search

xiaoruiDong / RDMC

Reaction Data and Molecular Conformers (RDMC) is a package dealing with reactions, molecules, conformers, majorly in 3D.
https://xiaoruidong.github.io/RDMC/
MIT License
22 stars 1 forks source link

Improve the TorsionalSampler class #41

Closed shihchengli closed 1 year ago

shihchengli commented 1 year ago
  1. If n_dimension is assigned -1, the n_dimension would be the number of rotatable bonds for a given conformer.
  2. The set of rotatable bonds can be manually assigned when TorsionalSampler is called.
  3. Whether the conformer is TS is determined by the argument of is_ts in __call__() function of TorsionalSampler object. Previously, it is automatically determined by whether the rxn_smiles is provided or not, which is not good.
  4. Whether the dangling bonds in TS are rotatable or not can be controlled by using the attribute of no_sample_dangling_bonds in __call__() in TorsionalSampler.
  5. Using no_greedy as True during torsional sampling, all the sampled points would be passed to the following steps (opt, freq, IRC...) without using the greedy algorithm to find some local minima of them.
  6. Save the sampled geometries in a .sdf file.
xiaoruiDong commented 1 year ago

@shihchengli Thanks! I went through the changes, and they look good to me. Since the "no greedy" approach doesn't involve any QM software, it would be great if you could write some unit tests regarding the functionality?

shihchengli commented 1 year ago

@xiaoruiDong Thanks for reviewing this! I just added a unittest file. Feel free to let me know if other tests should also be included.

shihchengli commented 1 year ago

I removed point 3 since the purpose for determining whether this conformer is TS or not is due to the input of the optimizer for stable species and TS being different. (which should be fixed in the other PR)