[mlir][sparse] Calling `SparseTensorType::get{Pos,Crd}Type` directly

[wrengr]

Summary

The SparseTensorType versions of these methods have some special handling to ensure that they work for unannotated tensors; whereas, the stt.getEncoding().get{Pos,Crd}Type() idiom can cause segfaults.

mlir/lib/Dialect/SparseTensor/Transforms/CodegenUtils.cpp

Value sparse_tensor::reshapeValuesToLevels(OpBuilder &builder, Location loc,
   const Type elemTp = getMemRefType(valuesBuffer).getElementType();
   const auto resTp = MemRefType::get(resShape, elemTp);
   return builder.create<memref::ReshapeOp>(loc, resTp, valuesBuffer, lvlCoords);
 }
 
 Value sparse_tensor::genToPositions(OpBuilder &builder, Location loc,
                                     Value tensor, Level lvl) {
   const auto srcTp = getSparseTensorType(tensor);
-  const Type posTp = srcTp.getEncoding().getPosType();
+  const Type posTp = srcTp.getPosType();
   const Type memTp = get1DMemRefType(posTp, /*withLayout=*/false);
   return builder.create<ToPositionsOp>(loc, memTp, tensor,
                                        builder.getIndexAttr(lvl));
 }
 
 Value sparse_tensor::genToCoordinates(OpBuilder &builder, Location loc,
                                       Value tensor, Level lvl, Level cooStart) {
   const auto srcTp = getSparseTensorType(tensor);
-  const Type crdTp = srcTp.getEncoding().getCrdType();
+  const Type crdTp = srcTp.getCrdType();
   const Type memTp = get1DMemRefType(crdTp, /*withLayout=*/lvl >= cooStart);
   return builder.create<ToCoordinatesOp>(loc, memTp, tensor,
                                          builder.getIndexAttr(lvl));
 }
 
 Value sparse_tensor::genToCoordinatesBuffer(OpBuilder &builder, Location loc,
                                             Value tensor) {
   const auto srcTp = getSparseTensorType(tensor);
-  const Type crdTp = srcTp.getEncoding().getCrdType();
+  const Type crdTp = srcTp.getCrdType();
   const Type memTp = get1DMemRefType(crdTp, /*withLayout=*/false);
   return builder.create<ToCoordinatesBufferOp>(loc, memTp, tensor);
 }
 
 Value sparse_tensor::genToValues(OpBuilder &builder, Location loc,
                                  Value tensor) {
   RankedTensorType srcTp = getRankedTensorType(tensor);
   Type valTp = get1DMemRefType(srcTp.getElementType(),