def fingerprint_distance_generator():
"""
Get FingerprintDistanceGenerator instance with dummy data.
Returns
-------
kissim.similarity.FingerprintDistanceGenerator
Fingerprint distance for multiple fingerprint pairs.
"""
data = pd.DataFrame({
"structure.1": ["pdbA", "pdbA", "pdbB"],
"structure.2": ["pdbB", "pdbC", "pdbC"],
"kinase.1": ["kinaseA", "kinaseA", "kinaseA"],
"kinase.2": ["kinaseA", "kinaseB", "kinaseB"],
"distance": np.array([0.75, 1.0, 0.8]),
"bit_coverage": np.array([1.0, 1.0, 1.0]),
})
# FingerprintDistanceGenerator
fingerprint_distance_generator = FingerprintDistanceGenerator()
fingerprint_distance_generator.data = data
fingerprint_distance_generator.structure_kinase_ids = [
("pdbA", "kinaseA"),
("pdbB", "kinaseA"),
("pdbC", "kinaseB"),
]
return fingerprint_distance_generator
def weight_feature_distances(
feature_distances_generator, output_filepath=None, feature_weights=None
):
"""
Weight feature distances: Generates per fingerprint pair a fingerprint distance.
Parameters
----------
feature_distances_generator : kissim.encoding.FeatureDistancesGenerator
Feature distances.
output_path : str
Path to output folder.
feature_weights : None or list of float
Feature weights of the following form:
(i) None
Default feature weights: All features equally distributed to 1/15
(15 features in total).
(ii) By feature (list of 15 floats):
Features to be set in the following order: size, hbd, hba, charge, aromatic,
aliphatic, sco, exposure, distance_to_centroid, distance_to_hinge_region,
distance_to_dfg_region, distance_to_front_pocket, moment1, moment2, and moment3.
All floats must sum up to 1.0.
"""
fingerprint_distance_generator = FingerprintDistanceGenerator.from_feature_distances_generator(
feature_distances_generator, feature_weights
)
if output_filepath:
output_filepath = Path(output_filepath)
fingerprint_distance_generator.to_csv(output_filepath)
return fingerprint_distance_generator
def test_from_fingerprint_generator(self, fingerprint_generator):
"""
Test FingerprintDistanceGenerator generation from a fingerprint generator.
"""
fingerprint_distance_generator = FingerprintDistanceGenerator.from_fingerprint_generator(
fingerprint_generator
)
assert isinstance(fingerprint_distance_generator, FingerprintDistanceGenerator)
def test_from_structure_klifs_ids(self, structure_klifs_ids, klifs_session, feature_weights):
"""
Test FingerprintDistanceGenerator generation from structure KLIFS IDs.
"""
fingerprint_distance_generator = FingerprintDistanceGenerator.from_structure_klifs_ids(
structure_klifs_ids, klifs_session, feature_weights
)
assert isinstance(fingerprint_distance_generator, FingerprintDistanceGenerator)
def test_from_structure_klifs_ids(self, structure_klifs_ids, klifs_session, feature_weights):
"""
Test FeatureDistancesGenerator class attributes.
"""
# Test FeatureDistancesGenerator class attributes
feature_distances_generator = FingerprintDistanceGenerator.from_structure_klifs_ids(
structure_klifs_ids, klifs_session, feature_weights
)
assert isinstance(feature_distances_generator, FingerprintDistanceGenerator)
def weight_feature_distances(
feature_distances_generator, output_filepath=None, feature_weights=None
):
"""
Weight feature distances: Generates per fingerprint pair a fingerprint distance.
Parameters
----------
feature_distances_generator : kissim.encoding.FeatureDistancesGenerator
Feature distances.
output_path : str
Path to output file.
feature_weights : None or list of float
Feature weights of the following form:
(i) None
Default feature weights: All features equally distributed to 1/15
(15 features in total).
(ii) By feature (list of 15 floats):
Features to be set in the following order: size, hbd, hba, charge, aromatic,
aliphatic, sco, exposure, distance_to_centroid, distance_to_hinge_region,
distance_to_dfg_region, distance_to_front_pocket, moment1, moment2, and moment3.
All floats must sum up to 1.0.
"""
fingerprint_distance_generator = FingerprintDistanceGenerator.from_feature_distances_generator(
feature_distances_generator, feature_weights
)
if output_filepath:
output_filepath = Path(output_filepath)
# Write fingerprint distances to file
fingerprint_distances_filepath = output_filepath
fingerprint_distance_generator.to_csv(fingerprint_distances_filepath)
# Write default kinase distances to file
kinase_distances_filepath = (
output_filepath.parent / f"{output_filepath.stem}_to_kinase_matrix.csv"
)
kinase_matrix = fingerprint_distance_generator.kinase_distance_matrix()
kinase_matrix.index.name = None
kinase_matrix.columns.name = None
kinase_matrix.to_csv(kinase_distances_filepath, index=True)
print(kinase_matrix)
# Write default tree to file
tree_filepath = output_filepath.parent / f"{output_filepath.stem}_to_kinase_clusters.tree"
annotation_filepath = output_filepath.parent / "kinase_annotation.csv"
tree.from_distance_matrix(kinase_matrix, tree_filepath, annotation_filepath)
return fingerprint_distance_generator
def weights(feature_distances_path, feature_weights=None, fingerprint_distances_path=None):
"""
Apply feature distances weighting to calculate fingerprint distances.
Parameters
----------
feature_distances_path : str or pathlib.Path
Path to feature distances CSV file.
feature_weights : None or list of float
Feature weights of the following form:
(i) None
Default feature weights: All features equally distributed to 1/15
(15 features in total).
(ii) By feature (list of 15 floats):
Features to be set in the following order: size, hbd, hba, charge, aromatic,
aliphatic, sco, exposure, distance_to_centroid, distance_to_hinge_region,
distance_to_dfg_region, distance_to_front_pocket, moment1, moment2, and moment3.
All floats must sum up to 1.0.
fingerprint_distances_path : None or str or pathlib.Path
Path to output fingerprint distances CSV file.
Returns
-------
kissim.comparison.FingerprintDistanceGenerator
Pairwise fingerprint distances.
"""
# Load feature distances
feature_distances_path = Path(feature_distances_path)
logger.info(f"Read feature distances from {feature_distances_path}...")
feature_distances_generator = FeatureDistancesGenerator.from_csv(feature_distances_path)
# Calculate fingerprint distances
logger.info(f"Feature weights: {feature_weights}")
fingerprint_distance_generator = FingerprintDistanceGenerator.from_feature_distances_generator(
feature_distances_generator, feature_weights
)
# Optionally: Save to file
if fingerprint_distances_path is not None:
fingerprint_distances_path = Path(fingerprint_distances_path)
logger.info(f"To file {fingerprint_distances_path}")
fingerprint_distance_generator.to_csv(fingerprint_distances_path)
return fingerprint_distance_generator
def test_from_feature_distances_generator(
self,
feature_distances_generator,
feature_weights,
structure_kinase_ids,
structure_pair_ids,
kinase_pair_ids,
structure_ids,
kinase_ids,
):
"""
Test FingerprintDistanceGenerator class attributes.
"""
# FingerprintDistanceGenerator
fingerprint_distance_generator = (
FingerprintDistanceGenerator.from_feature_distances_generator(
feature_distances_generator
)
)
# Test attributes
assert isinstance(fingerprint_distance_generator.data, pd.DataFrame)
data_columns = [
"structure.1",
"structure.2",
"kinase.1",
"kinase.2",
"distance",
"bit_coverage",
]
assert fingerprint_distance_generator.data.columns.to_list() == data_columns
assert fingerprint_distance_generator.structure_kinase_ids == structure_kinase_ids
# Test properties
assert fingerprint_distance_generator.structure_pair_ids == structure_pair_ids
assert fingerprint_distance_generator.kinase_pair_ids == kinase_pair_ids
assert fingerprint_distance_generator.structure_ids == structure_ids
assert fingerprint_distance_generator.kinase_ids == kinase_ids
assert isinstance(fingerprint_distance_generator.distances, np.ndarray)
assert isinstance(fingerprint_distance_generator.bit_coverages, np.ndarray)
def test_main_compare(fingerprint_generator, args):
"""
Test CLI for encoding using subprocesses.
"""
args = args.split()
input_filepath = Path(args[3])
output_path = Path(args[5])
# Generate
fingerprint_generator.to_json(input_filepath)
subprocess.run(args, check=True)
### Feature distances generator
# CSV file there?
feature_distances_filepath = output_path / "feature_distances.csv"
assert feature_distances_filepath.exists()
# CSV file can be loaded as FeatureDistancesGeneration object?
feature_distances_generator = FeatureDistancesGenerator.from_csv(
feature_distances_filepath)
assert isinstance(feature_distances_generator, FeatureDistancesGenerator)
assert isinstance(feature_distances_generator.data, pd.DataFrame)
### Fingerprint distance generator
# CSV file there?
fingerprint_distance_filepath = output_path / "fingerprint_distances.csv"
assert fingerprint_distance_filepath.exists()
# CSV file can be loaded as FingerprintDistanceGeneration object?
fingerprint_distance_generator = FingerprintDistanceGenerator.from_csv(
fingerprint_distance_filepath)
assert isinstance(fingerprint_distance_generator,
FingerprintDistanceGenerator)
assert isinstance(fingerprint_distance_generator.data, pd.DataFrame)
# Delete file - cannot be done within enter_tmp_directory, since temporary files
# apparently cannot be read from CLI
input_filepath.unlink()
feature_distances_filepath.unlink()
fingerprint_distance_filepath.unlink()