In SMARTY remove any initial types that match zero atoms
Closed this issue · 3 comments
bannanc commented
Example output after 9 iterations
Iteration 9 / 10
Attempting to create new subtype: '[#17]' (c_chlorine) + 'X1' (connections-1) -> '[#17X1]' (c_chlorine connections-1)
Atom type '[#17X1]' (c_chlorine connections-1) unused in dataset; rejecting.
Rejected.
INDEX ATOMS MOLECULES TYPE NAME SMARTS REF TYPE FRACTION OF REF TYPED MOLECULES MATCHED
1 : 464 42 | c_hydrogen [#1] HC 244 / 244 (100.000%)
2 : 232 42 | c_carbon [#6] CT 232 / 232 (100.000%)
3 : 0 0 | c_nitrogen [#7]
4 : 107 42 | c_oxygen [#8] OH 68 / 68 (100.000%)
5 : 0 0 | c_phosphorous [#15]
6 : 0 0 | c_sulfur [#16]
7 : 0 0 | c_chlorine [#17]
8 : 0 0 | c_iodine [#53]
TOTAL : 803 42 | 544 / 803 match (67.746 %)
Atom type hierarchy:
[#1]
[#16]
[#8]
[#53]
[#15]
[#6]
[#7]
[#17]
We shouldn't be wasting time creating atomtypes that don't exist in the molecule set.
I will address this will all the other things I'm fixing in pull request #198
davidlmobley commented
I had thought I'd fixed this long ago, but it seems like (from looking at the issue tracker) what I really fixed was to allow unused base types to get destroyed. So, these should eventually go away (and then no longer be suggested) but certainly you can save some effort by just getting rid of them initially, @bannanc .
bannanc commented
@davidlmobley I'm updating the smarty sampler as we speak, I already added a few lines to remove unused initial types.