Skip to content

Use logging framework for output #12

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
timogilvie wants to merge 2 commits into
base: master
Choose a base branch
from
Open

Use logging framework for output #12

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
fd5efe4
First pass using logging framework to output champion positions to ou…
Feb 2, 2018
2e7804b
Terminal handler added
Feb 2, 2018
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
54 changes: 43 additions & 11 deletions YAD2K/test_deep_league.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,8 @@
import os
from subprocess import call
import random
import logging
import sys

import numpy as np
from keras import backend as K
Expand Down Expand Up @@ -64,6 +66,12 @@
help='to help avoid bad predictions, tell DeepLeague the 10 champions in the game of the VOD you are passing',
default= "")

parser.add_argument(
'-log',
'--log_path',
type=str,
help='path to logger filename. Default=stdout')

subparsers = parser.add_subparsers(dest='subcommand')

youtube_option = subparsers.add_parser('youtube')
Expand Down Expand Up @@ -143,9 +151,30 @@
start_time = os.path.expanduser(args.start_time)
end_time = os.path.expanduser(args.end_time)

# Using logging framework for champ position messages
logger = logging.getLogger()
logger.setLevel(logging.DEBUG)
formatter = logging.Formatter('%(levelname)s,%(message)s')
terminal = logging.StreamHandler(sys.stdout)
terminal.setFormatter(formatter)
if args.log_path is not None:
# If a log file is specified, send DEBUG messages to file and INFO to stdout
logfile = logging.FileHandler(filename=os.path.expanduser(args.log_path))
logfile.setFormatter(formatter)
logfile.setLevel(logging.DEBUG)
terminal.setLevel(logging.INFO)
logger.addHandler(logfile)
logger.addHandler(terminal)
# Write header row
logger.debug('Type,Frame filename,Champion,Confidence,X Coordinate,Y Coordinate')
else:
# Otherwise, send it all to stdout
terminal.setLevel(logging.DEBUG)
logger.addHandler(terminal)


if not os.path.exists(output_path):
print('Creating output path {}'.format(output_path))
logger.info('Creating output path {}'.format(output_path))
os.mkdir(output_path)

sess = K.get_session() # TODO: Remove dependence on Tensorflow session.
Expand All @@ -172,7 +201,7 @@
'Mismatch between model and given anchor and class sizes. ' \
'Specify matching anchors and classes with --anchors_path and ' \
'--classes_path flags.'
print('{} model, anchors, and classes loaded.'.format(model_path))
logger.info('{} model, anchors, and classes loaded.'.format(model_path))

# Check if model is fully convolutional, assuming channel last order.
model_image_size = yolo_model.layers[0].input_shape[1:3]
Expand Down Expand Up @@ -211,7 +240,7 @@ def test_yolo(image, image_file_name):
image.height - (image.height % 32))
resized_image = image.resize(new_image_size, Image.BICUBIC)
image_data = np.array(resized_image, dtype='float32')
print(image_data.shape)
logger.info(image_data.shape)

image_data /= 255.
image_data = np.expand_dims(image_data, 0) # Add batch dimension.
Expand All @@ -224,7 +253,7 @@ def test_yolo(image, image_file_name):
K.learning_phase(): 0
})

print('Found {} boxes for {}'.format(len(out_boxes), image_file_name))
logger.info('Found {} boxes for {}'.format(len(out_boxes), image_file_name))

font = ImageFont.truetype(
font='font/FiraMono-Medium.otf',
Expand All @@ -249,7 +278,10 @@ def test_yolo(image, image_file_name):
left = max(0, np.floor(left + 0.5).astype('int32'))
bottom = min(image.size[1], np.floor(bottom + 0.5).astype('int32'))
right = min(image.size[0], np.floor(right + 0.5).astype('int32'))
print(label, (left, top), (right, bottom))

x_coord = int((left + right) / 2)
y_coord = int((top + bottom) / 2)
logger.debug("{},{},{},{}".format(image_file_name, label, x_coord, y_coord))

if top - label_size[1] >= 0:
text_origin = np.array([left, top - label_size[1]])
Expand All @@ -271,8 +303,8 @@ def test_yolo(image, image_file_name):

def process_mp4(test_mp4_vod_path):
video = cv2.VideoCapture(test_mp4_vod_path)
print("Opened ", test_mp4_vod_path)
print("Processing MP4 frame by frame")
logger.info("Opened {}".format(test_mp4_vod_path))
logger.info("Processing MP4 frame by frame")

# forward over to the frames you want to start reading from.
# manually set this, fps * time in seconds you wanna start from
Expand All @@ -283,14 +315,14 @@ def process_mp4(test_mp4_vod_path):
success = True
fps = int(video.get(cv2.CAP_PROP_FPS))
total_frame_count = int(video.get(cv2.CAP_PROP_FRAME_COUNT))
print("Loading video %d seconds long with FPS %d and total frame count %d " % (total_frame_count/fps, fps, total_frame_count))
logger.info("Loading video %d seconds long with FPS %d and total frame count %d " % (total_frame_count/fps, fps, total_frame_count))

while success:
success, frame = video.read()
if not success:
break
if count % 1000 == 0:
print("Currently at frame ", count)
logger.info("Currently at frame {}".format(count))

# i save once every fps, which comes out to 1 frames per second.
# i think anymore than 2 FPS leads to to much repeat data.
Expand Down Expand Up @@ -335,10 +367,10 @@ def _main():
with youtube_dl.YoutubeDL(ydl_opts) as ydl:
ydl.download([test_youtube_link])
# TODO test this on other systems.
print("Calling ffmpeg to cut up video")
logger.info("Calling ffmpeg to cut up video")
call(['ffmpeg', '-i', youtube_download_path + '/vod_full.mp4', '-ss', start_time, '-to', end_time, '-c', 'copy', youtube_download_path + '/vod.mp4'])
os.remove(youtube_download_path + '/vod_full.mp4')
print("Done with ffmpeg")
logger.info("Done with ffmpeg")
process_mp4(youtube_download_path + '/vod.mp4')

sess.close()
Expand Down