[mlir][sparse] Migrate tests to use new syntax (llvm#66543)

**COO**
`lvlTypes = [ "compressed_nu", "singleton" ]` to `map = (d0, d1) -> (d0
: compressed(nonunique), d1 : singleton)`
`lvlTypes = [ "compressed_nu_no", "singleton_no" ]` to `map = (d0, d1)
-> (d0 : compressed(nonunique, nonordered), d1 : singleton(nonordered))`

**SortedCOO**
`lvlTypes = [ "compressed_nu", "singleton" ]` to `map = (d0, d1) -> (d0
: compressed(nonunique), d1 : singleton)`

**BCOO**
`lvlTypes = [ "dense", "compressed_hi_nu", "singleton" ]` to `map = (d0,
d1, d2) -> (d0 : dense, d1 : compressed(nonunique, high), d2 :
singleton)`

**BCSR**
`lvlTypes = [ "compressed", "compressed", "dense", "dense" ], dimToLvl =
affine_map<(d0, d1) -> (d0 floordiv 2, d1 floordiv 3, d0 mod 2, d1 mod
3)>` to
`map = ( i, j ) ->
      ( i floordiv 2 : compressed,
        j floordiv 3 : compressed,
        i mod 2 : dense,
        j mod 3 : dense
      )`

**Tensor and other supported formats(e.g. CCC, CDC, CCCC)**

Currently, ELL and slice are not supported yet in the new syntax and the
CHECK tests will be updated once printing is set to output the new
syntax.

Previous PRs: llvm#66146, llvm#66309, llvm#66443

Author: yinying-lisa-li

mlir/include/mlir/Dialect/SparseTensor/IR/SparseTensorAttrDefs.td

@@ -200,7 +200,7 @@ def SparseTensorEncodingAttr : SparseTensor_Attr<"SparseTensorEncoding",
 
     // Sorted Coordinate Scheme.
     #SortedCOO = #sparse_tensor.encoding<{
-      lvlTypes = [ "compressed_nu", "singleton" ]
+      map = (d0, d1) -> (d0 : compressed(nonunique), d1 : singleton)
     }>
     ... tensor<?x?xf64, #SortedCOO> ...
 
@@ -214,8 +214,12 @@ def SparseTensorEncodingAttr : SparseTensor_Attr<"SparseTensorEncoding",
 
     // Block sparse row storage (2x3 blocks).
     #BCSR = #sparse_tensor.encoding<{
-      lvlTypes = [ "compressed", "compressed", "dense", "dense" ],
-      dimToLvl = affine_map<(i, j) -> (i floordiv 2, j floordiv 3, i mod 2, j mod 3)>
+      map = ( i, j ) ->
+      ( i floordiv 2 : compressed,
+        j floordiv 3 : compressed,
+        i mod 2      : dense,
+        j mod 3      : dense
+      )
     }>
     ... tensor<20x30xf32, #BCSR> ...
 

mlir/test/Dialect/SparseTensor/GPU/gpu_matvec_lib.mlir

@@ -2,7 +2,7 @@
 // RUN:             --sparsification="enable-gpu-libgen" | FileCheck %s
 
 #SortedCOO = #sparse_tensor.encoding<{
-  lvlTypes = [ "compressed_nu", "singleton" ]
+  map = (d0, d1) -> (d0 : compressed(nonunique), d1 : singleton)
 }>
 
 module {

mlir/test/Dialect/SparseTensor/codegen.mlir

@@ -27,7 +27,7 @@
 }>
 
 #UCSR = #sparse_tensor.encoding<{
-  lvlTypes = [ "dense", "compressed_no" ]
+  map = (d0, d1) -> (d0 : dense, d1 : compressed(nonordered))
 }>
 
 #CSC = #sparse_tensor.encoding<{
@@ -41,21 +41,19 @@
 }>
 
 #Dense3D = #sparse_tensor.encoding<{
-  lvlTypes = [ "dense", "dense", "dense" ],
-  dimToLvl = affine_map<(i, j, k) -> (k, i, j)>
+  map = (d0, d1, d2) -> (d2 : dense, d0 : dense, d1 : dense)
 }>
 
 #Coo = #sparse_tensor.encoding<{
-  lvlTypes = [ "compressed_nu", "singleton" ]
+  map = (d0, d1) -> (d0 : compressed(nonunique), d1 : singleton)
 }>
 
 #CooPNo = #sparse_tensor.encoding<{
-  lvlTypes = [ "compressed_nu", "singleton_no" ],
-  dimToLvl = affine_map<(i, j) -> (j, i)>
+  map = (d0, d1) -> (d1 : compressed(nonunique), d0 : singleton(nonordered))
 }>
 
 #ccoo = #sparse_tensor.encoding<{
-  lvlTypes = [ "compressed", "compressed_nu", "singleton" ]
+  map = (d0, d1, d2) -> (d0 : compressed, d1 : compressed(nonunique), d2 : singleton)
 }>
 
 // CHECK-LABEL: func @sparse_nop(

mlir/test/Dialect/SparseTensor/codegen_sparse_alloc.mlir

@@ -1,7 +1,7 @@
 // RUN: mlir-opt %s --sparse-tensor-codegen --canonicalize --cse | FileCheck %s
 
 #CSR = #sparse_tensor.encoding<{ map = (d0, d1) -> (d0 : dense, d1 : compressed)}>
-#COO = #sparse_tensor.encoding<{ lvlTypes = ["compressed_nu", "singleton"]}>
+#COO = #sparse_tensor.encoding<{ map = (d0, d1) -> (d0 : compressed(nonunique), d1 : singleton)}>
 
 // CHECK-LABEL:   func.func @sparse_alloc_copy_CSR(
 // CHECK-SAME:      %[[VAL_0:.*0]]: memref<?xindex>,

mlir/test/Dialect/SparseTensor/conversion.mlir

@@ -25,8 +25,7 @@
 }>
 
 #SparseTensor = #sparse_tensor.encoding<{
-  lvlTypes = ["dense", "compressed", "compressed"],
-  dimToLvl = affine_map<(i,j,k) -> (k,i,j)>
+  map = (d0, d1, d2) -> (d2 : dense, d0 : compressed, d1 : compressed)
 }>
 
 // CHECK-LABEL: func @sparse_nop(

mlir/test/Dialect/SparseTensor/convert_dense2sparse.mlir

@@ -15,8 +15,7 @@
 }>
 
 #SparseTensor = #sparse_tensor.encoding<{
-  lvlTypes = ["dense", "compressed", "compressed"],
-  dimToLvl = affine_map<(i,j,k) -> (k,i,j)>
+  map = (d0, d1, d2) -> (d2 : dense, d0 : compressed, d1 : compressed)
 }>
 
 // CHECK-LABEL: func @sparse_convert_1d(

mlir/test/Dialect/SparseTensor/convert_sparse2dense.mlir

@@ -12,8 +12,7 @@
 }>
 
 #SparseTensor = #sparse_tensor.encoding<{
-  lvlTypes = ["dense", "compressed", "compressed"],
-  dimToLvl = affine_map<(i,j,k) -> (k,i,j)>
+  map = (d0, d1, d2) -> (d2 : dense, d0 : compressed, d1 : compressed)
 }>
 
 // CHECK-LABEL: func @sparse_convert_1d(

mlir/test/Dialect/SparseTensor/convert_sparse2sparse.mlir

@@ -26,17 +26,16 @@
 }>
 
 #SortedCOO2D = #sparse_tensor.encoding<{
-  lvlTypes = [ "compressed_nu", "singleton" ],
+  map = (d0, d1) -> (d0 : compressed(nonunique), d1 : singleton),
 }>
 
 #SortedCOO3D = #sparse_tensor.encoding<{
-  lvlTypes = [ "compressed_nu", "singleton_nu", "singleton" ]
+  map = (d0, d1, d2) -> (d0 : compressed(nonunique), d1 : singleton(nonunique), d2 : singleton)
 
 }>
 
 #TsssPermuted = #sparse_tensor.encoding<{
-  lvlTypes = [ "compressed", "compressed", "compressed" ],
-  dimToLvl = affine_map<(i,j,k) -> (k,i,j)>
+  map = (d0, d1, d2) -> (d2 : compressed, d0 : compressed, d1 : compressed)
 }>
 
 #COOSlice = #sparse_tensor.encoding<{
@@ -115,13 +114,13 @@ func.func @sparse_convert(%arg0: tensor<?xf32, #SparseVector64>) -> tensor<?xf32
 }
 
 #SparseSingleton64 = #sparse_tensor.encoding<{
-  lvlTypes = ["singleton"],
+  map = (d0) -> (d0 : singleton),
   posWidth = 64,
   crdWidth = 64
 }>
 
 #SparseSingleton32 = #sparse_tensor.encoding<{
-  lvlTypes = ["singleton"],
+  map = (d0) -> (d0 : singleton),
   posWidth = 32,
   crdWidth = 32
 }>

mlir/test/Dialect/SparseTensor/invalid.mlir

@@ -32,7 +32,7 @@ func.func @invalid_pack_type(%values: tensor<6xf64>, %pos: tensor<2xi32>, %coord
 
 // -----
 
-#SparseVector = #sparse_tensor.encoding<{lvlTypes = ["compressed_nu", "singleton"], posWidth=32, crdWidth=32}>
+#SparseVector = #sparse_tensor.encoding<{map = (d0, d1) -> (d0 : compressed(nonunique), d1 : singleton), posWidth=32, crdWidth=32}>
 
 func.func @invalid_pack_type(%values: tensor<6xf64>, %pos: tensor<2xi32>, %coordinates: tensor<6x3xi32>)
                             -> tensor<100x2xf64, #SparseVector> {
@@ -68,7 +68,7 @@ func.func @invalid_unpack_type(%sp: tensor<100xf32, #SparseVector>, %values: ten
 
 // -----
 
-#SparseVector = #sparse_tensor.encoding<{lvlTypes = ["compressed_nu", "singleton"], posWidth=32, crdWidth=32}>
+#SparseVector = #sparse_tensor.encoding<{map = (d0, d1) -> (d0 : compressed(nonunique), d1 : singleton), posWidth=32, crdWidth=32}>
 
 func.func @invalid_unpack_type(%sp: tensor<100x2xf64, #SparseVector>, %values: tensor<6xf64>, %pos: tensor<2xi32>, %coordinates: tensor<6x3xi32>) {
   // expected-error@+1 {{input/output trailing COO level-ranks don't match}}
@@ -270,7 +270,7 @@ func.func @sparse_get_md(%arg0: !sparse_tensor.storage_specifier<#SparseVector>)
 
 // -----
 
-#COO = #sparse_tensor.encoding<{lvlTypes = ["compressed_nu", "singleton"]}>
+#COO = #sparse_tensor.encoding<{map = (d0, d1) -> (d0 : compressed(nonunique), d1 : singleton)}>
 
 func.func @sparse_get_md(%arg0: !sparse_tensor.storage_specifier<#COO>) -> index {
   // expected-error@+1 {{requested position memory size on a singleton level}}
@@ -658,7 +658,7 @@ func.func @invalid_concat_dim(%arg0: tensor<2x4xf64, #DC>,
 
 #C = #sparse_tensor.encoding<{map = (d0) -> (d0 : compressed)}>
 #DC = #sparse_tensor.encoding<{map = (d0, d1) -> (d0 : dense, d1 : compressed)}>
-#DCC = #sparse_tensor.encoding<{lvlTypes = ["dense", "compressed", "compressed"]}>
+#DCC = #sparse_tensor.encoding<{map = (d0, d1, d2) -> (d0 : dense, d1 : compressed, d2 : compressed)}>
 func.func @invalid_concat_rank_mismatch(%arg0: tensor<2xf64, #C>,
                                         %arg1: tensor<3x4xf64, #DC>,
                                         %arg2: tensor<4x4x4xf64, #DCC>) -> tensor<9x4xf64, #DC> {

mlir/test/Dialect/SparseTensor/pre_rewriting.mlir

@@ -5,7 +5,7 @@
 }>
 
 #SortedCOO = #sparse_tensor.encoding<{
-  lvlTypes = [ "compressed_nu", "singleton" ]
+  map = (d0, d1) -> (d0 : compressed(nonunique), d1 : singleton)
 }>
 
 #DCSR = #sparse_tensor.encoding<{