def main():
filename = sys.argv[1]
df = read_detection_csv.get_df(filename)
A, Ainv = read_detection_csv.get_A_Ainv(filename)
dt = calc_dt(df)
frame = None
observations = []
results_filename = os.path.splitext(filename)[0] + '.kalmanized.csv'
tracker = Tracker(dt, results_filename, meters_to_pixels=Ainv)
for _, row in df.iterrows():
# print('======= new row\n',row)
if frame is None:
# first frame
frame = int(row['frame'])
frame0 = frame
# print('-------------- first frame')
if row['frame'] > frame:
# next frame
tracker.handle_frame_data(frame, observations)
# print('-------------- frame advance')
observations = []
frame = int(row['frame'])
# print('-------------- within frame')
observations.append((row['x'], row['y']))
# current frame
# if frame >= frame0+100:
# break
tracker.handle_frame_data(frame, observations)
tracker.close()
def get_dt(kalman_filename):
fd = open(kalman_filename, mode='r')
buf = fd.read(100)
if buf.startswith('#'):
line0 = buf.split('\n')[0]
assert line0.startswith('# ')
yaml_buf = line0[2:]
yaml_data = yaml.safe_load(yaml_buf)
dt = yaml_data['dt']
else:
base1 = os.path.splitext(kalman_filename)[0]
base2 = os.path.splitext(base1)[0]
csv_filename = base2 + '.csv'
df = read_detection_csv.get_df(csv_filename)
dt = read_detection_csv.calculate_dt_1(df['timestamp'].values)
return dt
else:
print('******************* no on-food bout detected: %s' %
fname)
if skip_indiv_plots:
continue
if count > 5:
continue
count += 1
# ----------------------------------------
if 1:
raw_df = read_detection_csv.get_df(csv_filename)
plt.figure()
plt.imshow(jpeg,
interpolation='nearest',
cmap='gray',
zorder=-100)
plt.colorbar()
for obj_id, obj_df in kalman_df.groupby('obj_id'):
plt.plot(obj_df['pos_x_pix'], obj_df['pos_y_pix'], '-')
plt.plot(raw_df['x_px'],
raw_df['y_px'],
'k.',
ms=5,
zorder=-99)
plt.xlabel('x (px)')
import pandas as pd
import sys
import matplotlib.pyplot as plt
import numpy as np
import os
import scipy.misc
import read_detection_csv
fname = sys.argv[1]
df = pd.read_csv(fname, comment='#')
base1 = os.path.splitext(fname)[0]
base2 = os.path.splitext(base1)[0]
raw_df = read_detection_csv.get_df(base2 + '.csv')
if 1:
jpeg_fname = base2 + '.jpg'
jpeg = scipy.misc.imread(
jpeg_fname) # install python pillow package if AttributeError
plt.figure()
plt.imshow(jpeg, interpolation='nearest', cmap='gray', zorder=-100)
plt.colorbar()
plt.xlabel('x (px)')
plt.ylabel('y (px)')
for (obj_id, fly_df) in df.groupby('obj_id'):
plt.plot(fly_df['pos_x_pix'],
fly_df['pos_y_pix'],
lw=2,
label='id %d' % obj_id)
ax_cntl = fig.add_subplot(2, 1, 2)
for (category, category_df) in metadata_df.groupby('category'):
print("category: %s ------" % category)
for (genotype, genotype_df) in category_df.groupby('genotype'):
print(" genotype: %s ------" % genotype)
for (food_name, food_df) in genotype_df.groupby('food_type'):
print(" food_name: %s ------" % food_name)
for (fname, fly_df) in food_df.groupby('filename'):
print(" fname: %s" % fname)
food_x = fly_df.iloc[0]['food_x']
food_y = fly_df.iloc[0]['food_y']
jpeg_fname = os.path.splitext(fname)[0] + '.jpg'
jpeg = scipy.misc.imread(jpeg_fname)
df = read_detection_csv.get_df(fname)
food_distance = np.sqrt((df['x'] - food_x)**2 +
(df['y'] - food_y)**2)
food_dist_threshold = fly_df['food_radius'].iloc[0]
first_valid_frame = fly_df['first_valid_frame'].iloc[0]
print('first_valid_frame', first_valid_frame)
on_food_condition = food_distance < food_dist_threshold
on_food_idxs = np.nonzero(on_food_condition)[0]
print(on_food_idxs.shape)
if len(on_food_idxs) >= 1:
if not np.isnan(first_valid_frame):
# We had some bad tracking prior to this time, do not consider