How is the result_start_mols_task1_seed_0.csv generated?

[yukichenyunchi]

I followed the repo's instructions while removing all the csv under MCTS/libs/gsk3b_jnk3_stage1/ to run the entire code from scratch. However, the results from stage 1 cannot be passed to stage 2, since there is no such csv generated from stage 1 (which serves as start_mols_fn in stage 2).

I found that the txt generated in frag_mcts_mo_stage1.py (gsk3b_jnk3_maxchild_5_sim_20_scalar_0.7_idx_0_seed_0.txt) might be the one that is similar to the aforementioned required csv. However, there are just several (~10) molecules in the txt, which is far away from the molecules in result_start_mols_task1_seed_0.csv (~2000).

How is the result_start_mols_task1_seed_0.csv generated? Or how to increase the number of molecules in stage 1?

Thanks!