Apparently, the current method relies on some empirical rules based on the column name and type. It would be great to detect molecules automatically, no matter what the column name is.

Apparently, the current method relies on some empirical rules based on the column name and type. It would be great to detect molecules automatically, no matter what the column name is.
