def get_events(path):
events = {}
event_table = pd.read_csv(str(path))
for idx, event_row in event_table.iterrows():
if event_row["actually_built"] is True:
pandda_processed_dir = Path(event_row["event_map_path"]).parent
print(pandda_processed_dir)
if pandda_processed_dir.exists():
try:
event_row["ligand_smiles_path"] = get_ligand_smiles(
Path(event_row["event_map_path"]).parent)
print("\tLigand smiles path:{}".format(
event_row["ligand_smiles_path"]))
event = Event.from_record(event_row)
# events.append(event)
events[(event.pandda_name, event.dtag,
event.event_idx)] = event
except Exception as e:
print(e)
continue
else:
continue
return events
def get_events(events_csv_path):
events_table = pd.read_csv(str(events_csv_path))
events: List[Event] = []
for idx, event_record in events_table.iterrows():
# event: Event = Event(dtag=str(event_record["dtag"]),
# event_idx=int(event_record["event_idx"]),
# occupancy=event_record["occupancy"],
# analysed_resolution=event_record["analysed_resolution"],
# high_resolution=event_record["high_resolution"],
# interesting=event_record["interesting"],
# ligand_placed=event_record["ligand_placed"],
# ligand_confidence=event_record["ligand_confidence"],
# viewed=event_record["viewed"],
# initial_model_path=Path(event_record["initial_model_path"]),
# data_path=Path(event_record["data_path"]),
# final_model_path=Path(event_record["final_model_path"]),
# event_map_path=Path(event_record["event_map_path"]),
# actually_built=event_record["actually_built"],
# x=event_record["x"],
# y=event_record["y"],
# z=event_record["z"],
# )
event = Event.from_record(event_record)
events.append(event)
return events
def get_events(pandda: PanDDA):
events: Dict[Tuple[str, int], Event] = {}
inspect_table: pd.DataFrame = pd.read_csv(str(pandda.event_table_path))
processed_datasets_path: Path = pandda.dir / "processed_datasets"
for idx, event_record in inspect_table.iterrows():
dataset_path = processed_datasets_path / "{}".format(event_record["dtag"])
modelled_structures_path = dataset_path / "{}".format("modelled_structures")
initial_model_path = dataset_path / "{}-pandda-input.pdb".format(event_record["dtag"])
data_path = dataset_path / "{}-pandda-input.mtz".format(event_record["dtag"])
final_model_path = modelled_structures_path / "{}-pandda-model.pdb".format(event_record["dtag"])
# final_model_path = modelled_structures_path / "fitted-v0001".format(event_record["dtag"])
event_map_path = dataset_path / "{}-event_{}_1-BDC_{}_map.native.ccp4".format(event_record["dtag"],
event_record["event_idx"],
event_record["1-BDC"],
)
ligand_smiles = get_ligand_smiles(dataset_path)
actually_built = is_actually_built(final_model_path)
if actually_built:
distance_to_ligand_model = get_distance_to_ligand_model(np.array([event_record["x"],
event_record["y"],
event_record["z"]]
),
final_model_path)
# print("\t\tDistance to model is: {}".format(distance_to_ligand_model))
else:
distance_to_ligand_model = 0
event: Event = Event(dtag=str(event_record["dtag"]),
event_idx=int(event_record["event_idx"]),
occupancy=event_record["1-BDC"],
analysed_resolution=event_record["analysed_resolution"],
high_resolution=event_record["high_resolution"],
interesting=event_record["Interesting"],
ligand_placed=event_record["Ligand Placed"],
ligand_confidence=event_record["Ligand Confidence"],
viewed=event_record["Viewed"],
initial_model_path=initial_model_path,
data_path=data_path,
model_dir=pandda.model_dir,
pandda_dir=pandda.dir,
pandda_name=pandda.dir.name,
final_model_path=final_model_path,
event_map_path=event_map_path,
actually_built=actually_built,
ligand_smiles_path=ligand_smiles,
x=event_record["x"],
y=event_record["y"],
z=event_record["z"],
distance_to_ligand_model=distance_to_ligand_model,
event_size=event_record["cluster_size"],
)
events[(event.dtag, event.event_idx)] = event
return events
def get_events(events_table):
events: List[Event] = []
for idx, event_record in events_table.iterrows():
event: Event = Event(
dtag=str(event_record["dtag"]),
event_idx=int(event_record["event_idx"]),
occupancy=event_record["occupancy"],
analysed_resolution=event_record["analysed_resolution"],
high_resolution=event_record["high_resolution"],
interesting=event_record["interesting"],
ligand_placed=event_record["ligand_placed"],
ligand_confidence=event_record["ligand_confidence"],
viewed=event_record["viewed"],
initial_model_path=Path(event_record["initial_model_path"]),
data_path=Path(event_record["data_path"]),
final_model_path=Path(event_record["final_model_path"]),
event_map_path=Path(event_record["event_map_path"]),
actually_built=event_record["actually_built"],
x=event_record["x"],
y=event_record["y"],
z=event_record["z"],
ligand_smiles_path=event_record["ligand_smiles_path"],
model_dir=event_record["model_dir"],
pandda_dir=event_record["pandda_dir"],
pandda_name=event_record["pandda_name"],
)
events.append(event)
return events
def get_events(events_table: pd.DataFrame):
events = {}
for idx, row in events_table.iterrows():
event = Event.from_record(row)
event_id = EventID(row["dtag"], row["event_idx"])
events[event_id] = event
return events
def get_events(event_talbe_path):
event_table = pd.read_csv(str(event_talbe_path))
events = {}
for event_id, row in event_table.iterrows():
event = Event.from_record(row)
events[(event.pandda_name, event.dtag, event.event_idx)] = event
return events
def get_events(events_csv_path):
events_table = pd.read_csv(str(events_csv_path))
events: List[Event] = []
for idx, event_record in events_table.iterrows():
event = Event.from_record(event_record)
events.append(event)
return events, events_table
def get_event_table(path):
events = []
event_table = pd.read_csv(str(path))
for idx, event_row in event_table.iterrows():
if event_row["actually_built"] is True:
event = Event.from_record(event_row)
events.append(event)
else:
continue
return events
def from_pandda_path(path: Path):
pandda_dir = path
events_csv_path = path / "analyses" / "pandda_analyse_events.csv"
pandda_json_path = path / "pandda.json"
with open(str(pandda_json_path), "r") as f:
pandda_json_string = f.read()
pandda_json_dict = json.loads(pandda_json_string)
data_dirs = Path(pandda_json_dict["data_dirs"]).parent
event_table = pd.read_csv(str(events_csv_path))
events = Event.from_pandda_path(pandda_dir,
data_dirs,
)
pandda = PanDDA(data_dir=data_dirs,
pandda_dir=pandda_dir,
events_table_path=events_csv_path,
events=events,
)
return pandda
def __getitem__(self, item):
event_record = self.table.iloc[item]
event = Event.from_record(event_record)
# print(event.viewed)
# print(type(event.viewed))
if bool(event.viewed) is False:
sample_dict = {
"id": {
"pandda_name": event.pandda_name,
"dtag": event.dtag,
"event_idx": event.event_idx,
},
"data":
np.zeros(
(
2,
self.sample_shape[0],
self.sample_shape[1],
self.sample_shape[2],
),
dtype=np.float32,
),
"label":
np.array([1, 0], dtype=np.float32),
"event_map_path":
str(event.event_map_path),
"model_path":
str(event.initial_model_path),
"coords":
str([event.x, event.y, event.z])
}
return sample_dict
try:
event_centroid = np.array([
event.x,
event.y,
event.z,
])
data_map = get_map_from_mtz_path(event.data_path)
event_map: gemmi.Grid = gemmi.read_ccp4_map(event.event_map_path)
rotation = sample_rotation()
translation = sample_translation()
data_map_layer = sample_map(
data_map,
event_centroid,
rotation,
translation,
self.sample_shape,
)
event_map_layer = sample_event_map(
event_map,
event_centroid,
rotation,
translation,
self.sample_shape,
)
data = np.stack(
[data_map_layer, event_map_layer],
axis=0,
)
label = get_label(event)
sample_dict = {
"id": {
"pandda_name": event.pandda_name,
"dtag": event.dtag,
"event_idx": event.event_idx,
},
"data": data,
"label": label,
"event_map_path": str(event.event_map_path),
"model_path": str(event.initial_model_path),
"coords": str([event.x, event.y, event.z])
}
return sample_dict
except Exception as e:
# Null result
# print(e)
sample_dict = {
"id": {
"pandda_name": event.pandda_name,
"dtag": event.dtag,
"event_idx": event.event_idx,
# "initial_model_dir": ,
},
"data":
np.zeros(
(
2,
self.sample_shape[0],
self.sample_shape[1],
self.sample_shape[2],
),
dtype=np.float32,
),
"label":
np.array([1, 0], dtype=np.float32),
"event_map_path":
str(event.event_map_path),
"model_path":
str(event.initial_model_path),
"coords":
str([event.x, event.y, event.z])
}
return sample_dict
def __getitem__(self, item):
event_record = self.table.iloc[item]
event = Event.from_record(event_record)
rscc = self.rscc_dict[(event.pandda_name, event.dtag, event.event_idx)]
# if bool(event.viewed) is False:
# sample_dict = {"id": {"pandda_name": event.pandda_name,
# "dtag": event.dtag,
# "event_idx": event.event_idx,
# },
# "data": np.zeros((2,
# self.sample_shape[0],
# self.sample_shape[1],
# self.sample_shape[2],
# ),
# dtype=np.float32,
# ),
# "label": np.array([1, 0], dtype=np.float32),
# "event_map_path": str(event.event_map_path),
# "model_path": str(event.initial_model_path),
# "coords": str([event.x, event.y, event.z]),
# "rscc": 0,
# }
#
# return sample_dict
if float(rscc) == 0.0:
sample_dict = {"id": {"pandda_name": event.pandda_name,
"dtag": event.dtag,
"event_idx": event.event_idx,
},
"data": np.zeros((2,
self.sample_shape[0],
self.sample_shape[1],
self.sample_shape[2],
),
dtype=np.float32,
),
"label": np.array([1, 0], dtype=np.float32),
"event_map_path": str(event.event_map_path),
"model_path": str(event.initial_model_path),
"coords": str([event.x, event.y, event.z]),
"rscc": 0.0,
"rscc_class": np.array([1.0,0.0], dtype=np.float32),
}
return sample_dict
try:
event_centroid = np.array([event.x,
event.y,
event.z,
])
data_map = get_map_from_mtz_path(event.data_path)
event_map: gemmi.Grid = gemmi.read_ccp4_map(event.event_map_path)
rotation = sample_rotation()
translation = sample_translation()
data_map_layer = sample_map(data_map,
event_centroid,
rotation,
translation,
self.sample_shape,
)
event_map_layer = sample_event_map(event_map,
event_centroid,
rotation,
translation,
self.sample_shape,
)
data = np.stack([data_map_layer, event_map_layer],
axis=0,
)
label = get_label(event)
if rscc <0.7:
rscc_class = np.array([1.0, 0.0], dtype=np.float32)
else:
rscc_class = np.array([0.0,1.0], dtype=np.float32)
sample_dict = {"id": {"pandda_name": event.pandda_name,
"dtag": event.dtag,
"event_idx": event.event_idx,
},
"data": data,
"label": label,
"event_map_path": str(event.event_map_path),
"model_path": str(event.initial_model_path),
"coords": str([event.x, event.y, event.z]),
"rscc": rscc,
"rscc_class": rscc_class,
}
return sample_dict
except Exception as e:
# Null result
# print(e)
sample_dict = {"id": {"pandda_name": event.pandda_name,
"dtag": event.dtag,
"event_idx": event.event_idx,
# "initial_model_dir": ,
},
"data": np.zeros((2,
self.sample_shape[0],
self.sample_shape[1],
self.sample_shape[2],
),
dtype=np.float32,
),
"label": np.array([1, 0], dtype=np.float32),
"event_map_path": str(event.event_map_path),
"model_path": str(event.initial_model_path),
"coords": str([event.x, event.y, event.z]),
"rscc": 0.0,
"rscc_class": np.array([1, 0], dtype=np.float32),
}
return sample_dict
def __getitem__(self, item):
event_record = self.table.iloc[item]
event = Event.from_record(event_record)
# print(event.viewed)
# print(type(event.viewed))
if bool(event.viewed) is False:
sample_dict = {
"id": {
"pandda_name": event.pandda_name,
"dtag": event.dtag,
"event_idx": event.event_idx,
},
"data": np.zeros(
self.sample_shape,
dtype=np.float32,
),
"label": np.array([1, 0], dtype=np.float32),
}
return sample_dict
try:
event_centroid = np.array([
event.x,
event.y,
event.z,
])
event_map: gemmi.Grid = gemmi.read_ccp4_map(event.event_map_path)
rotation = sample_rotation()
translation = sample_translation()
event_map_layer = sample_event_map(
event_map,
event_centroid,
rotation,
translation,
self.sample_shape,
)
data = event_map_layer
label = get_label(event)
sample_dict = {
"id": {
"pandda_name": event.pandda_name,
"dtag": event.dtag,
"event_idx": event.event_idx,
},
"data": data,
"label": label,
}
return sample_dict
except:
# Null result
sample_dict = {
"id": {
"pandda_name": event.pandda_name,
"dtag": event.dtag,
"event_idx": event.event_idx,
},
"data": np.zeros(
self.sample_shape,
dtype=np.float32,
),
"label": np.array([1, 0], dtype=np.float32),
}
return sample_dict