adjust for deleted files and validate UUIDs

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: tgR.data
 Type: Package
 Title: Data Management Support for Fred Hutch Translational Genomics
-Version: 0.0.2
+Version: 0.0.3
 Author: Amy Paguirigan
 Maintainer: Amy Paguirigan <[email protected]>
 Description: Supports the curation of scientific metadata associated with files stored

NAMESPACE

@@ -5,6 +5,7 @@ export(check_credentials)
 export(commit_data_provenance)
 export(delete_file_entries)
 export(dropWhen)
+export(get_all_deleted_files)
 export(get_data_provenance)
 export(get_dataset_annotations)
 export(get_deleted_files)
@@ -13,4 +14,5 @@ export(mark_file_for_deletion)
 export(set_credentials)
 export(tag_and_commit_data_provenance)
 export(tag_file)
+export(validate_uuid)
 importFrom(magrittr,"%>%")

R/delete_file_entries.R

@@ -6,7 +6,6 @@
 delete_file_entries <- function(uuid = NULL) {
   suppressMessages(check_credentials())
   if(is.null(uuid)) {stop("Please provide uuid(s) for file entries to delete.")}
-  message("Deleting records cannot be undone. ")
   formData <- list("token"=Sys.getenv("S3META"),
                    content='record',
                    action='delete',

R/get_deleted_files.R

@@ -24,10 +24,11 @@ get_deleted_files <- function(bucket = NULL, prefix = NULL, DAG=NULL, file_type
                    format='csv',
                    type='flat',
                    csvDelimiter=',',
-                   'fields[0]'='object_prefix',
-                   'fields[1]'='bucket_name',
-                   'fields[2]'='bucket_prefix',
-                   'fields[3]'='deleted_object',
+                   'fields[0]'='uuid',
+                   'fields[1]'='object_prefix',
+                   'fields[2]'='bucket_name',
+                   'fields[3]'='bucket_prefix',
+                   'fields[4]'='deleted_object',
                    'forms[0]'='data_provenance',
                    rawOrLabel='raw',
                    rawOrLabelHeaders='raw',
@@ -46,3 +47,40 @@ get_deleted_files <- function(bucket = NULL, prefix = NULL, DAG=NULL, file_type
   } else {results <- data.frame()}
   return(results)
 }
+
+#' Find all files that used to be in S3 that are marked for deletion of their data provenance
+#'
+#' @return A data frame containing metadata about all files marked for deletion of data provenance. Requesting data from a specific prefix can speed up the request.
+#' @details Requires admin credentials.
+#' @export
+
+
+get_all_deleted_files <- function() {
+  check_credentials()
+  formData <- list("token"=Sys.getenv("S3META"),
+                   content='record',
+                   action='export',
+                   format='csv',
+                   type='flat',
+                   csvDelimiter=',',
+                   'fields[0]'='uuid',
+                   'fields[1]'='object_prefix',
+                   'fields[2]'='bucket_name',
+                   'fields[3]'='bucket_prefix',
+                   'fields[4]'='deleted_object',
+                   'forms[0]'='data_provenance',
+                   rawOrLabel='raw',
+                   rawOrLabelHeaders='raw',
+                   exportCheckboxLabel='false',
+                   exportSurveyFields='false',
+                   exportDataAccessGroups='true',
+                   returnFormat='csv',
+                   filterLogic="[deleted_object] = '1'"
+  )
+  results <- suppressMessages(httr::content(httr::POST(url = Sys.getenv("REDURI"), body = formData, encode = "form")))
+  message(paste0("Retrieved ", nrow(results), " records."))
+  if(nrow(results)>0) {
+    results <- results %>% dplyr::select(-dplyr::contains("_complete"))
+  } else {results <- data.frame()}
+  return(results)
+}

R/validate_uuid.R

@@ -0,0 +1,33 @@
+#' Checks to ensure a uuid is not already used
+#'
+#' @param uuid (Required) uuid you'd like to check
+#' @return Returns NA if all uuids are valid, and an array of invalid uuids if any are invalid
+#' @details Note: Only works for admin users.
+#' @export
+#'
+validate_uuid <- function(uuid = NULL) {
+  if(is.null(uuid)) {stop("Please provide at least one uuid.")}
+  suppressMessages(check_credentials())
+  message("Checking for existing record(s) with the uuid(s) provided.")
+  logic <- paste0("[uuid] = '", paste(uuid, collapse = "' OR [uuid] = '"), "'")
+  formData <- list("token"=Sys.getenv("S3META"),
+                   content='record',
+                   action='export',
+                   format='csv',
+                   type='flat',
+                   csvDelimiter=',',
+                   'fields[0]'='uuid',
+                   rawOrLabel='raw',
+                   rawOrLabelHeaders='raw',
+                   exportCheckboxLabel='false',
+                   exportSurveyFields='false',
+                   exportDataAccessGroups='false',
+                   returnFormat='csv',
+                   filterLogic=logic)
+  alreadyUsed <- suppressMessages(httr::content(
+    httr::POST(url = Sys.getenv("REDURI"),
+               body = formData, encode = "form", show_col_types = FALSE)))
+
+  if (nrow(alreadyUsed)>0) {result <- alreadyUsed$uuid} else {result <- NA}
+  return(result)
+}

docker/dockerfile

@@ -1,4 +1,4 @@
-### vortexing/r_tgr.data:v0.0.2
+### vortexing/r_tgr.data:v0.0.3
 ## Pull the R base image
 FROM rocker/r-base:4.1.1
 # Update
@@ -7,6 +7,6 @@ RUN apt-get install -y libssl-dev libxml2-dev libcurl4-openssl-dev
 # Install required packages
 RUN R -e "install.packages('remotes', dependencies = TRUE)"
 # Install from github
-RUN R -e "remotes::install_github('FredHutch/[email protected].2', dependencies = TRUE)"
+RUN R -e "remotes::install_github('FredHutch/[email protected].3', dependencies = TRUE)"
 # Override the R CMD for workflow managers
-CMD ["/bin/bash"]
+CMD ["/bin/bash"]