Source code for genepy3d.io.trackmate
"""This module supports reading tracked file from `TrackMate <https://imagej.net/plugins/trackmate/>`_.
"""
import pandas as pd
import numpy as np
from genepy3d.objext.simpletracks import SimpleTrack
[docs]
def read_spots(filepath):
"""Read tracked spots table exported from TrackMate.
Args:
filepath (str): path to the csv file.
Returns:
Spots object.
"""
spotsdf = pd.read_csv(filepath)
spotsdf = spotsdf.iloc[3:] # remove first 3 usefuless rows
spotsdf.columns = [str.lower(col) for col in spotsdf.columns] # convert column names to lowercase
del spotsdf['label'] # remove label column
spotsdf = spotsdf.apply(pd.to_numeric) # convert data to numeric type
return Spots(spotsdf)
[docs]
class Spots:
"""Handle tracked spots table.
Attributes:
spots (dataframe): Spots table.
trackid_lst (numpy array): list of trackid from the Spots table.
nb_tracks (int): number of tracks.
Notes:
The spots table must be imported from ``trackmate.read_spots()``.
Examples:
.. code-block:: python
import numpy as np
from genepy3d.io import trackmate
# Csv file of spots table from Trackmate
spots = trackmate.read_spots("path/to/spots/csv/file")
"""
def __init__(self,spotsdf):
self.spots = spotsdf
self.trackid_lst = spotsdf['track_id'].unique()
self.nb_tracks = len(self.trackid_lst)
[docs]
def tracks_nb_spots(self,ascending=False):
"""Return number of spots per track.
Args:
ascending (bool): if True, sort by ascending order.
Returns:
A Pandas series index by trackid.
"""
return self.spots.groupby("track_id")['id'].count().sort_values(ascending=ascending)
[docs]
def get_tracks(self,trackid_lst,return_type="simpletrack",x_col="position_x",y_col="position_y",z_col="position_z",t_col="frame"):
"""Return tracks data given by trackid_lst.
Support two returns: "dataframe" as a Pandas dataframe and "simpletrack" as ``SimpleTrack`` objects.
Args:
trackid_lst (numpy array or int): a single trackid or list of trackid.
return_type (str): if "dataframe" then return a Pandas dataframe, if "simpletrack" then return ``SimpleTrack`` objects.
x_col (str): x column name.
y_col (str): y column name.
z_col (str): z column name.
t_col (str): time column name.
Returns:
A Pandas dataframe or SimpleTrack objects depend on condition.
Notes:
The parameters `x_col`, `y_col`, `z_col` and `t_col` are only needed if return_type = "simpletrack".
"""
if isinstance(trackid_lst,(int,np.integer)):
trackid_lst_new = np.array([trackid_lst])
else:
trackid_lst_new = trackid_lst
if return_type == "dataframe":
tracks = self.spots[self.spots["track_id"].isin(trackid_lst_new)].copy()
return tracks
elif return_type == "simpletrack":
simtrack_lst = []
for trackid in trackid_lst_new:
track = self.spots[self.spots["track_id"]==trackid].copy()
track.sort_values("frame",inplace=True)
coors = track[[x_col,y_col,z_col]].values
t = track[t_col].values
new_traj = SimpleTrack(coors,t)
if np.sum(coors[:,2])==0: # 2D trajectories
new_traj.dim = 2
else:
new_traj.dim = 3
simtrack_lst.append(new_traj)
if len(simtrack_lst)==1:
return simtrack_lst[0] # no need to return list if only one item.
else:
return simtrack_lst
else:
raise Exception("return_type must be either 'dataframe' or 'simpletrack'.")
[docs]
def remove_tracks(self,trackid_lst_excluded):
"""Remove tracks from the table.
Args:
trackid_lst_excluded (numpy array): list of trackids to be removed.
Returns:
Itself. Tracks with given trackids will be removed from the spots table.
"""
new_trackid_lst = np.setdiff1d(self.trackid_lst,trackid_lst_excluded)
self.trackid_lst = new_trackid_lst
self.nb_tracks = len(new_trackid_lst)
spots = self.spots
spots = spots[spots['track_id'].isin(new_trackid_lst)]
self.spots = spots # maybe no need this line
