[Lang] MatrixNdarray refactor part13: Add scalarization for TernaryOpStmt

[jim19930609]

Related issue = #5873, #5819

This PR is working "Part ④" in #5873.

Before scalarization:

TensorType<4 x i32> val = TernaryStmt(TensorType<4 x i32> cond,
                                           TensorType<4 x i32> lhs,
                                           TensorType<4 x i32> rhs)

After scalarization:

i32 val0 = TernaryStmt(cond->cast<MatrixInitStmt>()->val[0],
                       lhs->cast<MatrixInitStmt>()->val[0],
                       rhs->cast<MatrixInitStmt>()->val[0])

i32 val1 = TernaryStmt(cond->cast<MatrixInitStmt>()->val[1],
                       lhs->cast<MatrixInitStmt>()->val[1],
                       rhs->cast<MatrixInitStmt>()->val[1])

i32 val2 = TernaryStmt(cond->cast<MatrixInitStmt>()->val[2],
                       lhs->cast<MatrixInitStmt>()->val[2],
                       rhs->cast<MatrixInitStmt>()->val[2])

i32 val3 = TernaryStmt(cond->cast<MatrixInitStmt>()->val[3],
                       lhs->cast<MatrixInitStmt>()->val[3],
                       rhs->cast<MatrixInitStmt>()->val[3])

tmp = MatrixInitStmt(val0, val1, val2, val3)

stmt->replace_all_usages_with(tmp)

[netlify]

✅ Deploy Preview for docsite-preview canceled.

Name	Link
🔨 Latest commit	8080e7b
🔍 Latest deploy log	https://app.netlify.com/sites/docsite-preview/deploys/634cdf6405d4e10009901b9a