refactor: tune performance

r_package/R/data-helpers.R

@@ -75,7 +75,7 @@ generateCombinationsImpl = function(sets,
   cc = colorLookup(colors)
 
   mergeUnion = function(a, b) {
-    ab_sets = sort(union(a$setNames, b$setNames))
+    ab_sets = union(a$setNames, b$setNames)
     ab_name = paste(ab_sets, collapse = symbol)
     ab_elems = c()
     if (a$cardinality == 0) {
@@ -89,7 +89,7 @@ generateCombinationsImpl = function(sets,
   }
 
   mergeIntersect = function(a, b) {
-    ab_sets = sort(union(a$setNames, b$setNames))
+    ab_sets = union(a$setNames, b$setNames)
     ab_name = paste(ab_sets, collapse = symbol)
     ab_elems = c()
     if (a$cardinality > 0 && b$cardinality > 0) {
@@ -104,6 +104,9 @@ generateCombinationsImpl = function(sets,
     if (s$degree < min || (!is.null(max) && s$degree > max) || (s$cardinality == 0 && !empty)) {
       return()
     }
+    if (!store.elems) {
+      s <<- asCombination(s$name, c(), 'distinctIntersection', s$setNames, cardinality=s$cardinality, color=s$color)
+    }
     if (!distinct || s$degree == 1) {
       combinations <<- c(combinations, list(s))
       return()
@@ -125,7 +128,7 @@ generateCombinationsImpl = function(sets,
       return()
     }
 
-    sDistinct = asCombination(s$name, dElems, 'distinctIntersection', s$setNames, color=s$color)
+    sDistinct = asCombination(s$name, if (store.elems) { dElems } else { c() }, 'distinctIntersection', s$setNames, cardinality=length(dElems), color=s$color)
     if (sDistinct$cardinality > 0 || empty) {
       combinations <<- c(combinations, list(sDistinct))
     }
@@ -185,24 +188,30 @@ extractCombinationsImpl = function(df,
 
   elems = rownames(df)
   df_sets = df[, all_set_names]
-  c_name = apply(df_sets, 1, function(r) paste(all_set_names[as.logical(r)], collapse=symbol))
-
+  c_name = apply(df_sets, 1, function(r) {
+    nn = all_set_names[as.logical(r)]
+    if (length(nn) == 1) {
+      nn
+    } else {
+      paste(nn, collapse=symbol)
+    }
+  })
   dd = aggregate(elems, list(c_name = c_name), function(r) {
     r
   })
   set_names = strsplit(dd$c_name, symbol, fixed = TRUE)
   set_colors = cc(dd$c_name)
 
   combinations = lapply(1:nrow(dd), function(i) {
-    elems = as.character(unlist(dd[i, 'x']))
+    elems = as.character(dd[i, 'x'][[1]])
     structure(
       list(
         name = dd[i, 'c_name'],
         color = set_colors[i],
         type = 'distinctIntersection',
-        elems = elems,
+        elems = if (store.elems) elems else c(),
         cardinality = length(elems),
-        setNames = unlist(set_names[i])
+        setNames = set_names[i][[1]]
       ),
       class = "upsetjs_combination"
     )

r_package/R/data.R

@@ -111,33 +111,29 @@ fromList = function(upsetjs,
   }
 
   sorted_sets = sortSets(sets, order.by = order.by, limit = limit)
-  gen_sets = if (is.null(limit))
-    sets
-  else
-    sorted_sets
 
   gen = if (!is.null(c_type) && c_type == "none") {
     list() 
   } else if (isVennDiagram(upsetjs) || isKarnaughMap(upsetjs)) {
     generateCombinationsImpl(
-      gen_sets,
+      sorted_sets,
       ifelse(is.null(c_type), 'distinctIntersection', c_type),
       0,
       NULL,
       TRUE,
       'degree',
-      NULL,
+      limit,
       colors
     )
   } else {
     generateCombinationsImpl(
-      gen_sets,
+      sorted_sets,
       ifelse(is.null(c_type), 'intersection', c_type),
       0,
       NULL,
       FALSE,
       order.by,
-      NULL,
+      limit,
       colors
     )
   }
@@ -326,14 +322,15 @@ fromDataFrame = function(upsetjs,
 
   cc = colorLookup(colors)
 
-  sorted_sets = extractSetsFromDataFrame(df, attributes, order.by, limit, colors, store.elems = store.elems)
+  gen_type = ifelse(!is.null(c_type), c_type, ifelse(isVennDiagram(upsetjs) || isKarnaughMap(upsetjs), 'distinctIntersection', 'intersection'))
+  sorted_sets = extractSetsFromDataFrame(df, attributes, order.by, limit, colors, store.elems = store.elems || gen_type != 'distinctIntersection')
 
   elems = rownames(df)
 
   gen = if (!is.null(c_type) && c_type == "none") {
     list() 
   } else if (isVennDiagram(upsetjs) || isKarnaughMap(upsetjs)) {
-    if (is.null(c_type) || c_type == 'distinctIntersection') {
+    if (gen_type == 'distinctIntersection') {
       extractCombinationsImpl(
         df,
         sorted_sets,
@@ -346,7 +343,7 @@ fromDataFrame = function(upsetjs,
     } else {
       generateCombinationsImpl(
         sorted_sets,
-        ifelse(is.null(c_type), 'distinctIntersection', c_type),
+        gen_type,
         0,
         NULL,
         TRUE,
@@ -356,7 +353,7 @@ fromDataFrame = function(upsetjs,
         store.elems = store.elems
       )
     }
-  } else if (!is.null(c_type) && c_type == 'distinctIntersection') {
+  } else if (gen_type == 'distinctIntersection') {
     extractCombinationsImpl(
       df,
       sorted_sets,
@@ -369,7 +366,7 @@ fromDataFrame = function(upsetjs,
   } else {
     generateCombinationsImpl(
       sorted_sets,
-      ifelse(is.null(c_type), 'intersection', c_type),
+      gen_type,
       0,
       NULL,
       FALSE,
@@ -379,6 +376,14 @@ fromDataFrame = function(upsetjs,
       store.elems = store.elems
     )
   }
+
+  if (!store.elems && gen_type != 'distinctIntersection') {
+    # delete
+    for(i in 1:length(sorted_sets)) {
+      sorted_sets[[i]]$elems = c()
+    }
+  }
+
   props = list(
     sets = sorted_sets,
     combinations = gen,

r_package/tests/testthat/helpers.R

@@ -17,7 +17,10 @@ upsetjs_mock = function() {
   r
 }
 
-expect_set = function(s, name, cardinality) {
+expect_set = function(s, name, cardinality, check.length=TRUE) {
   expect_equal(s$name, name)
   expect_equal(s$cardinality, cardinality)
+  if (check.length) {
+    expect_equal(length(s$elems), cardinality)
+  }
 }

r_package/tests/testthat/test-data.R

@@ -74,19 +74,19 @@ test_that("fromExpression", {
   u = upsetjs_mock() %>% fromExpression(expressionInput, order.by='degree')
   sets = u %>% getSets()
   expect_equal(length(sets), 3)
-  expect_set(sets[[1]], 'one', 9)
-  expect_set(sets[[2]], 'three', 9)
-  expect_set(sets[[3]], 'two', 5)
+  expect_set(sets[[1]], 'one', 9, check.length=FALSE)
+  expect_set(sets[[2]], 'three', 9, check.length=FALSE)
+  expect_set(sets[[3]], 'two', 5, check.length=FALSE)
 
   cc = u %>% getCombinations()
   expect_equal(length(cc), 7)
-  expect_set(cc[[1]], 'one', 9)
-  expect_set(cc[[2]], 'three', 9)
-  expect_set(cc[[3]], 'two', 5)
-  expect_set(cc[[4]], 'one&three', 6)
-  expect_set(cc[[5]], 'one&two', 3)
-  expect_set(cc[[6]], 'two&three', 3)
-  expect_set(cc[[7]], 'one&two&three', 2)
+  expect_set(cc[[1]], 'one', 9, check.length=FALSE)
+  expect_set(cc[[2]], 'three', 9, check.length=FALSE)
+  expect_set(cc[[3]], 'two', 5, check.length=FALSE)
+  expect_set(cc[[4]], 'one&three', 6, check.length=FALSE)
+  expect_set(cc[[5]], 'one&two', 3, check.length=FALSE)
+  expect_set(cc[[6]], 'two&three', 3, check.length=FALSE)
+  expect_set(cc[[7]], 'one&two&three', 2, check.length=FALSE)
 })
 
 test_that("fromDataFrame", {