From fd9c1ddd33eba8aeba97d3041cbe5b48ba40241c Mon Sep 17 00:00:00 2001
From: lakikowolfe <loren.wolfe@sagebase.org>
Date: Thu, 25 Apr 2024 15:12:10 -0700
Subject: [PATCH] fix scheduled job update fxns

---
 R/manifest.R | 49 +++++++++++++++++++------------------------------
 1 file changed, 19 insertions(+), 30 deletions(-)

diff --git a/R/manifest.R b/R/manifest.R
index ab45174..9073cea 100644
--- a/R/manifest.R
+++ b/R/manifest.R
@@ -206,21 +206,22 @@ fill_dataflow_manifest <- function(dataflow_manifest_chunk,
         access_token = access_token
         )
     })
-  }
 
-  # Convert the named vector into a dataframe
-  study_status_df <- data.frame(
-    contributor_id = names(study_status),
-    study_status = study_status
-  )
+    # Convert the named vector into a dataframe
+    study_status_df <- data.frame(
+      contributor_id = names(study_status),
+      study_status = study_status
+    )
 
-  # merge study status into dataflow manifest chunk
-  dataflow_manifest_chunk <- merge(
-    x = dataflow_manifest_chunk,
-    y = study_status_df,
-    by = "contributor_id",
-    all.x = TRUE
+    # merge study status into dataflow manifest chunk
+    dataflow_manifest_chunk <- merge(
+      x = dataflow_manifest_chunk,
+      y = study_status_df,
+      by = "contributor_id",
+      all.x = TRUE
     )
+  }
+
 
   # find attributes that are not present in provided manifest chunk
   missing_attributes_df <- attributes_df[!attributes_df$Attribute %in% names(dataflow_manifest_chunk), ]
@@ -246,6 +247,10 @@ fill_dataflow_manifest <- function(dataflow_manifest_chunk,
   # FIXME: is this what we want to happen if data flow schema changes?
   dataflow_manifest <- dplyr::bind_cols(dataflow_manifest_chunk, missing_attributes_filled)
 
+  # REMOVE contributor_id column
+  contributor_id_idx <- grep("contributor_id", names(dataflow_manifest))
+  dataflow_manifest <- dataflow_manifest[, -contributor_id_idx]
+
   # return filled columns with original manifest chunk
   return(dataflow_manifest)
 }
@@ -324,22 +329,6 @@ update_data_flow_manifest <- function(asset_view,
     }
   )
 
-  # synapse_manifests <- tryCatch(
-  #   {
-  #     get_all_manifests(
-  #       asset_view = asset_view,
-  #       na_replace = na_replace,
-  #       access_token = access_token,
-  #       base_url = base_url,
-  #       verbose = FALSE
-  #     )
-  #   },
-  #   error = function(e) {
-  #     message("get_all_manifests failed")
-  #     message(e)
-  #   }
-  # )
-
   # check synapse for new datasets
   dataflow_manifest_updated <- update_manifest_add_datasets(
     dataflow_manifest = dataflow_manifest,
@@ -463,7 +452,7 @@ update_manifest_add_datasets <- function(dataflow_manifest,
       }
     )
 
-    new_datasets$num_items <- num_items
+    new_datasets$num_items <- as.integer(num_items)
 
     # fill data flow manifest rows for missing datasets
     new_datasets <- fill_dataflow_manifest(
@@ -537,7 +526,7 @@ update_manifest_column <- function(dataflow_manifest,
   idx <- dataflow_manifest[, update_column] != get_all_manifests_out[, update_column]
 
   # if any items have changed update dataset type column
-  if (any(idx)) {
+  if (any(isTRUE(idx))) {
     n_changed <- sum(idx)
     print(paste0("Making ", n_changed, " update(s) to ", update_column, " column"))
     dataflow_manifest[idx, update_column] <- get_all_manifests_out[idx, update_column]