Skip to content

Fix subtle logic bugs and remove duplicate definitions #2

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
chrishwiggins wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Fix subtle logic bugs and remove duplicate definitions #2

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
39 changes: 9 additions & 30 deletions exatrack.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,7 @@
from scipy.stats import gamma
from tensorflow.keras.optimizers.schedules import LearningRateSchedule

import warnings
import pandas as pd
from glob import glob

Expand Down Expand Up @@ -208,7 +209,7 @@ def anomalous_diff_2D(track_len=20,
angles[i] = angles[i-1] + d_angles[i]

for i in range(len(positions)-1):
angle = angles[i-1]
angle = angles[i]
pesistent_disp = np.array([np.cos(angle), np.sin(angle)]).T * velocity/nb_sub_steps
positions[i+1] = positions[i] + pesistent_disp + disps[i]
positions[i+1] = (1-conf_sub_force) * positions[i+1] + conf_sub_force * anchor_positions[i]
Expand Down Expand Up @@ -333,7 +334,7 @@ def padding(track_list, frame_list):
nb_tracks = len(track_list)
padded_tracks = np.zeros((nb_tracks, max_len, track_list[0].shape[1]), dtype = track_list[0].dtype)
padded_frames = np.zeros((nb_tracks, max_len), dtype = frame_list[0].dtype)
mask = np.zeros((nb_tracks, max_len), dtype = track.dtype)
mask = np.zeros((nb_tracks, max_len), dtype = track_list[0].dtype)

for i, track in enumerate(track_list):
if track.shape[0]>=start_len:
Expand All @@ -343,7 +344,7 @@ def padding(track_list, frame_list):
mask[i, :cur_len] = 1
padded_frames[i, :cur_len] = frames
else:
raise Warning('The minimal track length supported is 2 time points. Tracks of 1 time point were discarded.')
warnings.warn('The minimal track length supported is 2 time points. Tracks of 1 time point were discarded.')

return padded_tracks, padded_frames, mask

Expand All @@ -367,7 +368,9 @@ def read_table(paths, # path of the file to read or list of paths to read multip

if type(paths) == str or type(paths) == np.str_:
paths = [paths]


colnames = list(colnames) # avoid mutating the caller's list

tracks = []
frames = []
track_IDs = []
Expand Down Expand Up @@ -437,8 +440,8 @@ def read_table(paths, # path of the file to read or list of paths to read multip
for m in opt_colnames:
opt_metrics[m].append(track[m].values[:l])

except :
print('problem with file :', path)
except Exception as e:
print('problem with file :', path, '-', e)

if zero_disp_tracks and not remove_no_disp:
print('Warning: some tracks show no displacements. To be checked if normal or not. These tracks can be removed with remove_no_disp = True')
Expand Down Expand Up @@ -1902,21 +1905,6 @@ def transition_param_function(transition_shapes, transition_rates, density, Fs,

return new_transition_shapes, new_transition_rates

class transpose_layer(tf.keras.layers.Layer):
def __init__(
self,
**kwargs):
super().__init__(**kwargs)

def build(self, input_shape):
self.built = True

def call(self, x, perm):
'''
input dimensions: time point, gaussian, track, state, observed variable
'''
return tf.transpose(x, perm = perm)

def get_model_params(model):
'''
Function to get the parameters from the model
Expand Down Expand Up @@ -2076,15 +2064,6 @@ def __call__(self, step):
decay_step = tf.reduce_max([step-self.decay_start, 0])
return self.peak_lr*(1-tf.math.exp(-step/self.warmup_steps))*tf.math.exp(-self.decay_rate*decay_step)

def MLE_loss(y_true, y_pred): # y_pred = log likelihood of the tracks shape (None, 1)
#print(y_pred)

max_LP = tf.math.reduce_max(y_pred, 1, keepdims = True)
reduced_LP = y_pred - max_LP
pred = tf.math.log(tf.math.reduce_sum(tf.math.exp(reduced_LP), 1, keepdims = True)) + max_LP

return - tf.math.reduce_mean(pred) # sum over the spatial dimensions axis

def logit(x):
return -np.log(1/x-1)

Expand Down