def test_get_episode(self):
''' Tests getting episode from video's field. '''
configure_for_unittest()
from model import Video, Show, Season, Episode
show = Show(title='House of Cards').save()
season1 = Season(show=show, number=1).save()
season2 = Season(show=show, number=2).save()
episode1 = Episode(season=season1,
number=1,
release_date=dt(2010, 1, 1)).save()
episode2 = Episode(season=season1, number=2).save()
episode3 = Episode(season=season2, number=1).save()
video = Video(link='vk.com').save()
show.seasons.connect(season1)
show.seasons.connect(season2)
season1.episodes.connect(episode1)
season1.episodes.connect(episode2)
season2.episodes.connect(episode3)
episode1.videos.connect(video)
video.refresh()
self.assertEqual(video.episode.get().number, 1)
def main():
with TransactionManager.transaction():
User.init_table()
Video.init_table()
Config.init_table()
print("Succeeded to init db.")
return 0
def extract_video_tweets(twapi_search_response: dict) -> List[Video]:
video_list = []
for tweet in twapi_search_response:
selected_video = select_video_url_from_tweet(tweet)
if selected_video is None:
continue
else:
video_url = selected_video[0]
video_type = selected_video[1]
if "retweeted_status" in tweet or "quoted_status" in tweet:
if "retweeted_status" in tweet:
video_source_tweet = tweet["retweeted_status"]
elif "quoted_status" in tweet:
video_source_tweet = tweet["quoted_status"]
author = User(
user_id=video_source_tweet["user"]["id"],
name=video_source_tweet["user"]["name"],
screen_name=video_source_tweet["user"]["screen_name"],
thumbnail_url=video_source_tweet["user"]["profile_image_url_https"])
retweeter = User(
user_id=tweet["user"]["id"],
name=tweet["user"]["name"],
screen_name=tweet["user"]["screen_name"],
thumbnail_url=tweet["user"]["profile_image_url_https"])
vid = Video(
tweet_id=tweet["id"],
author=author,
retweeted_user=retweeter,
body=video_source_tweet["full_text"],
created_at=twitter_date_to_datetime(tweet["created_at"]),
video_url=video_url,
video_type=video_type)
else:
author = User(
user_id=tweet["user"]["id"],
name=tweet["user"]["name"],
screen_name=tweet["user"]["screen_name"],
thumbnail_url=tweet["user"]["profile_image_url_https"])
vid = Video(
tweet_id=tweet["id"],
author=author,
retweeted_user=None,
body=tweet["full_text"],
created_at=twitter_date_to_datetime(tweet["created_at"]),
video_url=video_url,
video_type=video_type)
video_list.append(vid)
return video_list
def test_create_duplicates(self):
''' Tests raise exception with trying to create duplicates. '''
configure_for_unittest()
from model import Video
from neomodel import UniqueProperty
video1 = Video(link='https://vk.com/blablabla').save()
video2 = Video(link='vk.com/blablabla')
with self.assertRaises(UniqueProperty):
video2.save()
def test_create_duplicates(self):
''' Tests raise exception with trying to create duplicates. '''
configure_for_unittest()
from model import Video
from neomodel import UniqueProperty
video1 = Video(link='https://vk.com/blablabla').save()
video2 = Video(link='vk.com/blablabla')
with self.assertRaises(UniqueProperty):
video2.save()
def stabilization(optical_flow_method, debug: bool = False, **kwargs):
"""
Perform video stabilization using the given optical flow method.
Idea: test some metric using a known logo. Using ORB matching we could detect if it moves.
:param optical_flow_method: the optical flow method to use
:param debug: whether to show debug plots
"""
video = Video('../datasets/stabilization/piano')
feature_params = dict(maxCorners=500,
qualityLevel=0.3,
minDistance=7,
blockSize=7)
previous_frame = None
accum_flow = np.zeros(2)
count = 0
for i, frame in tqdm(enumerate(video.get_frames()),
total=len(video),
file=sys.stdout):
rows, cols, _ = frame.shape
if previous_frame is not None:
if i % 4 == 0:
p0 = cv2.goodFeaturesToTrack(cv2.cvtColor(
previous_frame, cv2.COLOR_BGR2GRAY),
mask=None,
**feature_params)
flow = optical_flow_method(previous_frame, frame, p0)
if debug:
show_optical_flow_arrows(previous_frame, flow)
m = np.mean(flow[np.logical_or(flow[:, :, 0] != 0,
flow[:, :, 1] != 0)],
axis=(0, 1))
if not np.isnan(accum_flow).any():
accum_flow += -m
transform = np.float32([[1, 0, accum_flow[0]],
[0, 1, accum_flow[1]]])
frame2 = cv2.warpAffine(frame, transform, (cols, rows))
if debug:
plt.figure()
plt.imshow(cv2.cvtColor(frame2, cv2.COLOR_BGR2RGB))
plt.axis('off')
plt.show()
cv2.imwrite("../video/block/OrigianlFrame%04d.jpg" % count,
frame) # save frame as JPEG file
cv2.imwrite("../video/block/StabilizedFrame%04d.jpg" % count,
frame2) # save frame as JPEG file
count += 1
previous_frame = frame
def upload():
if 'video' not in request.files:
return json_error("Video file not found")
video_file = request.files['video']
filename = secure_filename(video_file.filename)
ext = filename.split('.')[-1]
new_filename = generate_random_string() + '.' + ext
video_file.save(os.path.join(app.config['UPLOAD_DIR'], new_filename))
video = Video(filename=new_filename)
video.save()
return jsonify(video.to_dict())
def gaussian_model(video: Video,
frame_start: int,
background_mean: np.ndarray,
background_std: np.ndarray,
alpha: float = 2.5,
pixel_value: PixelValue = PixelValue.GRAY,
total_frames: int = None,
disable_tqdm=False) -> Iterator[np.ndarray]:
for im in tqdm(video.get_frames(frame_start),
total=total_frames,
file=sys.stdout,
desc="Non-adaptive gaussian model...",
disable=disable_tqdm):
if pixel_value == PixelValue.GRAY:
im_values = np.mean(im, axis=-1) / 255
elif PixelValue.HSV:
im_values = cv2.cvtColor(im, cv2.COLOR_BGR2HSV)[:, :, 0] / 180
else:
raise Exception
mask = (np.abs(im_values) - background_mean) >= (alpha *
(background_std +
(5 / 255)))
yield im, mask.astype(np.uint8) * 255
def get_video(vid):
video = Video.get_or_none(id=vid)
if video is None:
return json_error("Video not found", 404)
return jsonify(video.to_dict())
def gaussian_model_adaptive(video: Video,
train_stop_frame: int,
background_mean: np.ndarray,
background_std: np.ndarray,
alpha: float = 2.5,
rho: float = 0.1,
pixel_value: PixelValue = PixelValue.GRAY,
total_frames: int = None,
disable_tqdm=False) -> Iterator[np.ndarray]:
for im in tqdm(video.get_frames(train_stop_frame),
total=total_frames,
file=sys.stdout,
desc='Adaptive gaussian model...',
disable=disable_tqdm):
if pixel_value == PixelValue.GRAY:
im_values = np.mean(im, axis=-1) / 255
elif PixelValue.HSV:
im_values = cv2.cvtColor(im, cv2.COLOR_BGR2HSV)[:, :, 0] / 180
else:
raise Exception
mask = (np.abs(im_values) -
background_mean) >= (alpha * (background_std + 5 / 255))
background_mean = rho * im_values + (1 - rho) * background_mean
background_std = np.sqrt(rho *
np.power((im_values - background_mean), 2) +
(1 - rho) * np.power(background_std, 2))
yield im, mask.astype(np.uint8) * 255
def get_background_model(video: Video,
train_stop_frame: int,
total_frames: int = None,
pixel_value: PixelValue = PixelValue.GRAY,
disable_tqdm=False) -> (np.ndarray, np.ndarray):
background_list = None
i = 0
for im in tqdm(video.get_frames(0, train_stop_frame),
total=total_frames,
file=sys.stdout,
desc='Training model...',
disable=disable_tqdm):
if background_list is None:
background_list = np.zeros(
(im.shape[0], im.shape[1], train_stop_frame), dtype=np.int16)
if pixel_value == PixelValue.GRAY:
background_list[:, :, i] = np.mean(im, axis=-1)
elif PixelValue.HSV:
background_list[:, :, i] = cv2.cvtColor(im,
cv2.COLOR_BGR2HSV)[:, :, 0]
else:
raise Exception
i += 1
if pixel_value == PixelValue.GRAY:
background_mean = np.mean(background_list, axis=-1) / 255
background_std = np.std(background_list, axis=-1) / 255
elif PixelValue.HSV:
background_mean = np.mean(background_list, axis=-1) / 180
background_std = np.std(background_list, axis=-1) / 180
else:
raise Exception
return background_mean, background_std
def import_from_lines(self, lines):
"""
从多行数据录入数据库
:param lines: 待处理数据
:return: 读取到的记录数量,成功录入的记录数量
"""
(total, succ) = (0, 0)
for line in lines:
total += 1
items = line.strip().split('\t')
# 如果所读取的文件编码不为 utf-8,则需要 line.encode('<the_coding_of_str>').decode('utf-8').strip().split('\t')
if len(items) < 7:
CrawlerLogger.logger.info(
'line format error: {0}'.format(line))
continue
tmp_video = Video(0, items[0], items[1], items[2], items[3],
items[4], items[5], items[6])
if not self.insert_one_video(tmp_video):
CrawlerLogger.logger.info(
'insert line failed: {0}'.format(line))
else:
succ += 1
return total, succ
def put_video_in_queue():
if 'video_id' not in request.json:
return json_error("Video not found")
if 'qualities' not in request.json:
return json_error("Target qualities not found")
qualities = request.json['qualities']
video_id = request.json['video_id']
video = Video.get_or_none(id=video_id)
if video is None:
return json_error("Video not found")
video_queue = WaitingQueue.get_or_none(WaitingQueue.video == video)
if video_queue is not None:
return json_error("Already in queue")
video_queue = WaitingQueue(video=video)
video_queue.save()
success = save_video_qualities(video, qualities)
if not success:
video_queue.delete_instance()
return jsonify(video.to_dict())
def add(user_id, video_id):
#Taking the old video ti present it to the user and also to mention the video last publisher to post it
#Interactivly
user = session.query(User).filter_by(id = user_id).first()
curr_video = session.query(Video).filter_by(id = video_id).first()
if request.method == 'GET':
return render_template('add.html', user_id = user_id, video_id = video_id)
#Adding the video with refrence to the old one
else:
video = request.form.get('video')
description = request.form.get('description')
owner = user.id
curr_video.other_video = video
publish = False
new_vid = Video(video = video, description = description, publish = publish, owner = owner)
session.add(new_vid)
session.commit()
return redirect(url_for('homepage',user_id = user_id))
def load_videos():
"""Load videos from seed data into database"""
with open("seed_data/videos.txt") as videos:
for row in videos:
video = row.rstrip().split("|")
youtube = True if video[4] == "True" else False
hidden = True if video[5] == "True" else False
kwargs = dict(
video_id = video[0],
user_id = video[1],
video = video[2],
category = video[3],
youtube = youtube,
hidden = hidden
)
keys_to_remove = []
for key in kwargs.keys():
if kwargs[key] == "":
keys_to_remove.append(key)
for key in keys_to_remove:
del kwargs[key]
video = Video(**kwargs)
db.session.add(video)
db.session.commit()
def openFile(self, filename):
filetype = filename.split('.')[-1]
self.generateButton.setText("Loading...")
self.infoLabel.setText("File opened: " + filename)
print("Opening " + filename)
# start the video player or start a worker thread depending on file type
if filetype in self.videofiletypes:
self.currentFile = Video(filepath=filename, colortype="rgb")
self.generateButton.setText("Generate images from video")
self.generateButton.setDisabled(False)
self.startVideoPlayer()
elif filetype in self.metadatatypes:
self.metafilename = filename
self.startWorker(ip.loadCsv, self.setCurrentFrames,
self.fillGallery, self.metafilename,
self.currentFrames)
elif filetype == 'pkl':
self.currentFile = filename
self.startWorker(ip.loadPickle, self.setCurrentFrames,
self.fillGallery, self.currentFile)
elif filetype == 'h5':
self.currentFile = filename
self.startWorker(ip.loadhdf5, self.setCurrentFrames,
self.fillGallery, self.currentFile,
self.currentFrames)
else:
# invalid file type selected
self.showWarning('FileType')
def main():
alpha_values = np.linspace(1.5, 3, 20)
rho_values = np.logspace(-2, -0.1, 20)
# Ensure cache
video = Video("../datasets/AICity_data/train/S03/c010/frames")
get_background_model(video,
int(2141 * 0.25),
total_frames=int(2141 * 0.25),
disable_tqdm=False)
# Best alpha: 1.75
# mAP_list = Parallel(n_jobs=4)(delayed(w2_map_alpha)(alpha) for alpha in tqdm(alpha_values))
mAP_list = Parallel(n_jobs=3)(delayed(w2_map_alpha)(1.75, rho)
for rho in tqdm(rho_values))
"""mAP_list = [0.18656629994209614, 0.23257430508572247, 0.2333781161367368, 0.18301435406698566,
0.1773032336790726, 0.1762025561112319, 0.12792207792207794, 0.17066218427456575,
0.12438077386530996, 0.12091293755609694, 0.11872632575757576, 0.1189064558629776,
0.15132634758802985, 0.157589106928314, 0.26284443191338397, 0.39380709780347006,
0.43192630414348515, 0.357941584643725, 0.3186317361126976, 0.20596422790608496]"""
plt.figure()
plt.plot(rho_values, mAP_list)
plt.xlabel(r'$\rho$ threshold')
plt.ylabel('mAP')
plt.show()
def create_video(link, title, notes = ""):
"""Create and return a video."""
video = Video(link = link, title = title, notes = notes)
db.session.add(video)
db.session.commit()
return video
def get(self):
q = Video.all()
template_values = {
'list_videos':q.run(limit=10)
}
template = jinja_environment.get_template('/templates/video_list.html')
self.response.out.write(template.render(template_values))
def thead_list(play_list_id):
play = PlayList.get_by_id(play_list_id)
while True:
first_video = play.videos\
.where(Video.is_progress == 0) \
.where(Video.is_completed == 0) \
.first()
if first_video is None:
break
updated = Video.update(is_progress=1) \
.where(Video.is_progress == 0) \
.where(Video.id == first_video.id).execute()
if updated == 1:
client = Bilibili()
client.download_by_id(first_video.id)
Video.update(is_completed=True).where(
Video.id == first_video.id).execute()
def download_by_id(self, video_id):
video = Video.select().where(Video.id == video_id).first()
play = PlayList.select().where(
PlayList.id == video.play_list_id).first()
print("开始下载" + video.title + "，视频集：" + play.title)
download_path = Bilibili.download_path + '/' + play.title
if os.path.exists(download_path) is False:
os.mkdir(download_path)
self.get_url(video.cid, video.title, download_path)
def create_video(video_path, description, date_posted):
"""Create new video (info)."""
video = Video(video_path=video_path,
description=description,
date_posted=date_posted)
db.session.add(video)
db.session.commit()
return video
def videolist_since():
select_params = {}
if "since_id" in flask.request.args:
select_params["since_id"] = int(flask.request.args["since_id"])
if "count" in flask.request.args:
select_params["count"] = int(flask.request.args["count"])
videos = Video.select_since(**select_params)
videos_dict = [video.to_dict() for video in videos]
videos_dict = FilterPluginExecutor.get_instance()(videos_dict)
return flask.Response(response=json.dumps(videos_dict))
def upload_video():
"""Handle file uploads"""
# if it's a get - render registration form
if request.method == "GET":
"""Load form for user to upload new video"""
case_id = request.args.get('case_id')
return render_template('upload-video.html', case_id=case_id)
# otherwise register user in db
if request.method == "POST":
"""Uploads video to aws"""
print "\n\n\n\n\n\n\n", request, "\n\n\n\n\n\n"
video_file = request.files['media']
print "\n\n\n\n\n\n\n", video_file, "\n\n\n\n\n\n"
case_id = request.form.get('case_id')
# video_file = request.files.get("rawvid")
video_name = video_file.filename
try:
transcript_file = request.files.get("tscript")
except:
transcript_file = None
user_id = g.current_user.user_id
#get deponent name and recorded date
deponent = request.form.get('name')
recorded_at = request.form.get('date-taken')
# add the video to the db
date_added = datetime.now()
new_vid = Video(case_id=case_id,
vid_name=video_name,
added_by=user_id,
added_at=date_added,
deponent=deponent,
recorded_at=recorded_at)
db.session.add(new_vid)
db.session.commit()
if transcript_file:
script_text = transcript_file.readlines()
new_script = Transcript(vid_id=new_vid.vid_id, text=script_text)
db.session.add(new_script)
db.session.commit()
# send the upload to a separate thread to upload while the user moves on
upload = threading.Thread(target=upload_aws_db,
args=(video_file, video_name, case_id,
user_id, socketio)).start()
return jsonify(case_id)
def update_episode_urls(episode):
urls, new_urls = get_episode_urls(episode), []
for url in urls:
try:
v = Video(link=url).save()
episode.videos.connect(v)
new_urls.append(urls)
except neomodel.UniqueProperty:
pass
return new_urls
def create_video(channel_name, web_title, youtube_title):
video = Video(channel_name=channel_name,
web_title=web_title,
youtube_title=youtube_title)
db.session.add(video)
db.session.commit()
return video
def get_url(self, cid, filename, path):
"""获取每个链接并下载"""
html = self.get_response_by_cid(cid=cid)
print(html)
video_list = []
if len(html['durl']) == 1:
# 如果只有一个链接，则表示单视频
print(html['durl'][0])
Video.update(size=html['durl'][0]['size']).where(
Video.cid == cid).execute()
self.download(html['durl'][0]['url'],
path + '/' + filename + '.mp4', self.next_headers)
else:
# 否则是列表
temps = []
for i in html['durl']:
print(i)
exit()
temp = path + '/' + filename + '.tmp'
temps.append(temp)
self.download(i['url'], temp, self.next_headers)
return video_list
def w2_map_alpha(alpha):
video = Video("../datasets/AICity_data/train/S03/c010/frames")
frames = []
ious = []
for im, mask, frame in week2_nonadaptive(video, alpha, disable_tqdm=True):
frames.append(frame)
ious.append(frame.get_detection_iou_mean(ignore_classes=True))
mAP = mean_average_precision(frames)
print('alpha', alpha, 'mAP', mAP, 'mean IoU', np.mean(ious))
return mAP
def add_video():
"""add video to database"""
user_id = session["current_user"]
category = request.form.get("category")
youtube = request.form.get("youtube")
print(category)
print(youtube)
if youtube:
kwargs = dict(user_id=user_id,
video=youtube,
category=category,
youtube=True)
db.session.add(Video(**kwargs))
db.session.commit()
print("added youtube")
else:
video = functions.save_photo("video")
print(video)
kwargs = dict(user_id=user_id,
video=video,
category=category,
youtube=False)
db.session.add(Video(**kwargs))
db.session.commit()
print("added vid")
user = User.query.get(user_id)
for vid in user.videos:
print(vid.video, vid.category, vid.youtube)
return redirect("/users/{}/my_videos".format(user_id))
def create_video(video_title, video_duration, video_url, playlist_id):
"""Create and return a new video to a playlist"""
video = Video(
video_title=video_title,
video_duration=video_duration,
video_url=video_url,
playlist_id=playlist_id,
)
db.session.add(video)
db.session.commit()
return video
def post(self):
title = TString.trim( self.request.get('txtTitle') )
link = TString.trim( self.request.get('txtLink') )
priority = TString.trim( self.request.get('txtPriority') )
#self.response.out.write(title + ";" + link + ";" + priority)
m = re.search('(http:\/\/)?\w\w\w\.[a-z]+\.[a-z]*.*',link)
link = m.group(0).split("v=")[1]
vc = RegisterContext()
template = jinja_environment.get_template('/templates/video_form.html')
if priority.isdigit():
priority = int(priority)
vc.success_flag = ( priority >= 1 and priority<=5 )
else:
vc.success_flag = False
if vc.success_flag:
new_v = Video(vd_title=title,vd_link=link,vd_priority=int(priority) )
new_v.put()
self.response.out.write(template.render({'video_ctx':vc}))
def test_get_episode(self):
''' Tests getting episode from video's field. '''
configure_for_unittest()
from model import Video, Show, Season, Episode
show = Show(title='House of Cards').save()
season1 = Season(show=show, number=1).save()
season2 = Season(show=show, number=2).save()
episode1 = Episode(season=season1, number=1, release_date=dt(2010,1,1)).save()
episode2 = Episode(season=season1, number=2).save()
episode3 = Episode(season=season2, number=1).save()
video = Video(link='vk.com').save()
show.seasons.connect(season1)
show.seasons.connect(season2)
season1.episodes.connect(episode1)
season1.episodes.connect(episode2)
season2.episodes.connect(episode3)
episode1.videos.connect(video)
video.refresh()
self.assertEqual(video.episode.get().number, 1)
def play(play_id):
play_list = PlayList.select().where(PlayList.id == play_id).first()
videos = Video.select().where(Video.play_list_id == play_id)
video_dict = models_to_dict(videos)
client = Bilibili()
for video in video_dict:
file = Bilibili.download_path + '/' + play_list.title + "/" + video[
'title'] + '.mp4'
if video['size'] == 0:
html = client.get_response_by_cid(video['cid'])
if 'durl' in html.keys() and len(html['durl']) == 1:
# 如果只有一个链接，则表示单视频
print(html['durl'][0])
Video.update(size=html['durl'][0]['size']).where(
Video.cid == video['cid']).execute()
print(html)
if os.path.exists(file) is True:
video['file_size'] = os.path.getsize(file)
else:
video['file_size'] = 0
return render_template("play.html", play=play_list, videos=video_dict)
def fine_tune_yolo(debug=False):
video = Video("../datasets/AICity_data/train/S03/c010/frames")
detection_transform = DetectionTransform()
classes = utils.load_classes('../config/coco.names')
hyperparams = parse_model_config('../config/yolov3.cfg')[0]
learning_rate = float(hyperparams["learning_rate"])
momentum = float(hyperparams["momentum"])
decay = float(hyperparams["decay"])
burn_in = int(hyperparams["burn_in"])
model = Darknet('../config/yolov3.cfg')
print(model)
model.load_weights('../weights/yolov3.weights')
model.train()
for module_def, module in zip(model.module_defs, model.module_list):
if module_def["type"] == "yolo":
break
module.train(False)
if torch.cuda.is_available():
model = model.cuda()
optimizer = Adam(filter(lambda p: p.requires_grad, model.parameters()),
lr=1e-5)
gt = read_annotations(
'../datasets/AICity_data/train/S03/c010/m6-full_annotation.xml')
dataset = YoloDataset(video, gt, classes, transforms=detection_transform)
data_loader = DataLoader(dataset,
batch_size=16,
shuffle=True,
num_workers=4)
for epoch in tqdm(range(10), file=sys.stdout, desc='Fine tuning'):
for images, targets in tqdm(data_loader,
file=sys.stdout,
desc='Running epoch'):
if torch.cuda.is_available():
images = images.cuda()
targets = targets.cuda()
optimizer.zero_grad()
loss = model(images, targets)
loss.backward()
optimizer.step()
print('Training finished. Saving weights...')
model.save_weights('../weights/fine_tuned_yolo_freeze.weights')
print('Saved weights')
