def _process_multi_smiles(filename: str, finder: AiZynthFinder,
output_name: str, do_clustering: bool) -> None:
output_name = output_name or "output.hdf5"
with open(filename, "r") as fileobj:
smiles = [line.strip() for line in fileobj.readlines()]
results = defaultdict(list)
for smi in smiles:
finder.target_smiles = smi
finder.prepare_tree()
search_time = finder.tree_search()
finder.build_routes()
stats = finder.extract_statistics()
logger().info(f"Done with {smi} in {search_time:.3} s")
if do_clustering:
_do_clustering(finder, stats, detailed_results=True)
for key, value in stats.items():
results[key].append(value)
results["top_scores"].append(", ".join(
"%.4f" % score for score in finder.routes.scores))
results["trees"].append(finder.routes.dicts)
data = pd.DataFrame.from_dict(results)
with warnings.catch_warnings(): # This wil suppress a PerformanceWarning
warnings.simplefilter("ignore")
data.to_hdf(output_name, key="table", mode="w")
logger().info(f"Output saved to {output_name}")
def _process_single_smiles(
smiles: str,
finder: AiZynthFinder,
output_name: str,
do_clustering: bool,
route_distance_model: str = None,
) -> None:
output_name = output_name or "trees.json"
finder.target_smiles = smiles
finder.prepare_tree()
finder.tree_search(show_progress=True)
finder.build_routes()
with open(output_name, "w") as fileobj:
json.dump(finder.routes.dicts, fileobj, indent=2)
logger().info(f"Trees saved to {output_name}")
scores = ", ".join("%.4f" % score for score in finder.routes.scores)
logger().info(f"Scores for best routes: {scores}")
stats = finder.extract_statistics()
if do_clustering:
_do_clustering(finder,
stats,
detailed_results=False,
model_path=route_distance_model)
stats_str = "\n".join(f"{key.replace('_', ' ')}: {value}"
for key, value in stats.items())
logger().info(stats_str)
def process_smiles_list(finder: AiZynthFinder,
target_smiles: List[str]) -> List[Dict[str, str]]:
stats_list: List[Dict[str, str]] = []
for smiles in target_smiles:
finder.target_smiles = smiles
finder.tree_search()
finder.build_routes()
stats_list.append(finder.extract_statistics())
return stats_list
