Merge pull request #1020 from cmu-delphi/main

Deploy Quidel Fixes to prod

Author: chinandrew

facebook/delphiFacebook/NAMESPACE

@@ -63,6 +63,7 @@ importFrom(dplyr,across)
 importFrom(dplyr,all_of)
 importFrom(dplyr,anti_join)
 importFrom(dplyr,arrange)
+importFrom(dplyr,bind_cols)
 importFrom(dplyr,bind_rows)
 importFrom(dplyr,case_when)
 importFrom(dplyr,coalesce)

facebook/delphiFacebook/R/contingency_aggregate.R

@@ -264,15 +264,10 @@ summarize_aggs <- function(df, crosswalk_data, aggregations, geo_level, params)
   }
 
   ## Find all unique groups and associated frequencies, saved in column `Freq`.
-  # Keep rows with missing values initially so that we get the correct column
-  # names. Explicitly drop groups with missing values in second step.
   unique_groups_counts <- as.data.frame(
     table(df[, group_vars, with=FALSE], exclude=NULL, dnn=group_vars), 
     stringsAsFactors=FALSE
   )
-  unique_groups_counts <- unique_groups_counts[
-    complete.cases(unique_groups_counts[, group_vars]),
-  ]
 
   # Drop groups with less than threshold sample size.
   unique_groups_counts <- filter(unique_groups_counts, Freq >= params$num_filter)
@@ -327,9 +322,10 @@ summarize_aggs <- function(df, crosswalk_data, aggregations, geo_level, params)
     aggregation <- aggregations$id[row]
     group_vars <- aggregations$group_by[[row]]
     post_fn <- aggregations$post_fn[[row]]
-
+    
+    # Keep only aggregations where the main value, `val`, is present.
     dfs_out[[aggregation]] <- dfs_out[[aggregation]][
-      rowSums(is.na(dfs_out[[aggregation]][, c("val", "sample_size", group_vars)])) == 0,
+      rowSums(is.na(dfs_out[[aggregation]][, c("val", "sample_size")])) == 0,
     ]
 
     dfs_out[[aggregation]] <- apply_privacy_censoring(dfs_out[[aggregation]], params)

facebook/delphiFacebook/R/contingency_calculate.R

@@ -13,12 +13,11 @@
 #'   be a non-integer
 #' 
 #' @return a list of named means and other descriptive statistics
-compute_numeric <- function(response, weight, sample_size, total_represented)
+compute_household_binary <- function(response, weight, sample_size, total_represented)
 {
   response_mean <- compute_count_response(response, weight, sample_size)
   response_mean$sample_size <- sample_size
   response_mean$represented <- total_represented
-  response_mean$se <- NA_real_
 
   return(response_mean)
 }