faster CountTable sort(), optional SortOrder (#11010)

* use existing sort for CountTable, and
add SortOrder options to CountTable, OrderedTable sort(s)

* add some tests, runnables, etc.

* fix runnable imports

lib/pure/collections/tables.nim

@@ -216,7 +216,7 @@
 ## * `hashes module<hashes.html>`_ for helper functions for hashing
 
 
-import hashes, math
+import hashes, math, algorithm
 
 include "system/inclrtl"
 
@@ -1467,7 +1467,7 @@ proc clear*[A, B](t: var OrderedTable[A, B]) =
   t.first = -1
   t.last = -1
 
-proc sort*[A, B](t: var OrderedTable[A, B], cmp: proc (x,y: (A, B)): int) =
+proc sort*[A, B](t: var OrderedTable[A, B], cmp: proc (x,y: (A, B)): int, order = SortOrder.Ascending) =
   ## Sorts ``t`` according to the function ``cmp``.
   ##
   ## This modifies the internal list
@@ -1475,12 +1475,15 @@ proc sort*[A, B](t: var OrderedTable[A, B], cmp: proc (x,y: (A, B)): int) =
   ## call but key lookup and insertions remain possible after ``sort`` (in
   ## contrast to the `sort proc<#sort,CountTable[A]>`_ for count tables).
   runnableExamples:
+    import algorithm
     var a = initOrderedTable[char, int]()
     for i, c in "cab":
       a[c] = 10*i
     doAssert a == {'c': 0, 'a': 10, 'b': 20}.toOrderedTable
     a.sort(system.cmp)
     doAssert a == {'a': 10, 'b': 20, 'c': 0}.toOrderedTable
+    a.sort(system.cmp, order=SortOrder.Descending)
+    doAssert a == {'c': 0, 'b': 20, 'a': 10}.toOrderedTable
 
   var list = t.first
   var
@@ -1508,7 +1511,7 @@ proc sort*[A, B](t: var OrderedTable[A, B], cmp: proc (x,y: (A, B)): int) =
         elif qsize == 0 or q < 0:
           e = p; p = t.data[p].next; dec(psize)
         elif cmp((t.data[p].key, t.data[p].val),
-                 (t.data[q].key, t.data[q].val)) <= 0:
+                 (t.data[q].key, t.data[q].val)) * order <= 0:
           e = p; p = t.data[p].next; dec(psize)
         else:
           e = q; q = t.data[q].next; dec(qsize)
@@ -1875,7 +1878,7 @@ proc clear*[A, B](t: var OrderedTableRef[A, B]) =
     doAssert len(a) == 0
   clear(t[])
 
-proc sort*[A, B](t: OrderedTableRef[A, B], cmp: proc (x,y: (A, B)): int) =
+proc sort*[A, B](t: OrderedTableRef[A, B], cmp: proc (x,y: (A, B)): int, order = SortOrder.Ascending) =
   ## Sorts ``t`` according to the function ``cmp``.
   ##
   ## This modifies the internal list
@@ -1883,13 +1886,16 @@ proc sort*[A, B](t: OrderedTableRef[A, B], cmp: proc (x,y: (A, B)): int) =
   ## call but key lookup and insertions remain possible after ``sort`` (in
   ## contrast to the `sort proc<#sort,CountTableRef[A]>`_ for count tables).
   runnableExamples:
+    import algorithm
     var a = newOrderedTable[char, int]()
     for i, c in "cab":
       a[c] = 10*i
     doAssert a == {'c': 0, 'a': 10, 'b': 20}.newOrderedTable
     a.sort(system.cmp)
     doAssert a == {'a': 10, 'b': 20, 'c': 0}.newOrderedTable
-  t[].sort(cmp)
+    a.sort(system.cmp, order=SortOrder.Descending)
+    doAssert a == {'c': 0, 'b': 20, 'a': 10}.newOrderedTable
+  t[].sort(cmp, order=order)
 
 proc `$`*[A, B](t: OrderedTableRef[A, B]): string =
   ## The ``$`` operator for hash tables. Used internally when calling `echo`
@@ -2199,30 +2205,27 @@ proc clear*[A](t: var CountTable[A]) =
   ## Resets the table so that it is empty.
   clearImpl()
 
-proc sort*[A](t: var CountTable[A]) =
-  ## Sorts the count table so that the entry with the highest counter comes
-  ## first.
+func ctCmp[T](a, b: tuple[key: T, val: int]): int =
+  result = system.cmp(a.val, b.val)
+
+proc sort*[A](t: var CountTable[A], order = SortOrder.Descending) =
+  ## Sorts the count table so that, by default, the entry with the
+  ## highest counter comes first.
   ##
   ## **This is destructive! You must not modify ``t`` afterwards!**
   ##
   ## You can use the iterators `pairs<#pairs.i,CountTable[A]>`_,
   ## `keys<#keys.i,CountTable[A]>`_, and `values<#values.i,CountTable[A]>`_
   ## to iterate over ``t`` in the sorted order.
-
-  # we use shellsort here; fast enough and simple
-  var h = 1
-  while true:
-    h = 3 * h + 1
-    if h >= high(t.data): break
-  while true:
-    h = h div 3
-    for i in countup(h, high(t.data)):
-      var j = i
-      while t.data[j-h].val <= t.data[j].val:
-        swap(t.data[j], t.data[j-h])
-        j = j-h
-        if j < h: break
-    if h == 1: break
+  runnableExamples:
+    import algorithm, sequtils
+    var a = toCountTable("abracadabra")
+    doAssert a == "aaaaabbrrcd".toCountTable
+    a.sort()
+    doAssert toSeq(a.values) == @[5, 2, 2, 1, 1]
+    a.sort(SortOrder.Ascending)
+    doAssert toSeq(a.values) == @[1, 1, 2, 2, 5]
+  t.data.sort(cmp=ctCmp, order=order)
 
 proc merge*[A](s: var CountTable[A], t: CountTable[A]) =
   ## Merges the second table into the first one (must be declared as `var`).
@@ -2467,16 +2470,16 @@ proc clear*[A](t: CountTableRef[A]) =
   ## Resets the table so that it is empty.
   clearImpl()
 
-proc sort*[A](t: CountTableRef[A]) =
-  ## Sorts the count table so that the entry with the highest counter comes
-  ## first.
+proc sort*[A](t: CountTableRef[A], order = SortOrder.Descending) =
+  ## Sorts the count table so that, by default, the entry with the
+  ## highest counter comes first.
   ##
   ## **This is destructive! You must not modify `t` afterwards!**
   ##
   ## You can use the iterators `pairs<#pairs.i,CountTableRef[A]>`_,
   ## `keys<#keys.i,CountTableRef[A]>`_, and `values<#values.i,CountTableRef[A]>`_
   ## to iterate over ``t`` in the sorted order.
-  t[].sort
+  t[].sort(order=order)
 
 proc merge*[A](s, t: CountTableRef[A]) =
   ## Merges the second table into the first one.

tests/collections/ttables.nim

@@ -8,7 +8,7 @@ And we get here
 '''
 joinable: false
 """
-import hashes, sequtils, tables
+import hashes, sequtils, tables, algorithm
 
 # test should not be joined because it takes too long.
 block tableadds:
@@ -274,22 +274,30 @@ block tablesref:
   block countTableTest1:
     var s = data.toTable
     var t = newCountTable[string]()
-    for k in s.keys: t.inc(k)
-    for k in t.keys: assert t[k] == 1
-    t.inc("90", 3)
-    t.inc("12", 2)
-    t.inc("34", 1)
+    var r = newCountTable[string]()
+    for x in [t, r]:
+      for k in s.keys:
+        x.inc(k)
+        assert x[k] == 1
+      x.inc("90", 3)
+      x.inc("12", 2)
+      x.inc("34", 1)
     assert t.largest()[0] == "90"
 
     t.sort()
-    var i = 0
-    for k, v in t.pairs:
-      case i
-      of 0: assert k == "90" and v == 4
-      of 1: assert k == "12" and v == 3
-      of 2: assert k == "34" and v == 2
-      else: break
-      inc i
+    r.sort(SortOrder.Ascending)
+    var ps1 = toSeq t.pairs
+    var ps2 = toSeq r.pairs
+    ps2.reverse()
+    for ps in [ps1, ps2]:
+      var i = 0
+      for (k, v) in ps:
+        case i
+        of 0: assert k == "90" and v == 4
+        of 1: assert k == "12" and v == 3
+        of 2: assert k == "34" and v == 2
+        else: break
+        inc i
 
   block SyntaxTest:
     var x = newTable[int, string]({:})
@@ -308,13 +316,20 @@ block tablesref:
     var t = newOrderedTable[string, int](2)
     for key, val in items(data): t[key] = val
     for key, val in items(data): assert t[key] == val
-    t.sort(proc (x, y: tuple[key: string, val: int]): int = cmp(x.key, y.key))
+    proc cmper(x, y: tuple[key: string, val: int]): int = cmp(x.key, y.key)
+    t.sort(cmper)
     var i = 0
     # `pairs` needs to yield in sorted order:
     for key, val in pairs(t):
       doAssert key == sorteddata[i][0]
       doAssert val == sorteddata[i][1]
       inc(i)
+    t.sort(cmper, order=SortOrder.Descending)
+    i = 0
+    for key, val in pairs(t):
+      doAssert key == sorteddata[high(data)-i][0]
+      doAssert val == sorteddata[high(data)-i][1]
+      inc(i)
 
     # check that lookup still works:
     for key, val in pairs(t):