CLR Normalization does not automatically handle zero values

[qqwxp]

After performing normalize(data.obj, method = "CLR") on the readcount data, I found that some taxa with small readcount values were transformed into negative numbers (which is normal), but the zero values remain. It is a completely incorrect data.

The function should automatically handle zero values (e.g., by adding 1 to all zeros), or provide some method parameters for handling zero values.