view elastic_tool.Rmd @ 7:01eaa66fcc2e draft

planemo upload

---
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'))
```