[IR] Add casting for min/max/put

[chhzh123]

Description

This PR adds casting operations for min/max/put. Specifically, the codegen for constant float values of HLS also changed, since 0.0 will be interpreted as double if not explicitly declaring the type.

Examples

def test_minmax_cast():
    def kernel(A: int8[2]) -> int32[2]:
        res: int32[2] = 0
        res[0] = min(A[0], 0)
        res[1] = max(A[1], 0.0)
        return res

    s = allo.customize(kernel)
    print(s.module)
    mod = s.build()
    np_A = np.random.randint(-64, 64, size=(2,)).astype(np.int8)
    allo_B = mod(np_A)
    assert allo_B[0] == min(np_A[0], 0)
    assert allo_B[1] == max(np_A[1], 0.0)
    mod = s.build(target="vhls")
    print(mod)
    assert "min" in mod.hls_code
    assert "max" in mod.hls_code
    assert "(float)" in mod.hls_code

void kernel(
  int8_t v0[2],
  int32_t v1[2]
) {     // L2
  for (int v2 = 0; v2 < 2; v2++) {      // L5
    v1[v2] = 0; // L5
  }
  int8_t v3 = v0[0];    // L6
  int32_t v4 = v3;      // L7
  int32_t v5 = min(v4, 0);      // L9
  v1[0] = v5;   // L10
  int8_t v6 = v0[1];    // L11
  float v7 = v6;        // L12
  float v8 = max(v7, (float)0.000000);  // L14
  int32_t v9 = v8;      // L15
  v1[1] = v9;   // L16
}

Checklist

• PR's title starts with a category (e.g. [Bugfix], [IR], [Builder], etc)
• Changes are complete (i.e. I finished coding on this PR)
• All changes have test coverage (It would be better to provide ~2 different test cases to test the robustness of your code)
• Code is well-documented