R/Main.R
In ohdsi-studies/dGEMcovid: A Skeleton for running distributed model studies
Defines functions
makeFriendlyNames
execute
Documented in
execute
# Copyright 2020 Observational Health Data Sciences and Informatics
#
# This file is part of dGEMcovid
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#' Execute the Study
#'
#' @details
#' This function executes the dGEMcovid Study.
#'
#' @param databaseDetails The connection details and OMOP CDM details. Created using \code{PatientLevelPrediction::createDatabaseDetails}.
#' @param siteId The name of your site (can be the university name, site with a number, etc.) tnis needs to be shared with the study administrator
#' @param outputFolder Name of local folder to place results - make sure control is in here; make sure to use forward slashes
#' (/). Do not use a folder on a network drive since this greatly impacts
#' performance.
#' @param createCohorts Create the cohortTable table with the target population and outcome cohorts?
#' @param createData Create the labelled data set (this is required before runAnalysis)
#' @param createPlot Plot the number of covid hospitalizations by month
#' @param sampleSize The number of patients in the target cohort to sample (if NULL uses all patients)
#' @param createControl (for the lead site only) Run this code to create the study control
#' @param siteIds (for the lead site only) vector with the names of all the sites contributing to the study. Required when createControl = TRUE
#' @param runInitialize Runs the initialization step - this require downloading the json control from https://pda-ota.pdamethods.org/
#' @param runDerive Runs the derive step - this require downloading the updated json control from https://pda-ota.pdamethods.org/
#' @param runEstimate Runs the estimate step - this require downloading the updated json control from https://pda-ota.pdamethods.org/
#' @param runSynthesize Once the site estimates are returned, it is now possible to apply each model to the data to calculate
#' predictions. This step requires downloading the updated json control from https://pda-ota.pdamethods.org/
#' @param verbosity Sets the level of the verbosity. If the log level is at or higher in priority than the logger threshold, a message will print. The levels are:
#' \itemize{
#' \item{DEBUG}{Highest verbosity showing all debug statements}
#' \item{TRACE}{Showing information about start and end of steps}
#' \item{INFO}{Show informative information (Default)}
#' \item{WARN}{Show warning messages}
#' \item{ERROR}{Show error messages}
#' \item{FATAL}{Be silent except for fatal errors}
#' }
#' @param alphaStart The start date for the alpha wave in the data in "yyyymmdd" format
#' @param alphaEnd The end date for the alpha wave in the data in "yyyymmdd" format
#' @param deltaStart The start date for the delta wave in the data in "yyyymmdd" format
#' @param deltaEnd The end date for the detla wave in the data in "yyyymmdd" format
#'
#' @examples
#' \dontrun{
#' connectionDetails <- createConnectionDetails(dbms = "postgresql",
#' user = "joe",
#' password = "secret",
#' server = "myserver")
#'
#' execute(databaseDetails,
#' siteId = 'site_1',
#' outputFolder = "c:/temp/study_results",
#' createCohorts = T,
#' createData = T,
#' sampleSize = 10000,
#' createControl = F,
#' siteIds = c('site_1', 'site_2'),
#' leadSiteNextStep = F,
#' runAnalysis = F,
#' runSynthesize = F,
#' verbosity = "INFO"
#' )
#' }
#'
#' @export
execute <- function(
databaseDetails,
siteId = 'site_name',
outputFolder,
createCohorts = F,
createData = F,
createPlot = createData,
sampleSize = NULL,
createControl = F,
siteIds = '', # only needed if lead site
leadSiteNextStep = F,
runInitialize = F,
runDerive = F,
runEstimate = F,
runSynthesize = F,
verbosity = "INFO",
alphaStart = '20201101',
alphaEnd = '20210201',
deltaStart = '20210701',
deltaEnd = '20211001'
) {
dates <- list(
alpha = list(startDate = alphaStart, endDate = alphaEnd),
delta = list(startDate = deltaStart, endDate = deltaEnd)
)
waves <- c('alpha','delta')
if (!file.exists(outputFolder)){
dir.create(outputFolder, recursive = TRUE)
}
website <- 'https://pda-ota.pdamethods.org/'
# load the analysis
analysisListFile <- system.file(
"settings",
"analysisList.json",
package = "dGEMcovid"
)
if(file.exists(analysisListFile)){
analysisList <- loadJson(analysisListFile)
analysisList <- fromJsonFormat(analysisList)
analysisList$studySettings$covariateSettings <- addCohortSettings(
covariateSettings = analysisList$studySettings$covariateSettings,
cohortDatabaseSchema = databaseDetails$cohortDatabaseSchema ,
cohortTable = databaseDetails$cohortTable
)
}else{
stop('No analysisList available')
}
if(createCohorts) {
ParallelLogger::logInfo("Creating cohorts")
createCohorts(
databaseDetails = databaseDetails,
outputFolder = outputFolder
)
}
if(createPlot){
ParallelLogger::logInfo('Generating date plot')
generateDatePlot(
databaseDetails = databaseDetails,
outputFolder = outputFolder
)
}
if(createData){
# save the settings
timeSettings <- data.frame(
wave = names(dates),
starts = unlist(lapply(dates, function(x) x$startDate)),
ends = unlist(lapply(dates, function(x) x$endDate))
)
write.csv(timeSettings, file.path(outputFolder, 'timeSettings.csv'), row.names = F)
for(wave in waves){
if(!dir.exists(file.path(outputFolder, wave))){
dir.create(file.path(outputFolder, wave), recursive = T)
}
# restrict to the input date:
ParallelLogger::logInfo(paste0(wave, " date: ",dates[[wave]]$startDate, "-", dates[[wave]]$endDate ))
analysisList$studySettings$restrictPlpDataSettings$studyStartDate <- dates[[wave]]$startDate
analysisList$studySettings$restrictPlpDataSettings$studyEndDate <- dates[[wave]]$endDate
ParallelLogger::logInfo("Creating labelled dataset")
databaseDetails$cohortId <- analysisList$studySettings$targetId
databaseDetails$outcomeIds <- analysisList$studySettings$outcomeId
plpData <- tryCatch(
{
PatientLevelPrediction::getPlpData(
databaseDetails = databaseDetails,
covariateSettings = analysisList$studySettings$covariateSettings,
restrictPlpDataSettings = analysisList$studySettings$restrictPlpDataSettings
)},
error = function(e){ParallelLogger::logInfo(e); return(NULL)}
)
if(!is.null(plpData)){
labels <- PatientLevelPrediction::createStudyPopulation(
plpData = plpData,
outcomeId = analysisList$studySettings$outcomeId,
populationSettings = analysisList$studySettings$populationSettings
)
# convert to matrix
dataObject <- PatientLevelPrediction::toSparseM(
plpData = plpData,
cohort = labels
)
#sparse matrix: dataObject$dataMatrix
#labels: dataObject$labels
columnDetails <- as.data.frame(dataObject$covariateRef)
cnames <- columnDetails$covariateName[order(columnDetails$columnId)]
ipMat <- as.matrix(dataObject$dataMatrix)
ipdata <- as.data.frame(ipMat)
colnames(ipdata) <- makeFriendlyNames(cnames)
ipdata$outcome <- dataObject$labels$outcomeCount
# modify the covariates
# Charlson comorbidity categories: 0-1, 2-4, and 5
if('Charlson_index___Romano_adaptation' %in% colnames(ipdata)){
ParallelLogger::logInfo('Processing Charlson index into categories')
# Charlson comorbidity categories: 0-1, 2-4, and 5
# make 0-1 the reference
##ipdata$Charlson_index___Romano_adaptation0_1 <- rep(0, nrow(ipdata))
##ipdata$Charlson_index___Romano_adaptation0_1[ipdata$Charlson_index___Romano_adaptation <= 1] <- 1
ipdata$Charlson_index___Romano_adaptation2_4 <- rep(0, nrow(ipdata))
ipdata$Charlson_index___Romano_adaptation2_4[ipdata$Charlson_index___Romano_adaptation <= 4 & ipdata$Charlson_index___Romano_adaptation >=2] <- 1
ipdata$Charlson_index___Romano_adaptation5p <- rep(0, nrow(ipdata))
ipdata$Charlson_index___Romano_adaptation5p[ipdata$Charlson_index___Romano_adaptation >= 5] <- 1
ipdata <- ipdata[!colnames(ipdata) == 'Charlson_index___Romano_adaptation']
}
# Age categories: 18-65, 65-80, and 80
if('age_in_years' %in% colnames(ipdata)){
ParallelLogger::logInfo('Processing age into categories')
# make 18-64 the reference
##ipdata$age18_64 <- rep(0, nrow(ipdata))
##ipdata$age18_64[ipdata$age_in_years >= 0 & ipdata$age_in_years <= 64] <- 1
ipdata$age65_80 <- rep(0, nrow(ipdata))
ipdata$age65_80[ipdata$age_in_years >= 65 & ipdata$age_in_years <= 79] <- 1
ipdata$age80p <- rep(0, nrow(ipdata))
ipdata$age80p[ipdata$age_in_years >= 80] <- 1
ipdata <- ipdata[,colnames(ipdata) != 'age_in_years']
}
# make the female gender the reference
ipdata <- ipdata[,colnames(ipdata) != 'gender___FEMALE']
# save the data:
utils::write.csv(
x = ipdata,
file = file.path(outputFolder,wave,'data.csv'),
row.names = F
)
# add the data covariates
if(!is.null(analysisList$studySettings$dataCovariateSettings)){
dataSummary <- do.call(
'dataSummary',
list(
dataCovariateSettings = analysisList$studySettings$dataCovariateSettings,
databaseDetails = databaseDetails
)
)
utils::write.csv(
x = dataSummary,
file = file.path(outputFolder,wave, 'dataSummary.csv'),
row.names = F
)
}
} # end if not null
} # end wave
}
# Step 1: lead site creates the controls
if(createControl){
for(wave in waves){
ParallelLogger::logInfo(paste0('Creating the control settings for wave ', wave))
#check the data exist to get the names
if(file.exists(file.path(outputFolder, wave, 'dataSummary.csv'))){
dataSummary <- utils::read.csv(file.path(outputFolder, wave, 'dataSummary.csv'))
} else{
dataSummary <- NULL
}
#check the data exist to get the names
if(file.exists(file.path(outputFolder, wave, 'data.csv'))){
data <- utils::read.csv(file.path(outputFolder, wave, 'data.csv'))
} else{
stop('Please generate data before creating control')
}
control <- list(
project_name = paste(analysisList$packageName, wave, sep = '_'),
step = 'initialize',
sites = siteIds,
heterogeneity = analysisList$studySettings$control$heterogeneity,
model = analysisList$studySettings$control$model,
family = analysisList$studySettings$control$family,
outcome = "outcome",
variables = colnames(data)[colnames(data)!='outcome'],
variables_site_level = names(dataSummary),
optim_maxit = analysisList$studySettings$control$optim_maxit,
lead_site = siteId,
upload_date = as.character(Sys.time())
)
control <- jsonlite::toJSON(
x = control,
pretty = T,
digits = 23,
auto_unbox=TRUE,
null = "null"
)
write(control, file.path(outputFolder, wave, 'control.json'))
ParallelLogger::logInfo(
paste0(
'Initial control created at: ',
file.path(outputFolder, wave, 'control.json'),
' please upload this to ', website
)
)
}
}
if(runInitialize | runDerive | runEstimate | runSynthesize){
# json control file
jsonFileLocation <- list(
alpha = file.path(outputFolder,'alpha', 'control.json'),
delta = file.path(outputFolder,'delta', 'control.json')
)
}
if(runInitialize){
for(wave in waves){
control <- tryCatch(
{readChar(jsonFileLocation[[wave]], file.info(jsonFileLocation[[wave]])$size)},
error= function(cond) {
ParallelLogger::logInfo('Issue with loading json file...');
ParallelLogger::logError(cond);
return(NULL)
})
if(!is.null(control)){
control <- jsonlite::fromJSON(control)
# check control name
if(length(grep(wave, control$project_name))==0){
ParallelLogger::logWarn('Control project_name suggests incorrect wave')
}
ipdata <- utils::read.csv(file.path(outputFolder, wave, 'data.csv'))
# get summary variables
if(file.exists(file.path(outputFolder, wave, 'dataSummary.csv'))){
dataSummary <- utils::read.csv(file.path(outputFolder, wave, 'dataSummary.csv'))
} else{
dataSummary <- NULL
}
if(!is.null(control)){
if(control$step == 'initialize'){
ParallelLogger::logInfo(paste0('At step ', control$step))
pda::pda(
ipdata = ipdata,
hosdata = dataSummary,
site_id = siteId,
dir = file.path(outputFolder, wave)
)
ParallelLogger::logInfo(paste0('result json ready to check in ', file.path(outputFolder, wave)))
ParallelLogger::logInfo(paste0('If you are happy to share, please upload this to ', website))
} # control exists
} else{
ParallelLogger::logInfo('Control is not for initialize - please check control stage')
}
} else {
ParallelLogger::logInfo(paste0('No control for wave ', wave))
}
} # end wave
}
if(runDerive){
for(wave in waves){
control <- tryCatch(
{readChar(jsonFileLocation[[wave]], file.info(jsonFileLocation[[wave]])$size)},
error= function(cond) {
ParallelLogger::logInfo('Issue with loading json file...');
ParallelLogger::logError(cond);
return(NULL)
})
if(!is.null(control)){
control <- jsonlite::fromJSON(control)
# check control name
if(length(grep(wave, control$project_name))==0){
ParallelLogger::logWarn('Control project_name suggests incorrect wave')
}
ipdata <- utils::read.csv(file.path(outputFolder, wave, 'data.csv'))
# get summary variables
if(file.exists(file.path(outputFolder, wave, 'dataSummary.csv'))){
dataSummary <- utils::read.csv(file.path(outputFolder, wave, 'dataSummary.csv'))
} else{
dataSummary <- NULL
}
if(!is.null(control)){
if(control$step == 'derive'){
ParallelLogger::logInfo(paste0('At step ', control$step))
pda::pda(
ipdata = ipdata,
hosdata = dataSummary,
site_id = siteId,
dir = file.path(outputFolder, wave)
)
ParallelLogger::logInfo(paste0('result json ready to check in ', file.path(outputFolder, wave)))
ParallelLogger::logInfo(paste0('If you are happy to share, please upload this to ', website))
} # control exists
} else{
ParallelLogger::logInfo('Control is not for derive - please check control stage')
}
} else {
ParallelLogger::logInfo(paste0('No control for wave ', wave))
}
} # end wave
}
if(runEstimate){
for(wave in waves){
control <- tryCatch(
{readChar(jsonFileLocation[[wave]], file.info(jsonFileLocation[[wave]])$size)},
error= function(cond) {
ParallelLogger::logInfo('Issue with loading json file...');
ParallelLogger::logError(cond);
return(NULL)
})
if(!is.null(control)){
control <- jsonlite::fromJSON(control)
# check control name
if(length(grep(wave, control$project_name))==0){
ParallelLogger::logWarn('Control project_name suggests incorrect wave')
}
ipdata <- utils::read.csv(file.path(outputFolder, wave, 'data.csv'))
# get summary variables
if(file.exists(file.path(outputFolder, wave, 'dataSummary.csv'))){
dataSummary <- utils::read.csv(file.path(outputFolder, wave, 'dataSummary.csv'))
} else{
dataSummary <- NULL
}
if(!is.null(control)){
if(control$step == 'estimate'){
ParallelLogger::logInfo(paste0('At step ', control$step))
pda::pda(
ipdata = ipdata,
hosdata = dataSummary,
site_id = siteId,
dir = file.path(outputFolder, wave)
)
ParallelLogger::logInfo(paste0('result json ready to check in ', file.path(outputFolder, wave)))
ParallelLogger::logInfo(paste0('If you are happy to share, please upload this to ', website))
} # control exists
} else{
ParallelLogger::logInfo('Control is not for estimate - please check control stage')
}
} else {
ParallelLogger::logInfo(paste0('No control for wave ', wave))
}
} # end wave
}
if(runSynthesize){
#ipdata <- utils::read.csv(file.path(outputFolder, 'data'))
# do something with this?
#cbind(b.odal=fit.odal$btilde,
# sd.odal=sqrt(diag(solve(fit.odal$Htilde)/nrow(ipdata))))
}
invisible(NULL)
}
makeFriendlyNames <- function(columnNames){
columnNames <- gsub("[[:punct:]]", " ", columnNames)
columnNames <- gsub(" ", "_", columnNames)
return(columnNames)
}
ohdsi-studies/dGEMcovid documentation built on May 10, 2023, 3:38 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions makeFriendlyNames execute
Documented in execute
# Copyright 2020 Observational Health Data Sciences and Informatics
#
# This file is part of dGEMcovid
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#' Execute the Study
#'
#' @details
#' This function executes the dGEMcovid Study.
#'
#' @param databaseDetails The connection details and OMOP CDM details. Created using \code{PatientLevelPrediction::createDatabaseDetails}.
#' @param siteId The name of your site (can be the university name, site with a number, etc.) tnis needs to be shared with the study administrator
#' @param outputFolder Name of local folder to place results - make sure control is in here; make sure to use forward slashes
#' (/). Do not use a folder on a network drive since this greatly impacts
#' performance.
#' @param createCohorts Create the cohortTable table with the target population and outcome cohorts?
#' @param createData Create the labelled data set (this is required before runAnalysis)
#' @param createPlot Plot the number of covid hospitalizations by month
#' @param sampleSize The number of patients in the target cohort to sample (if NULL uses all patients)
#' @param createControl (for the lead site only) Run this code to create the study control
#' @param siteIds (for the lead site only) vector with the names of all the sites contributing to the study. Required when createControl = TRUE
#' @param runInitialize Runs the initialization step - this require downloading the json control from https://pda-ota.pdamethods.org/
#' @param runDerive Runs the derive step - this require downloading the updated json control from https://pda-ota.pdamethods.org/
#' @param runEstimate Runs the estimate step - this require downloading the updated json control from https://pda-ota.pdamethods.org/
#' @param runSynthesize Once the site estimates are returned, it is now possible to apply each model to the data to calculate
#' predictions. This step requires downloading the updated json control from https://pda-ota.pdamethods.org/
#' @param verbosity Sets the level of the verbosity. If the log level is at or higher in priority than the logger threshold, a message will print. The levels are:
#' \itemize{
#' \item{DEBUG}{Highest verbosity showing all debug statements}
#' \item{TRACE}{Showing information about start and end of steps}
#' \item{INFO}{Show informative information (Default)}
#' \item{WARN}{Show warning messages}
#' \item{ERROR}{Show error messages}
#' \item{FATAL}{Be silent except for fatal errors}
#' }
#' @param alphaStart The start date for the alpha wave in the data in "yyyymmdd" format
#' @param alphaEnd The end date for the alpha wave in the data in "yyyymmdd" format
#' @param deltaStart The start date for the delta wave in the data in "yyyymmdd" format
#' @param deltaEnd The end date for the detla wave in the data in "yyyymmdd" format
#'
#' @examples
#' \dontrun{
#' connectionDetails <- createConnectionDetails(dbms = "postgresql",
#' user = "joe",
#' password = "secret",
#' server = "myserver")
#'
#' execute(databaseDetails,
#' siteId = 'site_1',
#' outputFolder = "c:/temp/study_results",
#' createCohorts = T,
#' createData = T,
#' sampleSize = 10000,
#' createControl = F,
#' siteIds = c('site_1', 'site_2'),
#' leadSiteNextStep = F,
#' runAnalysis = F,
#' runSynthesize = F,
#' verbosity = "INFO"
#' )
#' }
#'
#' @export
execute <- function(
databaseDetails,
siteId = 'site_name',
outputFolder,
createCohorts = F,
createData = F,
createPlot = createData,
sampleSize = NULL,
createControl = F,
siteIds = '', # only needed if lead site
leadSiteNextStep = F,
runInitialize = F,
runDerive = F,
runEstimate = F,
runSynthesize = F,
verbosity = "INFO",
alphaStart = '20201101',
alphaEnd = '20210201',
deltaStart = '20210701',
deltaEnd = '20211001'
) {
dates <- list(
alpha = list(startDate = alphaStart, endDate = alphaEnd),
delta = list(startDate = deltaStart, endDate = deltaEnd)
)
waves <- c('alpha','delta')
if (!file.exists(outputFolder)){
dir.create(outputFolder, recursive = TRUE)
}
website <- 'https://pda-ota.pdamethods.org/'
# load the analysis
analysisListFile <- system.file(
"settings",
"analysisList.json",
package = "dGEMcovid"
)
if(file.exists(analysisListFile)){
analysisList <- loadJson(analysisListFile)
analysisList <- fromJsonFormat(analysisList)
analysisList$studySettings$covariateSettings <- addCohortSettings(
covariateSettings = analysisList$studySettings$covariateSettings,
cohortDatabaseSchema = databaseDetails$cohortDatabaseSchema ,
cohortTable = databaseDetails$cohortTable
)
}else{
stop('No analysisList available')
}
if(createCohorts) {
ParallelLogger::logInfo("Creating cohorts")
createCohorts(
databaseDetails = databaseDetails,
outputFolder = outputFolder
)
}
if(createPlot){
ParallelLogger::logInfo('Generating date plot')
generateDatePlot(
databaseDetails = databaseDetails,
outputFolder = outputFolder
)
}
if(createData){
# save the settings
timeSettings <- data.frame(
wave = names(dates),
starts = unlist(lapply(dates, function(x) x$startDate)),
ends = unlist(lapply(dates, function(x) x$endDate))
)
write.csv(timeSettings, file.path(outputFolder, 'timeSettings.csv'), row.names = F)
for(wave in waves){
if(!dir.exists(file.path(outputFolder, wave))){
dir.create(file.path(outputFolder, wave), recursive = T)
}
# restrict to the input date:
ParallelLogger::logInfo(paste0(wave, " date: ",dates[[wave]]$startDate, "-", dates[[wave]]$endDate ))
analysisList$studySettings$restrictPlpDataSettings$studyStartDate <- dates[[wave]]$startDate
analysisList$studySettings$restrictPlpDataSettings$studyEndDate <- dates[[wave]]$endDate
ParallelLogger::logInfo("Creating labelled dataset")
databaseDetails$cohortId <- analysisList$studySettings$targetId
databaseDetails$outcomeIds <- analysisList$studySettings$outcomeId
plpData <- tryCatch(
{
PatientLevelPrediction::getPlpData(
databaseDetails = databaseDetails,
covariateSettings = analysisList$studySettings$covariateSettings,
restrictPlpDataSettings = analysisList$studySettings$restrictPlpDataSettings
)},
error = function(e){ParallelLogger::logInfo(e); return(NULL)}
)
if(!is.null(plpData)){
labels <- PatientLevelPrediction::createStudyPopulation(
plpData = plpData,
outcomeId = analysisList$studySettings$outcomeId,
populationSettings = analysisList$studySettings$populationSettings
)
# convert to matrix
dataObject <- PatientLevelPrediction::toSparseM(
plpData = plpData,
cohort = labels
)
#sparse matrix: dataObject$dataMatrix
#labels: dataObject$labels
columnDetails <- as.data.frame(dataObject$covariateRef)
cnames <- columnDetails$covariateName[order(columnDetails$columnId)]
ipMat <- as.matrix(dataObject$dataMatrix)
ipdata <- as.data.frame(ipMat)
colnames(ipdata) <- makeFriendlyNames(cnames)
ipdata$outcome <- dataObject$labels$outcomeCount
# modify the covariates
# Charlson comorbidity categories: 0-1, 2-4, and 5
if('Charlson_index___Romano_adaptation' %in% colnames(ipdata)){
ParallelLogger::logInfo('Processing Charlson index into categories')
# Charlson comorbidity categories: 0-1, 2-4, and 5
# make 0-1 the reference
##ipdata$Charlson_index___Romano_adaptation0_1 <- rep(0, nrow(ipdata))
##ipdata$Charlson_index___Romano_adaptation0_1[ipdata$Charlson_index___Romano_adaptation <= 1] <- 1
ipdata$Charlson_index___Romano_adaptation2_4 <- rep(0, nrow(ipdata))
ipdata$Charlson_index___Romano_adaptation2_4[ipdata$Charlson_index___Romano_adaptation <= 4 & ipdata$Charlson_index___Romano_adaptation >=2] <- 1
ipdata$Charlson_index___Romano_adaptation5p <- rep(0, nrow(ipdata))
ipdata$Charlson_index___Romano_adaptation5p[ipdata$Charlson_index___Romano_adaptation >= 5] <- 1
ipdata <- ipdata[!colnames(ipdata) == 'Charlson_index___Romano_adaptation']
}
# Age categories: 18-65, 65-80, and 80
if('age_in_years' %in% colnames(ipdata)){
ParallelLogger::logInfo('Processing age into categories')
# make 18-64 the reference
##ipdata$age18_64 <- rep(0, nrow(ipdata))
##ipdata$age18_64[ipdata$age_in_years >= 0 & ipdata$age_in_years <= 64] <- 1
ipdata$age65_80 <- rep(0, nrow(ipdata))
ipdata$age65_80[ipdata$age_in_years >= 65 & ipdata$age_in_years <= 79] <- 1
ipdata$age80p <- rep(0, nrow(ipdata))
ipdata$age80p[ipdata$age_in_years >= 80] <- 1
ipdata <- ipdata[,colnames(ipdata) != 'age_in_years']
}
# make the female gender the reference
ipdata <- ipdata[,colnames(ipdata) != 'gender___FEMALE']
# save the data:
utils::write.csv(
x = ipdata,
file = file.path(outputFolder,wave,'data.csv'),
row.names = F
)
# add the data covariates
if(!is.null(analysisList$studySettings$dataCovariateSettings)){
dataSummary <- do.call(
'dataSummary',
list(
dataCovariateSettings = analysisList$studySettings$dataCovariateSettings,
databaseDetails = databaseDetails
)
)
utils::write.csv(
x = dataSummary,
file = file.path(outputFolder,wave, 'dataSummary.csv'),
row.names = F
)
}
} # end if not null
} # end wave
}
# Step 1: lead site creates the controls
if(createControl){
for(wave in waves){
ParallelLogger::logInfo(paste0('Creating the control settings for wave ', wave))
#check the data exist to get the names
if(file.exists(file.path(outputFolder, wave, 'dataSummary.csv'))){
dataSummary <- utils::read.csv(file.path(outputFolder, wave, 'dataSummary.csv'))
} else{
dataSummary <- NULL
}
#check the data exist to get the names
if(file.exists(file.path(outputFolder, wave, 'data.csv'))){
data <- utils::read.csv(file.path(outputFolder, wave, 'data.csv'))
} else{
stop('Please generate data before creating control')
}
control <- list(
project_name = paste(analysisList$packageName, wave, sep = '_'),
step = 'initialize',
sites = siteIds,
heterogeneity = analysisList$studySettings$control$heterogeneity,
model = analysisList$studySettings$control$model,
family = analysisList$studySettings$control$family,
outcome = "outcome",
variables = colnames(data)[colnames(data)!='outcome'],
variables_site_level = names(dataSummary),
optim_maxit = analysisList$studySettings$control$optim_maxit,
lead_site = siteId,
upload_date = as.character(Sys.time())
)
control <- jsonlite::toJSON(
x = control,
pretty = T,
digits = 23,
auto_unbox=TRUE,
null = "null"
)
write(control, file.path(outputFolder, wave, 'control.json'))
ParallelLogger::logInfo(
paste0(
'Initial control created at: ',
file.path(outputFolder, wave, 'control.json'),
' please upload this to ', website
)
)
}
}
if(runInitialize | runDerive | runEstimate | runSynthesize){
# json control file
jsonFileLocation <- list(
alpha = file.path(outputFolder,'alpha', 'control.json'),
delta = file.path(outputFolder,'delta', 'control.json')
)
}
if(runInitialize){
for(wave in waves){
control <- tryCatch(
{readChar(jsonFileLocation[[wave]], file.info(jsonFileLocation[[wave]])$size)},
error= function(cond) {
ParallelLogger::logInfo('Issue with loading json file...');
ParallelLogger::logError(cond);
return(NULL)
})
if(!is.null(control)){
control <- jsonlite::fromJSON(control)
# check control name
if(length(grep(wave, control$project_name))==0){
ParallelLogger::logWarn('Control project_name suggests incorrect wave')
}
ipdata <- utils::read.csv(file.path(outputFolder, wave, 'data.csv'))
# get summary variables
if(file.exists(file.path(outputFolder, wave, 'dataSummary.csv'))){
dataSummary <- utils::read.csv(file.path(outputFolder, wave, 'dataSummary.csv'))
} else{
dataSummary <- NULL
}
if(!is.null(control)){
if(control$step == 'initialize'){
ParallelLogger::logInfo(paste0('At step ', control$step))
pda::pda(
ipdata = ipdata,
hosdata = dataSummary,
site_id = siteId,
dir = file.path(outputFolder, wave)
)
ParallelLogger::logInfo(paste0('result json ready to check in ', file.path(outputFolder, wave)))
ParallelLogger::logInfo(paste0('If you are happy to share, please upload this to ', website))
} # control exists
} else{
ParallelLogger::logInfo('Control is not for initialize - please check control stage')
}
} else {
ParallelLogger::logInfo(paste0('No control for wave ', wave))
}
} # end wave
}
if(runDerive){
for(wave in waves){
control <- tryCatch(
{readChar(jsonFileLocation[[wave]], file.info(jsonFileLocation[[wave]])$size)},
error= function(cond) {
ParallelLogger::logInfo('Issue with loading json file...');
ParallelLogger::logError(cond);
return(NULL)
})
if(!is.null(control)){
control <- jsonlite::fromJSON(control)
# check control name
if(length(grep(wave, control$project_name))==0){
ParallelLogger::logWarn('Control project_name suggests incorrect wave')
}
ipdata <- utils::read.csv(file.path(outputFolder, wave, 'data.csv'))
# get summary variables
if(file.exists(file.path(outputFolder, wave, 'dataSummary.csv'))){
dataSummary <- utils::read.csv(file.path(outputFolder, wave, 'dataSummary.csv'))
} else{
dataSummary <- NULL
}
if(!is.null(control)){
if(control$step == 'derive'){
ParallelLogger::logInfo(paste0('At step ', control$step))
pda::pda(
ipdata = ipdata,
hosdata = dataSummary,
site_id = siteId,
dir = file.path(outputFolder, wave)
)
ParallelLogger::logInfo(paste0('result json ready to check in ', file.path(outputFolder, wave)))
ParallelLogger::logInfo(paste0('If you are happy to share, please upload this to ', website))
} # control exists
} else{
ParallelLogger::logInfo('Control is not for derive - please check control stage')
}
} else {
ParallelLogger::logInfo(paste0('No control for wave ', wave))
}
} # end wave
}
if(runEstimate){
for(wave in waves){
control <- tryCatch(
{readChar(jsonFileLocation[[wave]], file.info(jsonFileLocation[[wave]])$size)},
error= function(cond) {
ParallelLogger::logInfo('Issue with loading json file...');
ParallelLogger::logError(cond);
return(NULL)
})
if(!is.null(control)){
control <- jsonlite::fromJSON(control)
# check control name
if(length(grep(wave, control$project_name))==0){
ParallelLogger::logWarn('Control project_name suggests incorrect wave')
}
ipdata <- utils::read.csv(file.path(outputFolder, wave, 'data.csv'))
# get summary variables
if(file.exists(file.path(outputFolder, wave, 'dataSummary.csv'))){
dataSummary <- utils::read.csv(file.path(outputFolder, wave, 'dataSummary.csv'))
} else{
dataSummary <- NULL
}
if(!is.null(control)){
if(control$step == 'estimate'){
ParallelLogger::logInfo(paste0('At step ', control$step))
pda::pda(
ipdata = ipdata,
hosdata = dataSummary,
site_id = siteId,
dir = file.path(outputFolder, wave)
)
ParallelLogger::logInfo(paste0('result json ready to check in ', file.path(outputFolder, wave)))
ParallelLogger::logInfo(paste0('If you are happy to share, please upload this to ', website))
} # control exists
} else{
ParallelLogger::logInfo('Control is not for estimate - please check control stage')
}
} else {
ParallelLogger::logInfo(paste0('No control for wave ', wave))
}
} # end wave
}
if(runSynthesize){
#ipdata <- utils::read.csv(file.path(outputFolder, 'data'))
# do something with this?
#cbind(b.odal=fit.odal$btilde,
# sd.odal=sqrt(diag(solve(fit.odal$Htilde)/nrow(ipdata))))
}
invisible(NULL)
}
makeFriendlyNames <- function(columnNames){
columnNames <- gsub("[[:punct:]]", " ", columnNames)
columnNames <- gsub(" ", "_", columnNames)
return(columnNames)
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.