Interface to empirical optimizers for reaction condition optimization.
docker compose up- OLYMPUS: https://github.com/aspuru-guzik-group/olympus
- Summit: https://github.com/sustainable-processes/summit
Input
{
"identifier": "example_olympus",
"continuous_parameter_space": {
"temperature": [75, 90],
"pd_mol": [0.5, 5],
"arbpin": [1, 1.8],
"k3po4": [1.5, 3]
},
"categorical_parameter_space": {},
"target_names": [
"yield"
],
"observations": [
{"temperature": 75, "pd_mol": 0.5, "arbpin": 1, "k3po4": 1.5, "yield": 2.4},
{"temperature": 75, "pd_mol": 0.5, "arbpin": 1.2, "k3po4": 1.5, "yield": 4.6}
],
"optimize_goal": "maximize",
"planner_name": "ConjugateGradient"
}Output
{
"arbpin": 1.6015094769562883, "k3po4": 2.9960569094796194,
"pd_mol": 4.371576658109785, "temperature": 85.93353741155443
}Access to the following reaction datasets is provided.
| DOI | Description | Parameters | Target | # of reactions |
|---|---|---|---|---|
| 10.1126/science.aap9112 | Suzuki-Miyaura | Categorical + Continuous | Yield | 5280 |
| 10.1126/science.aar5169 | Buchwald-Hartwig | Categorical | Yield | 4132 |
| 10.1021/jacs.8b01523 | Deoxyfluorination | Categorical | Yield | 80 |
| 10.1021/acscatal.0c02247 | Dehalogenation | Categorical | Yield | 1152 |
| 10.1038/s41586-021-03213-y | Arylation | Categorical + Continuous | Yield | 1728 |
After downloading them to the correct folder using download.py,
all datasets can be accessed as BenchmarkDataframe objects:
from reaction_datasets import BenchmarkDataframe
dehalogenation_dataset = BenchmarkDataframe.ds_3()An example using Summit for benchmarking can be found in this notebook
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent