_list) > 1) {
log_list[[i]] <- combine_summary_info(ld_list)
} else {
log_list[[i]] <- ld_list[[1]]
}
if (plot_type == 'simci-spatial') {
# if (pooling == "net" && study_details$map_type == "act") {
#   next
# } #testing
panel_list[[i]] <- create_plots(pd_list, plot_type = 'simci', effect_type = effect_size_type, do_multivariate = FALSE, add_description = add_plt_description, do_minimal_title = use_minimal_title, log_list[[i]], meta = plot_combination_style == "meta")
panel_list_2[[i]] <- create_plots(pd_list_2, plot_type = 'spatial', effect_type = effect_size_type, do_multivariate = FALSE, add_description = add_plt_description, do_minimal_title = use_minimal_title, log_list[[i]], meta = plot_combination_style == "meta")
} else {
panel_list[[i]] <- create_plots(pd_list, plot_type = plot_type, effect_type = effect_size_type, do_multivariate = FALSE, add_description = add_plt_description, do_minimal_title = use_minimal_title, log_list[[i]], meta = plot_combination_style == "meta")
}
}
}
}
if (make_plots) {
# General plot parameters
# TODO: figure out what we want to set up here vs. to pass or set up in
# create_plots, which gets passed to plot_sim_ci, etc.
# Should at least set all panel / canvas dimensions here
pp <- list()
pp$width_per_panel <- 7
pp$height_per_panel <- 6
pp$res <- 100
pp$units <- "in"
pp$title_size <- 20
if (plot_type == 'simci-spatial') {
pp$ncol <- 2
pp$nrow <- 1
} else if (plot_type == 'power') {
pp$ncol <- length(panel_list)
pp$nrow <- 1
} else {
pp$ncol <- 1
pp$nrow <- length(panel_list)
}
if (plot_output_style == 'manuscript') {
do_minimal_title <- TRUE
} else {
do_minimal_title <- FALSE
}
# prep: set up dir and file names
if (save_plots) {
if (plot_combination_style == 'meta') {
grouping_var_str <- paste0('_', grouping_var)
} else {
grouping_var_str <- ''
}
if (plot_output_style == 'manuscript') {
grouping_var_str <- paste0('_', plot_type)
} else {
grouping_var_str <- paste0('_', plot_type, '_', effect_size_type)
}
out_basename <- paste0(out_basename_basename, plot_output_style, '/', effect_size_type, '/motion_', motion, '/pooling_', pooling, '/', plot_combination_style, grouping_var_str)
if (plot_type == 'simci-spatial') { # use out_basename as dir, otherwise use as basename for concat plots
out_basename <- paste0(out_basename,'/')
}
actual_dir <- sub("/[^/]*$", "", out_basename)
if (!dir.exists(actual_dir)) {
dir.create(actual_dir, recursive = TRUE)
}
cat("Saving plots to...\n", out_basename, "\n", sep = "")
}
# plot multi panels
if (plot_type == 'simci-spatial') {
for (i in 1:length(panel_list)) {
if (is.null(panel_list[[i]]) || is.null(panel_list_2[[i]])) {
next
}
t <- list(panel_list[[i]], panel_list_2[[i]])
t_master_title <- t[[1]]$labels$title
t[[1]]$labels$title <- ""
t[[2]]$labels$title <- ""
multi_plot <- ggarrange(plotlist = t, ncol = pp$ncol, nrow = pp$nrow)
multi_plot <- annotate_figure(multi_plot,
top = text_grob(t_master_title, face = "bold", size = pp$title_size))
if (save_plots) {
study_name <- names(plot_info$idx)[[i]]
plot_fn <- paste0(out_basename, study_name, '.png')
ggsave(plot_fn, plot = multi_plot, width = pp$width_per_panel * pp$ncol, height = pp$height_per_panel * pp$nrow, units = pp$units, dpi = pp$res, bg = "white", device = "png")
if (save_logs) { # TODO: some logs are identical - see about saving only single
log_fn <- paste0(out_basename, study_name,'.txt')
writeLines(unlist(lapply(log_list, function(x) c(x$title_text, x$bottom_text, ""))), log_fn)
}
}
}
} else {
multi_plot <- ggarrange(plotlist = panel_list, ncol=pp$ncol, nrow=pp$row)
if (save_plots) {
plot_fn <- paste0(out_basename, '.png')
ggsave(plot_fn, plot = multi_plot, width = pp$width_per_panel * pp$ncol, height = pp$height_per_panel * pp$nrow, units = pp$units, dpi = pp$res, bg = "white", device = "png", limitsize = FALSE)
if (save_logs) { # TODO: some logs are identical - see about saving only single
log_fn <- paste0(out_basename, '.txt')
writeLines(unlist(lapply(log_list, function(x) c(x$title_text, x$bottom_text, ""))), log_fn)
}
}
}
}
## close loop over plot types and styles
} # effect_size_type
} # pooling
} # plot_combination_style
} # plot_type
} # motion
} # grouping_var
shiny::runApp('~/Google Drive/My Drive/Github/BrainEffeX')
library(useShinyjs)
library(shinyjs)
runApp('~/Google Drive/My Drive/Github/BrainEffeX')
runApp('~/Google Drive/My Drive/Github/BrainEffeX')
runApp('~/Google Drive/My Drive/Github/BrainEffeX')
runApp('~/Google Drive/My Drive/Github/BrainEffeX')
runApp('~/Google Drive/My Drive/Github/BrainEffeX')
runApp('~/Google Drive/My Drive/Github/BrainEffeX')
rsconnect::writeManifest()
getRversion()
runApp('~/Google Drive/My Drive/Github/BrainEffeX')
runApp()