Mercurial > repos > mingchen0919 > elastic_r_tidyverse
changeset 9:06f38d3febdc draft default tip
planemo upload
| author | mingchen0919 |
|---|---|
| date | Wed, 28 Mar 2018 23:34:03 -0400 |
| parents | ea3271940757 |
| children | |
| files | elastic_r_package.Rmd elastic_r_package_render.R elastic_r_package_wrappers_macros.xml elastic_tool.Rmd |
| diffstat | 4 files changed, 150 insertions(+), 178 deletions(-) [+] |
line wrap: on
line diff
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/elastic_r_package.Rmd Wed Mar 28 23:34:03 2018 -0400 @@ -0,0 +1,110 @@ +--- +title: 'Tool Report' +output: html_document +--- + +<style> +pre code, pre, code { + white-space: pre !important; + overflow-x: scroll !important; + word-break: keep-all !important; + word-wrap: initial !important; +} +</style> + +```{r setup, include=FALSE, warning=FALSE, message=FALSE} +knitr::opts_chunk$set(error = TRUE) +``` + +## User input + +```{r, echo=FALSE} +knitr::kable(arguments) +``` + + +```{r, echo=FALSE} +# each tool execution runs one or multiple chained functions but generate only one single object. +# we save this object to an rdata file and output this file to galaxy history so that it can be used by other tools +# we can use this rdata output file's dataset id as the variable name of the saved object. +job_script_path = paste0(Sys.getenv('REPORT_FILES_PATH'), '/job-script.R') +tool_rdata_output = Sys.getenv('TOOL_RDATA_OUTPUT') +dataset_id = tail(strsplit(tool_rdata_output, '/')[[1]], 1) +dataset_num = gsub("(.+_)([0-9]+)\\.dat", "\\2", dataset_id) +rdata_id = paste0('rdata_', dataset_num) + +## build script +# the first line of the job script is 'rdata_NUM = ', where 'NUM' is the dataset number of the output rdata. +write(paste0(rdata_id, ' = '), file = job_script_path) +# loop through argument data frame to build up the job script. +for (i in 1: (nrow(arguments)-1)) { + row_type = arguments[i, 'row_type'] + switch (row_type, + # if it's a function row, the line has format 'function_name(' + func = write(paste0(arguments[i, 'function_name'], '('), + file = job_script_path, + append = TRUE ), + + + argument = { + # if it's an argument row and the next row is not an operator row, + # the line has format ' argument_name=argument_value,' + if (arguments[i+1, 'operator'] == "") { + write(paste0(' ', arguments[i, 'argument_name'], '=', arguments[i, 'argument_value'], ','), + file = job_script_path, + append = TRUE ) + } else { + # if it's an argument row and the next row IS an operator row, + # the line has format ' argument_name=argument_value'. note that there is not comma at the end. + write(paste0(' ', arguments[i, 'argument_name'], '=', arguments[i, 'argument_value']), + file = job_script_path, + append = TRUE ) + } + }, + + # if it is an operator row, the line has format ') operator' + operator = write(paste0(') ', arguments[i, 'operator']), + file = job_script_path, + append = TRUE ) + ) +} + +# the last line is missing a ')' +write(')', file = job_script_path, append = TRUE) +``` + + +```{bash, 'display script', results='asis', echo=FALSE} +echo '## Job script' +echo '' +echo '' +echo '```r' +cat ${REPORT_FILES_PATH}/job-script.R +echo '```' +``` + + +## Result + +```{r, 'run job script', echo=FALSE} +source(job_script_path) +# display result. +eval(parse(text = rdata_id)) +``` + + +```{r, 'display output directory contents', results='asis', echo=FALSE} +## after the job is done, we list all files from the output directory. +## full relative path to the output directory needs to be displayed. + +cat('##All output files') +cat('\n\n') +all_files = list.files(path = Sys.getenv('REPORT_FILES_PATH'), + full.names = TRUE, + recursive = TRUE) + +for (f in sub(Sys.getenv('REPORT_FILES_PATH'), '.', all_files) ) { + cat('* [', f, '](', f, ')\n') +} +cat('\n') +``` \ No newline at end of file
--- a/elastic_r_package_render.R Tue Mar 27 23:42:53 2018 -0400 +++ b/elastic_r_package_render.R Wed Mar 28 23:34:03 2018 -0400 @@ -9,15 +9,16 @@ options(stringsAsFactors = FALSE) library(rmarkdown) +library(tidyverse) #------------------------------------------------ #------------------------------------------------ -options_and_arguments = read.table(paste0(Sys.getenv('REPORT_FILES_PATH'), '/options_and_arguments.txt'), +arguments = read.table(paste0(Sys.getenv('REPORT_FILES_PATH'), '/arguments.txt'), sep = '|', header = TRUE) #-----------------render Rmd-------------- -render(paste0(Sys.getenv('TOOL_INSTALL_DIR'), '/elastic_tool.Rmd'), +render(paste0(Sys.getenv('TOOL_INSTALL_DIR'), '/elastic_r_package.Rmd'), output_file = paste0(Sys.getenv('REPORT_FILES_PATH'), '/report.html')) # for some unknow reason, directly using REPORT as the input value for output_file parameter
--- a/elastic_r_package_wrappers_macros.xml Tue Mar 27 23:42:53 2018 -0400 +++ b/elastic_r_package_wrappers_macros.xml Wed Mar 28 23:34:03 2018 -0400 @@ -14,43 +14,49 @@ </xml> <xml name="function_components"> - <repeat name="function_component_repeat" title="Call a function" min="1" default="1"> + <repeat name="function_components" title="Call a function" min="1" default="1"> <param type="select" name="function_name" multiple="false" label="==== Call function ===="> <option value="ggplot" selected="false">ggplot</option> + <option value="geom_point" selected="false">geom_point</option> <option value="aes" selected="false">ggplot</option> <option value="read.table" selected="false">read.table</option> <option value="write.csv" selected="false">write.csv</option> </param> <expand macro="function_arguments" /> - <param type="select" name="operator" label="connect to next function"> - <option value="%>%" selected="false">pipe (%>%)</option> - <option value="+" selected="false">plus (+)</option> - </param> </repeat> </xml> <xml name="function_arguments"> - <repeat name="regular_argument_value" title="Argument value: from user input" min="0" default="0"> - <param type="text" name="argument" label="argument" /> - <param type="text" name="argument_value" label="argument value" /> - <param type="select" name="argument_type" label="argument type"> - <option value="string" selected="true">string</option> - <option value="numeric" selected="false">numeric</option> - <option value="boolean" selected="false">boolean</option> - <option value="variable" selected="false">variable</option> - </param> + <repeat name="function_arguments" title="argument name/value pairs" min="0" default="0"> + <conditional name="argument_type"> + <param name="argument_type_selector" type="select" label="Argument type"> + <option value="value" selected="false">user input value</option> + <option value="rdata" selected="false">value from RData output</option> + </param> + <when value="value"> + <param type="text" name="argument_name" label="argument name" /> + <param type="text" name="argument_value" label="argument value" /> + </when> + <when value="rdata"> + <param type="text" name="argument_name" label="argument name" /> + <param type="data" format="rdata" name="argument_value" optional="false" label="RData output" /> + </when> + </conditional> </repeat> - <repeat name="rdata_argument_value" title="Argument value: from RData" min="0" default="0"> - <param type="text" name="argument" label="argument" /> - <param type="data" format="rdata" name="argument_value" optional="false" label="RData output" /> - </repeat> + <param type="select" name="operator" label="connect to next function"> + <sanitizer> + <valid initial="string.printable"/> + </sanitizer> + <option value="%>%" selected="true">pipe (%>%)</option> + <option value="+" selected="false">plus (+)</option> + </param> </xml> <!--output set--> <xml name="output_set"> <data format="html" name="report" label="${tool.name} report"/> - <data format="txt" name="tool_help_doc" label="${tool.name} help doc"/> + <data format="rdata" name="tool_rdata_output" label="${tool.name} RData output ${on_string}"/> <data format="txt" name="tool_log" label="${tool.name} log" /> </xml> @@ -70,18 +76,20 @@ ############ 3. save user inputs into arguments.txt ############ 3.1 table header - echo 'function|input_type|argument|argument_value|argument_value_type|operator' > $report.files_path/arguments.txt && + echo 'row_type|function_name|argument_name|argument_value|argument_value_type|operator' > $report.files_path/arguments.txt && ############ 3.2 loop through function component repeat - #for $fun_component in $function_components: - #for reg_arg in $fun_component.regular_argument_value - echo '$fun_component.function_name|regular|$reg_arg.argument|$arg_arg.argument_value|$arg_arg.argument_value_type|$fun_component.operator' >> $report.files_path/arguments.txt && + #for fun_component in $function_components: + echo 'func|$fun_component.function_name||||' >> $report.files_path/arguments.txt && + + Rscript -e 'library(tidyverse); help($fun_component.function_name)' >> $report.files_path/$fun_component.function_name-help.txt && + + #for argument in $fun_component.function_arguments: + echo 'argument||$argument.argument_type.argument_name|$argument.argument_type.argument_value|$argument.argument_type.argument_type_selector|' >> $report.files_path/arguments.txt && #end for - - #for rdata_arg in $fun_component.rdata_argument_value - echo '$fun_component.function_name|rdata|$rdata_arg.argument|$arg_arg.argument_value|rdata|$fun_component.operator' >> $report.files_path/arguments.txt && - #end for + echo 'operator|||||$fun_component.operator' >> $report.files_path/arguments.txt && + #end for ############ output_set ## '$report' @@ -91,12 +99,12 @@ ####################### export REPORT='$report' && export REPORT_FILES_PATH='$report.files_path' && - export TOOL_HELP_DOC='$tool_help_doc' && + export TOOL_RDATA_OUTPUT='$tool_rdata_output' && export TOOL_LOG='$tool_log' && ############ run render R script to render R markdowns - Rscript '${__tool_directory__}/elastic_tool_render.R' + Rscript '${__tool_directory__}/elastic_r_package_render.R' ]]></command> </xml>
--- a/elastic_tool.Rmd Tue Mar 27 23:42:53 2018 -0400 +++ /dev/null Thu Jan 01 00:00:00 1970 +0000 @@ -1,147 +0,0 @@ ---- -title: 'Tool Report' -output: html_document ---- - -<style> -pre code, pre, code { - white-space: pre !important; - overflow-x: scroll !important; - word-break: keep-all !important; - word-wrap: initial !important; -} -</style> - -```{r setup, include=FALSE, warning=FALSE, message=FALSE} -knitr::opts_chunk$set(error = TRUE) -``` - -## User input - -```{r, 'display user input'} -# get user input and save it into a data frame. -df = read.table(paste0(Sys.getenv('REPORT_FILES_PATH'), '/options_and_arguments.txt'), - sep = '|', header = TRUE) - -# if the input type is 'path_relative_to_a_tool', prepend A_TOOL_OUTPUT_PATH to the value to make -# the value a full path. -if (nrow(df[df$type == 'path_relative_to_a_tool', ]) > 0) { - for (i in 1:nrow(df[df$type == 'path_relative_to_a_tool', ])) { - root_path = readLines(df[df$type == 'path_relative_to_a_tool', ][i, 'path_type'])[1] - df[df$type == 'path_relative_to_a_tool', ][i, 'value'] = paste(root_path, - df[df$type == 'path_relative_to_a_tool', ][i, 'value'], - sep = '/') - } -} - -## display user input as a table -knitr::kable(df) -``` - - -```{r, 'build script', echo=FALSE} -##-------- build script files ----------- - -# get tool name, the first line of the script is always the tool name. -tool_name = df[df$type == 'tool_name', 'value'] - -# if the number of option/argument pairs is larger than 0, build script file -df2 = df[df$type != 'tool_name', ] -if (nrow(df2) > 0) { - # write tool name as the first line of the script.sh - # before running the job, cd into the ${REPORT_FILES_PATH} directory - write(paste0(tool_name, ' \\'), - paste0(Sys.getenv('REPORT_FILES_PATH'), '/script.sh')) - df2 = df[df$type != 'tool_name', ] - write(paste(' ', df2$flag, df2$value, '\\', sep = ' '), - file = paste0(Sys.getenv('REPORT_FILES_PATH'), '/script.sh'), - append = TRUE ) - # remember that after writing option/argument lines to the script.sh, the last line - # has an extra newline character '\' which causes a problem. We can either remove that extra - # '\' or add a new line to the end. We choose to add a new line. - - # add an extra line to the end to redirect stdout to stdout.txt and stderr to stderr.txt - write(' > ${REPORT_FILES_PATH}/stdout.txt 2>${REPORT_FILES_PATH}/stderr.txt', - paste0(Sys.getenv('REPORT_FILES_PATH'), '/script.sh'), append = TRUE) -} else { - # if no option/argument input, simply display the help message - write(paste0(tool_name, ' -h'), - file = paste0(Sys.getenv('REPORT_FILES_PATH'), '/script.sh')) -} -``` - - -```{r, 'create paths if they do not exist', echo=FALSE} -## if the input type is 'path_relative_to_this_tool', that means -## we need to create a directory or file path. - -# create paths before running the job script -df_paths = df[df$type == 'path_relative_to_this_tool', ] - -if (nrow(df_paths) > 0) { - for (i in 1:nrow(df_paths)) { - path = paste0(Sys.getenv('REPORT_FILES_PATH'), '/', df_paths[i, 'value']) - path_type = df_paths[i, 'path_type'] - - # create file paths - if ((path_type == 'file_path') & !file.exists(path)) { - dir_path = paste(head(strsplit(path, '/')[[1]], -1), collapse = '/' ) - if (!dir.exists(dir_path)) { - dir.create(dir_path, recursive = TRUE) - } - file.create(path) - } - - # create dir paths - if ((path_type == 'dir_path') & !dir.exists(path)) { - dir.create(path, recursive = TRUE) - } - } -} -``` - - -```{bash, 'run jobs', echo=FALSE} -# run job script, always use absolute path. -# we want to run all jobs within the working path. -sh ${REPORT_FILES_PATH}/script.sh -``` - - -```{bash, 'display script', results='asis', echo=FALSE} -echo '## Job script' -echo '' -echo '' -echo '```bash' -cat ${REPORT_FILES_PATH}/script.sh -echo '```' -``` - - -```{r, 'display output directory contents', results='asis', echo=FALSE} -## after the job is done, we list all files from the output directory. -## full relative path to the output directory needs to be displayed. - -cat('##All output files') -cat('\n\n') -all_files = list.files(path = Sys.getenv('REPORT_FILES_PATH'), - full.names = TRUE, - recursive = TRUE) - -for (f in sub(Sys.getenv('REPORT_FILES_PATH'), '.', all_files) ) { - cat('* [', f, '](', f, ')\n') -} -cat('\n') -``` - - -```{r, 'save output directory of this tool', echo=FALSE} -## each elastic tool has a galaxy history output which contains the REPORT_FILES_PATH of this tool -## so that other tools can reference the outputs from this tool. - -## obtain REPORT_FILES_PAHT and save it to a galaxy output. -database_root = paste(head(strsplit(Sys.getenv('TOOL_LOG'), '/')[[1]], -1), collapse = '/') -tool_output_dir_id = tail(strsplit(Sys.getenv('REPORT_FILES_PATH'), '/')[[1]], 1) -tool_output_dir = paste0(database_root, '/', tool_output_dir_id) -write(tool_output_dir, Sys.getenv('TOOL_OUTPUT_DIR')) -```