ivy: matrix divide, mdiv

Dead easy now we have matrix invert.

Moore-Penrose inverse will enable more uses; see issue #146

doc.go

@@ -158,7 +158,7 @@ Binary operators
 	                                      In ivy: abs(A) gives count, A <= 0 inserts zero
 	Index of              A⍳B   iota      The location (index) of B in A; 1+⌈/⍳⍴A if not found
 	                                      In ivy: origin-1 if not found (i.e. 0 if one-indexed)
-	Matrix divide         A⌹B             Solution to system of linear equations Ax = B
+	Matrix divide         A⌹B   mdiv      Solution to system of linear equations Ax = B
 	Rotation              A⌽B   rot       The elements of B are rotated A positions left
 	Rotation              A⊖B   flip      The elements of B are rotated A positions along the first axis
 	Logarithm             A⍟B   log       Logarithm of B to base A

testdata/binary_matrix.ivy

@@ -132,6 +132,13 @@
 	1 2 3
 	4 5 6
 
+(2 2 rho 1 2 3 4) mdiv (2 2 rho 5 6 7 8)
+	 5  4
+	-4 -3
+
+(5 8) mdiv (2 2 rho 1 2 2 -1)
+	21/5 2/5
+
 (2 3 rho iota 10) ** 5
 	   1   32  243
 	1024 3125 7776

value/binary.go

@@ -355,6 +355,18 @@ func init() {
 			},
 		},
 
+		{ // Matrix division
+			name:        "mdiv",
+			elementwise: false,
+			whichType:   binaryArithType,
+			fn: [numType]binaryFn{
+				// TODO: Other types?
+				matrixType: func(c Context, u, v Value) Value {
+					return innerProduct(c, v.(*Matrix).inverse(c), "+", "*", u)
+				},
+			},
+		},
+
 		{ // Euclidean integer division.
 			name:        "div",
 			elementwise: true,