Skip to content

Update Emitter to use SimulariumIO AgentData from lists #22

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
ascibisz wants to merge 2 commits into
base: buckling
Choose a base branch
from
Open

Update Emitter to use SimulariumIO AgentData from lists #22

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
8dd67e7
Use simulariumio AgentData from lists functionality
ascibisz May 5, 2023
7a23c4b
remove unused input
ascibisz May 23, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
92 changes: 17 additions & 75 deletions vivarium_models/processes/simularium_emitter.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,6 @@
from vivarium.core.emitter import Emitter

import numpy as np
import pandas as pd
from simulariumio import (
TrajectoryConverter,
TrajectoryData,
Expand Down Expand Up @@ -110,83 +109,26 @@ def get_simularium_monomers(self, time, monomers, actin_radius, trajectory):
return trajectory

@staticmethod
def fill_df(df, fill):
def get_scaled_agent_data(trajectory, scale_factor) -> AgentData:
"""
Fill Nones in a DataFrame with a fill value
"""
# Create a dataframe of fill values
fill_array = [[fill] * df.shape[1]] * df.shape[0]
fill_df = pd.DataFrame(fill_array)
# Replace all entries with None with the fill
df[df.isna()] = fill_df
return df

@staticmethod
def jagged_3d_list_to_numpy_array(jagged_3d_list):
"""
Shape a jagged list with 3 dimensions to a numpy array
"""
df = SimulariumEmitter.fill_df(pd.DataFrame(jagged_3d_list), [0.0, 0.0, 0.0])
df_t = df.transpose()
exploded = [df_t[col].explode() for col in list(df_t.columns)]
return np.array(exploded).reshape((df.shape[0], df.shape[1], 3))

@staticmethod
def get_subpoints_numpy_array(trajectory) -> np.ndarray:
"""
Shape a 4 dimensional jagged list for subpoints into a numpy array
"""
frame_arrays = []
max_agents = 0
max_subpoints = 0
total_steps = len(trajectory["subpoints"])
for time_index in range(total_steps):
frame_array = SimulariumEmitter.jagged_3d_list_to_numpy_array(
trajectory["subpoints"][time_index]
)
if frame_array.shape[0] > max_agents:
max_agents = frame_array.shape[0]
if frame_array.shape[1] > max_subpoints:
max_subpoints = frame_array.shape[1]
frame_arrays.append(frame_array)
values_per_frame = max_agents * max_subpoints * 3
result = np.zeros(total_steps * values_per_frame)
for time_index, frame_array in enumerate(frame_arrays):
if frame_array.shape[1] < max_subpoints:
new_frame_array = np.zeros((frame_array.shape[0], max_subpoints, 3))
new_frame_array[:, : frame_array.shape[1]] = frame_array
frame_array = new_frame_array
flat_array = frame_array.flatten()
start_index = time_index * values_per_frame
result[start_index : start_index + flat_array.shape[0]] = flat_array
return result.reshape(total_steps, max_agents, max_subpoints, 3)

@staticmethod
def get_agent_data_from_jagged_lists(trajectory, scale_factor) -> AgentData:
"""
Shape a dictionary of jagged lists into a Simularium AgentData object
Build AgentData object, scaling appropriate values by scale_factor
"""
return AgentData(
times=np.arange(len(trajectory["times"])),
n_agents=np.array(trajectory["n_agents"]),
viz_types=SimulariumEmitter.fill_df(
pd.DataFrame(trajectory["viz_types"]), 1000.0
).to_numpy(),
unique_ids=SimulariumEmitter.fill_df(
pd.DataFrame(trajectory["unique_ids"]), 0
).to_numpy(dtype=int),
times=trajectory["times"],
n_agents=trajectory["n_agents"],
viz_types=trajectory["viz_types"],
unique_ids=trajectory["unique_ids"],
types=trajectory["type_names"],
positions=scale_factor
* SimulariumEmitter.jagged_3d_list_to_numpy_array(trajectory["positions"]),
radii=scale_factor
* SimulariumEmitter.fill_df(
pd.DataFrame(trajectory["radii"]), 0.0
).to_numpy(),
n_subpoints=SimulariumEmitter.fill_df(
pd.DataFrame(trajectory["n_subpoints"]), 0
).to_numpy(dtype=int),
subpoints=scale_factor
* SimulariumEmitter.get_subpoints_numpy_array(trajectory),
positions=[
[[k * scale_factor for k in j] for j in i]
for i in trajectory["positions"]
],
radii=[[j * scale_factor for j in i] for i in trajectory["radii"]],
n_subpoints=trajectory["n_subpoints"],
subpoints=[
[[k * scale_factor for k in j] for j in i]
for i in trajectory["subpoints"]
],
)

@staticmethod
Expand All @@ -204,7 +146,7 @@ def get_simularium_converter(
meta_data=MetaData(
box_size=scale_factor * box_dimensions,
),
agent_data=SimulariumEmitter.get_agent_data_from_jagged_lists(
agent_data=SimulariumEmitter.get_scaled_agent_data(
trajectory, scale_factor
),
time_units=UnitData("count"),
Expand Down