testscript_3d.py

"""
DeepLabCut2.0 Toolbox (deeplabcut.org)
© A. & M. Mathis Labs
https://github.com/AlexEMG/DeepLabCut
Please see AUTHORS for contributors.

https://github.com/AlexEMG/DeepLabCut/blob/master/AUTHORS
Licensed under GNU Lesser General Public License v3.0

This script tests various functionalities in an automatic way.
It produces nothing of interest scientifically.
"""


task = "TEST3D"  # Enter the name of your experiment Task
scorer = "Alex"  # Enter the name of the experimenter/labeler
num_cameras = 2  # Enter the number of cameras

import os, deeplabcut
import zipfile, urllib.request, shutil
from datetime import datetime as dt
import glob
from pathlib import Path
import subprocess

print("Imported DLC!")

basepath = os.path.dirname(os.path.abspath("testscript_3d.py"))
videoname = "reachingvideo1"

video = [
    os.path.join(
        basepath, "Reaching-Mackenzie-2018-08-30", "videos", videoname + ".avi"
    )
]

folder = "3Dtestviews_videos"
deeplabcut.auxiliaryfunctions.attempttomakefolder(folder)

# copying demo video from reaching data set and create two "views":
dst_videoname1 = "vid1_camera-1"
dst_videoname2 = "vid1_camera-2"
dst_videoname3 = "long_camera-2"
output1 = os.path.join(basepath, folder, dst_videoname1 + ".avi")
output2 = os.path.join(basepath, folder, dst_videoname2 + ".avi")

output3 = os.path.join(basepath, folder, dst_videoname3 + ".avi")
shutil.copyfile(video[0], output3)

vname = "brief"
try:  # you need ffmpeg command line interface
    subprocess.call(
        [
            "ffmpeg",
            "-i",
            video[0],
            "-ss",
            "00:00:00",
            "-to",
            "00:00:00.4",
            "-c",
            "copy",
            output1,
        ]
    )
    subprocess.call(
        [
            "ffmpeg",
            "-i",
            video[0],
            "-ss",
            "00:00:00",
            "-to",
            "00:00:00.4",
            "-c",
            "copy",
            output2,
        ]
    )
except:
    pass

"""
# copying demo video from reaching data set and create two "views":
dst_videoname1 = 'vid1_camera-1'
dst_videoname2 = 'vid1_camera-2'
output1 = os.path.join(basepath,folder,dst_videoname1+'.avi')
output2 = os.path.join(basepath,folder,dst_videoname2+'.avi')
shutil.copyfile(video[0], output1)
shutil.copyfile(video[0], output2)
"""
# checking if 2d test project is available
try:
    config = glob.glob(os.path.join(basepath, "TEST*", "config.yaml"))[-1]
except:
    raise RuntimeError("Please run the testscript.py first before testing for 3d")

dfolder = None

print("CREATING 3-D PROJECT")
path_config_file = deeplabcut.create_new_project_3d(task, scorer, num_cameras)

try:
    cfg = deeplabcut.auxiliaryfunctions.read_config(path_config_file)
    cfg["config_file_camera-1"] = config
    cfg["shuffle_camera-1"] = 1

    cfg["config_file_camera-2"] = config
    cfg["shuffle_camera-2"] = 2

    cfg["skeleton"] = [["bodypart1", "bodypart2"], ["objectA", "bodypart3"]]
    deeplabcut.auxiliaryfunctions.write_config_3d(path_config_file, cfg)
except:
    raise (
        "Please delete the project and re-try."
    )  # otherwise the cfg is an empty array!

"""
# Creating the name of the project
date = dt.today()
month = date.strftime("%B")
day = date.day
d = str(month[0:3]+str(day))
date = dt.today().strftime('%Y-%m-%d')
project_name = '{pn}-{exp}-{date}-{triangulate}'.format(pn=task, exp=scorer, date=date,triangulate='3d')
"""
project_name = path_config_file.split(os.sep)[-2]

os.chdir(os.path.join(basepath, project_name, "calibration_images"))
# Dowloading the calibration images
url = "http://www.vision.caltech.edu/bouguetj/calib_doc/htmls/stereo_example.zip"
file_name = "stereo_example.zip"
with urllib.request.urlopen(url) as response, open(file_name, "wb") as out_file:
    shutil.copyfileobj(response, out_file)
file_name = os.path.join(
    basepath, project_name, "calibration_images", "stereo_example.zip"
)
with zipfile.ZipFile(file_name) as zf:
    zf.extractall()

# Deleting unneccesary images; the ones whose corners are not detected and .mat files
cwd = os.getcwd()
[os.remove(file) for file in os.listdir(cwd) if not file.endswith(".jpg")]

# change the file names for calibration images to match the name of cameras in config.yaml file.i.e. camera-1 and camera-2
cam1_images = glob.glob(os.path.join(cwd, "left*.jpg"))
cam2_images = glob.glob(os.path.join(cwd, "right*.jpg"))
# Sorting images
cam1_images.sort(key=lambda f: int("".join(filter(str.isdigit, f))))
cam2_images.sort(key=lambda f: int("".join(filter(str.isdigit, f))))
for idx, name in enumerate(cam1_images):
    os.rename(
        name, os.path.join(cwd, str("camera-1_" + "{0:0=2d}".format(idx + 1) + ".jpg"))
    )

for idx, name in enumerate(cam2_images):
    os.rename(
        name, os.path.join(cwd, str("camera-2_" + "{0:0=2d}".format(idx + 1) + ".jpg"))
    )

# Removing some of the images where the corner was not detected
[os.remove(file) for file in glob.glob(os.path.join(cwd, "*06.jpg"))]
[os.remove(file) for file in glob.glob(os.path.join(cwd, "*01.jpg"))]

print("CALIBRATING THE CAMERAS")
deeplabcut.calibrate_cameras(path_config_file, calibrate=True)

print("CHECKING FOR UNDISTORTION")
deeplabcut.check_undistortion(path_config_file)

print("TRIANGULATING")
video_dir = os.path.join(basepath, folder)
deeplabcut.triangulate(path_config_file, video_dir, save_as_csv=True)


print("CREATING LABELED VIDEO 3-D")
deeplabcut.create_labeled_video_3d(path_config_file, [video_dir], start=5, end=10)

# output_path = [os.path.join(basepath,folder)]
# deeplabcut.create_labeled_video_3d(path_config_file,output_path,start=5,end=10)

print("ALL DONE!!! - default 3D cases are functional.")