def __init__(self, config: Config, dataset_label: str) -> None:
self.batch_size = config["batch_size"]
input_filename = config.filename(dataset_label + "_inputs")
label_filename = config.filename(dataset_label + "_labels")
self.input_matrix = self._load_data(input_filename)
self.label_matrix = self._load_data(label_filename)
self.input_dim = self.input_matrix.shape[1]
def test_update_single_setting(default_config, write_yaml):
filename = write_yaml({"fingerprint_len": 10})
config = Config(filename)
assert config["fingerprint_len"] == 10
assert config["template_occurance"] == default_config["template_occurance"]
def test_preprocess_rollout(write_yaml, shared_datadir, add_cli_arguments):
config_path = write_yaml(
{
"file_prefix": str(shared_datadir / "dummy"),
"split_size": {"training": 0.6, "testing": 0.2, "validation": 0.2},
}
)
add_cli_arguments(config_path)
rollout_main()
with open(shared_datadir / "dummy_template_library.csv", "r") as fileobj:
lines = fileobj.read().splitlines()
assert len(lines) == 10
with open(shared_datadir / "dummy_training.csv", "r") as fileobj:
lines = fileobj.read().splitlines()
assert len(lines) == 6
with open(shared_datadir / "dummy_testing.csv", "r") as fileobj:
lines = fileobj.read().splitlines()
assert len(lines) == 2
with open(shared_datadir / "dummy_validation.csv", "r") as fileobj:
lines = fileobj.read().splitlines()
assert len(lines) == 2
data = pd.read_hdf(shared_datadir / "dummy_unique_templates.hdf5", "table")
config = Config(config_path)
assert len(data) == 2
assert "retro_template" in data.columns
assert "library_occurence" in data.columns
for column in config["metadata_headers"]:
assert column in data.columns
def test_update_invalid_setting(default_config, write_yaml):
filename = write_yaml(
{"fingerprint_len": {"training": 0.8, "testing": 0.1, "validation": 0.1}}
)
config = Config(filename)
assert config["fingerprint_len"] == default_config["fingerprint_len"]
def _get_config() -> Config:
parser = argparse.ArgumentParser(
"Tool to pre-process a template library to be used in training a expansion network policy"
)
parser.add_argument("config", help="the filename to a configuration file")
args = parser.parse_args()
return Config(args.config)
def _get_config():
parser = argparse.ArgumentParser(
"Tool to pre-process a template library to be used to train a recommender network"
)
parser.add_argument("config", help="the filename to a configuration file")
args = parser.parse_args()
return Config(args.config)
def main():
""" Entry-point for the aizynth_training tool
"""
parser = argparse.ArgumentParser("Tool to train a network policy")
parser.add_argument("config", help="the filename to a configuration file")
args = parser.parse_args()
config = Config(args.config)
train_rollout_keras_model(config)
def test_update_nested_setting(default_config, write_yaml):
filename = write_yaml(
{"split_size": {"training": 0.8, "testing": 0.1, "validation": 0.1}}
)
config = Config(filename)
assert config["template_occurance"] == default_config["template_occurance"]
assert config["split_size"]["training"] == 0.8
assert config["split_size"]["testing"] == 0.1
assert config["split_size"]["validation"] == 0.1
def _save_unique_templates(dataset: pd.DataFrame, config: Config) -> None:
template_group = dataset.groupby("template_hash", sort=False).size()
dataset = dataset[["retro_template", "template_code"] +
config["metadata_headers"]]
if "classification" in dataset.columns:
dataset["classification"].fillna("-", inplace=True)
dataset = dataset.drop_duplicates(subset="template_code", keep="first")
dataset["library_occurence"] = template_group.values
dataset.set_index("template_code", inplace=True)
dataset = dataset.sort_index()
dataset.to_hdf(config.filename("unique_templates"), "table")
def _get_config():
parser = argparse.ArgumentParser(
"Tool to generate artifical negative reactions")
parser.add_argument("config", help="the filename to a configuration file")
parser.add_argument(
"method",
choices=["strict", "random", "recommender"],
help="the method to create random data",
)
args = parser.parse_args()
return Config(args.config), args.method
def _filter_dataset(config: Config) -> pd.DataFrame:
filename = config.filename("raw_library")
if not os.path.exists(filename):
raise FileNotFoundError(
f"The file {filename} is missing - cannot proceed without the full template library."
)
# Skipping the last header as it is not available in the raw data
full_data = pd.read_csv(
filename,
index_col=False,
header=None,
names=config["library_headers"][:-1],
)
if config["remove_unsanitizable_products"]:
products = full_data["products"].to_numpy()
idx = np.apply_along_axis(is_sanitizable, 0, [products])
full_data = full_data[idx]
full_data = full_data.drop_duplicates(subset="reaction_hash")
template_group = full_data.groupby("template_hash")
template_group = template_group.size().sort_values(ascending=False)
min_index = template_group[
template_group >= config["template_occurrence"]].index
dataset = full_data[full_data["template_hash"].isin(min_index)]
template_labels = LabelEncoder()
dataset = dataset.assign(
template_code=template_labels.fit_transform(dataset["template_hash"]))
dataset.to_csv(
config.filename("library"),
mode="w",
header=False,
index=False,
)
return dataset
def test_preprocess_expansion_no_class(write_yaml, shared_datadir,
add_cli_arguments):
config_path = write_yaml({
"library_headers": [
"index",
"ID",
"reaction_hash",
"reactants",
"products",
"retro_template",
"template_hash",
],
"metadata_headers": ["template_hash"],
"file_prefix":
str(shared_datadir / "dummy_noclass"),
"split_size": {
"training": 0.6,
"testing": 0.2,
"validation": 0.2
},
})
add_cli_arguments(config_path)
expansion_main()
with open(shared_datadir / "dummy_noclass_template_library.csv",
"r") as fileobj:
lines = fileobj.read().splitlines()
assert len(lines) == 10
with open(shared_datadir / "dummy_noclass_training.csv", "r") as fileobj:
lines = fileobj.read().splitlines()
assert len(lines) == 6
with open(shared_datadir / "dummy_noclass_testing.csv", "r") as fileobj:
lines = fileobj.read().splitlines()
assert len(lines) == 2
with open(shared_datadir / "dummy_noclass_validation.csv", "r") as fileobj:
lines = fileobj.read().splitlines()
assert len(lines) == 2
data = pd.read_hdf(shared_datadir / "dummy_noclass_unique_templates.hdf5",
"table")
config = Config(config_path)
assert len(data) == 2
assert "retro_template" in data.columns
assert "library_occurrence" in data.columns
for column in config["metadata_headers"]:
assert column in data.columns
def main() -> None:
"""Entry-point for the aizynth_training tool"""
parser = argparse.ArgumentParser("Tool to train a network policy")
parser.add_argument("config", help="the filename to a configuration file")
parser.add_argument(
"model",
choices=["expansion", "filter", "recommender"],
help="the model to train",
)
args = parser.parse_args()
config = Config(args.config)
if args.model == "expansion":
train_expansion_keras_model(config)
elif args.model == "filter":
train_filter_keras_model(config)
elif args.model == "recommender":
train_recommender_keras_model(config)
def _setup_callbacks(config: Config) -> List[Any]:
early_stopping = EarlyStopping(monitor="val_loss", patience=10)
csv_logger = CSVLogger(config.filename("_keras_training.log"), append=True)
checkpoint_path = os.path.join(config["output_path"], "checkpoints")
if not os.path.exists(checkpoint_path):
os.mkdir(checkpoint_path)
checkpoint = ModelCheckpoint(
os.path.join(checkpoint_path, "keras_model.hdf5"),
monitor="loss",
save_best_only=True,
)
reduce_lr = ReduceLROnPlateau(
monitor="val_loss",
factor=0.5,
patience=5,
verbose=0,
mode="auto",
min_delta=0.000001,
cooldown=0,
min_lr=0,
)
return [early_stopping, csv_logger, checkpoint, reduce_lr]
def __init__(self, config: Config, dataset_label: str) -> None:
super().__init__(config, dataset_label)
filename = config.filename(dataset_label + "_inputs2")
self.input_matrix2 = self._load_data(filename)
def default_config():
return Config()
def _save_unique_templates(dataset: pd.DataFrame, config: Config) -> None:
dataset = dataset[["retro_template", "template_code"]]
dataset = dataset.drop_duplicates(subset="template_code", keep="first")
dataset.set_index("template_code", inplace=True)
dataset = dataset.sort_index()
dataset.to_hdf(config.filename("unique_templates"), "table")
def test_empty_config(default_config, write_yaml):
filename = write_yaml({})
config = Config(filename)
assert config._config == default_config._config
