Add support for running some key bitops functions on integer Tensors

While we already supported most of nim's std/math features in Arraymancer, we did not support any of the std/bitops operators and procedures yet. These are very useful to implement some important algorithms such as gray coding and others.

This commit adds some of the most important std/bitops features. These will soon be used in `impulse` to implement some new algorithms.

src/arraymancer/tensor.nim

@@ -39,6 +39,7 @@ import  ./laser/dynamic_stack_arrays,
         ./tensor/math_functions,
         ./tensor/aggregate,
         ./tensor/algorithms,
+        ./tensor/bitops_functions,
         ./tensor/lapack,
         ./tensor/optim_ops_fusion,
         ./tensor/syntactic_sugar,
@@ -67,6 +68,7 @@ export  dynamic_stack_arrays,
         math_functions,
         aggregate,
         algorithms,
+        bitops_functions,
         lapack,
         optim_ops_fusion,
         syntactic_sugar,

src/arraymancer/tensor/bitops_functions.nim

@@ -0,0 +1,131 @@
+# Copyright 2017 the Arraymancer contributors
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import  ./data_structure,
+        ./higher_order_applymap,
+        ./shapeshifting,
+        ./ufunc
+import std / bitops
+
+export bitops
+
+proc `shr`*[T1, T2: SomeInteger](t: Tensor[T1], value: T2): Tensor[T1] {.noinit.} =
+  ## Broadcasted tensor-value `shr` (i.e. shift right) operator
+  ##
+  ## This is similar to numpy's `right_shift` and Matlab's `bitsra`
+  ## (or `bitshift` with a positive shift value).
+  t.map_inline(x shr value)
+
+proc `shr`*[T1, T2: SomeInteger](value: T1, t: Tensor[T2]): Tensor[T2] {.noinit.} =
+  ## Broadcasted value-tensor `shr` (i.e. shift right) operator
+  ##
+  ## This is similar to numpy's `right_shift` and Matlab's `bitsra`
+  ## (or `bitshift` with a positive shift value).
+  t.map_inline(value shr x)
+
+proc `shr`*[T: SomeInteger](t1, t2: Tensor[T]): Tensor[T] {.noinit.} =
+  ## Tensor element-wise `shr` (i.e. shift right) broadcasted operator
+  ##
+  ## This is similar to numpy's `right_shift` and Matlab's `bitsra`
+  ## (or `bitshift` with a positive shift value).
+  let (tmp1, tmp2) = broadcast2(t1, t2)
+  result = map2_inline(tmp1, tmp2, x shr y)
+
+proc `shl`*[T1, T2: SomeInteger](t: Tensor[T1], value: T2): Tensor[T1] {.noinit.} =
+  ## Broadcasted tensor-value `shl` (i.e. shift left) operator
+  ##
+  ## This is similar to numpy's `left_shift` and Matlab's `bitsla`
+  ## (or `bitshift` with a negative shift value).
+  t.map_inline(x shl value)
+
+proc `shl`*[T1, T2: SomeInteger](value: T1, t: Tensor[T2]): Tensor[T2] {.noinit.} =
+  ## Broadcasted value-tensor `shl` (i.e. shift left) operator
+  ##
+  ## This is similar to numpy's `left_shift` and Matlab's `bitsla`
+  ## (or `bitshift` with a negative shift value).
+  t.map_inline(value shl x)
+
+proc `shl`*[T: SomeInteger](t1, t2: Tensor[T]): Tensor[T] {.noinit.} =
+  ## Tensor element-wise `shl` (i.e. shift left) broadcasted operator
+  ##
+  ## This is similar to numpy's `left_shift` and Matlab's `bitsla`
+  ## (or `bitshift` with a negative shift value).
+  let (tmp1, tmp2) = broadcast2(t1, t2)
+  result = map2_inline(tmp1, tmp2, x shl y)
+
+makeUniversal(bitnot,
+  docSuffix="""Element-wise `bitnot` procedure
+
+This is similar to numpy's `bitwise_not` and Matlab's `bitnot`.""")
+
+proc bitand*[T](t: Tensor[T], value: T): Tensor[T] {.noinit.} =
+  ## Broadcasted tensor-value `bitand` procedure
+  ##
+  ## This is similar to numpy's `bitwise_and` and Matlab's `bitand`.
+  t.map_inline(bitand(x, value))
+
+proc bitand*[T](value: T, t: Tensor[T]): Tensor[T] {.noinit.} =
+  ## Broadcasted value-tensor `bitand` procedure
+  ##
+  ## This is similar to numpy's `bitwise_and` and Matlab's `bitand`.
+  t.map_inline(bitand(value, x))
+
+proc bitand*[T](t1, t2: Tensor[T]): Tensor[T] {.noinit.} =
+  ## Tensor element-wise `bitand` procedure
+  ##
+  ## This is similar to numpy's `bitwise_and` and Matlab's `bitand`.
+  let (tmp1, tmp2) = broadcast2(t1, t2)
+  result = map2_inline(tmp1, tmp2, bitand(x, y))
+
+
+proc bitor*[T](t: Tensor[T], value: T): Tensor[T] {.noinit.} =
+  ## Broadcasted tensor-value `bitor` procedure
+  ##
+  ## This is similar to numpy's `bitwise_or` and Matlab's `bitor`.
+  t.map_inline(bitor(x, value))
+
+proc bitor*[T](value: T, t: Tensor[T]): Tensor[T] {.noinit.} =
+  ## Broadcasted value-tensor `bitor` procedure
+  ##
+  ## This is similar to numpy's `bitwise_or` and Matlab's `bitor`.
+  t.map_inline(bitor(value, x))
+
+proc bitor*[T](t1, t2: Tensor[T]): Tensor[T] {.noinit.} =
+  ## Tensor element-wise `bitor` procedure
+  ##
+  ## This is similar to numpy's `bitwise_or` and Matlab's `bitor`.
+  let (tmp1, tmp2) = broadcast2(t1, t2)
+  map2_inline(tmp1, tmp2, bitor(x, y))
+
+proc bitxor*[T](t: Tensor[T], value: T): Tensor[T] {.noinit.} =
+  ## Broadcasted tensor-value `bitxor` procedure
+  ##
+  ## This is similar to numpy's `bitwise_xor` and Matlab's `bitxor`.
+  t.map_inline(bitxor(x, value))
+
+proc bitxor*[T](value: T, t: Tensor[T]): Tensor[T] {.noinit.} =
+  ## Broadcasted value-tensor `bitxor` procedure
+  ##
+  ## This is similar to numpy's `bitwise_xor` and Matlab's `bitxor`.
+  t.map_inline(bitxor(value, x))
+
+proc bitxor*[T](t1, t2: Tensor[T]): Tensor[T] {.noinit.} =
+  ## Tensor element-wise `bitxor` procedure
+  ##
+  ## This is similar to numpy's `bitwise_xor` and Matlab's `bitxor`.
+  let (tmp1, tmp2) = broadcast2(t1, t2)
+  map2_inline(tmp1, tmp2, bitxor(x, y))
+
+makeUniversal(reverseBits,
+  docSuffix="Element-wise `reverseBits` procedure")

tests/tensor/test_bitops_functions.nim

@@ -0,0 +1,69 @@
+# Copyright 2017 the Arraymancer contributors
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import ../../src/arraymancer
+import std / unittest
+
+proc main() =
+  suite "Bitops functions":
+    test "bitnot":
+      let t = [0, 1, 57, 1022, -100].toTensor
+      let expected = [-1, -2, -58, -1023, 99].toTensor
+      check: t.bitnot == expected
+
+    test "shr":
+      let t1 = [0, 1, 57, 1022, -100].toTensor
+      let t2 = [0, 1, 2, 3, 4].toTensor
+      check: t1 shr 3 == [0, 0, 7, 127, -13].toTensor
+      check: 1024 shr t2 == [1024, 512, 256, 128, 64].toTensor
+      check: t1 shr t2 == [0, 0, 14, 127, -7].toTensor
+
+    test "shl":
+      let t1 = [0, 1, 57, 1022, -100].toTensor
+      let t2 = [0, 1, 2, 3, 4].toTensor
+      check: t1 shl 3 == [0, 8, 456, 8176, -800].toTensor
+      check: 3 shl t2 == [3, 6, 12, 24, 48].toTensor
+      check: t1 shl t2 == [0, 2, 228, 8176, -1600].toTensor
+
+    test "bitand":
+      let t1 = [0, 1, 57, 1022, -100].toTensor
+      let t2 = [0, 2, 7, 15, 11].toTensor
+      check: bitand(t1, 0b010_110_101) == [0, 1, 49, 180, 148].toTensor
+      check: bitand(t1, 0b010_110_101) == bitand(0b010_110_101, t1)
+      check: bitand(t1, t2) == [0, 0, 1, 14, 8].toTensor
+      check: bitand(t1, t2) == bitand(t1, t2)
+
+    test "bitor":
+      let t1 = [0, 1, 57, 1022, -100].toTensor
+      let t2 = [0, 2, 7, 15, 11].toTensor
+      check: bitor(t1, 0b010_110_101) == [181, 181, 189, 1023, -67].toTensor
+      check: bitor(t1, 0b010_110_101) == bitor(0b010_110_101, t1)
+      check: bitor(t1, t2) == [0, 3, 63, 1023, -97].toTensor
+      check: bitor(t1, t2) == bitor(t1, t2)
+
+    test "bitxor":
+      let t1 = [0, 1, 57, 1022, -100].toTensor
+      let t2 = [0, 2, 7, 15, 11].toTensor
+      check: bitxor(t1, 0b010_110_101) == [181, 180, 140, 843, -215].toTensor
+      check: bitxor(t1, 0b010_110_101) == bitxor(0b010_110_101, t1)
+      check: bitxor(t1, t2) == [0, 3, 62, 1009, -105].toTensor
+      check: bitxor(t1, t2) == bitxor(t1, t2)
+
+    test "reverse_bits":
+      let t = [0, 1, 57, 1022].toTensor(uint16)
+      let expected = [0, 32768, 39936, 32704].toTensor(uint16)
+      check: t.reverse_bits == expected
+
+main()
+GC_fullCollect()