def streaming_annotate(stream_file):
"""Annotate a local video file through streaming API."""
client = videointelligence.StreamingVideoIntelligenceServiceClient()
# Set the chunk size to 5MB (recommended less than 10MB).
chunk_size = 5 * 1024 * 1024
# Open file.
with open(stream_file) as video_file:
requests = (
types.StreamingAnnotateVideoRequest(input_content=chunk)
for chunk in stream(video_file, chunk_size))
# Set streaming config.
config = types.StreamingVideoConfig(
feature=enums.StreamingFeature.STREAMING_EXPLICIT_CONTENT_DETECTION)
config_request = types.StreamingAnnotateVideoRequest(video_config=config)
# streaming_annotate_video returns a generator.
# timeout argument specifies the maximum allowable time duration between
# the time that the last packet is sent to Google video intelligence API
# and the time that an annotation result is returned from the API.
# timeout argument is represented in number of seconds.
responses = client.streaming_annotate_video(
config_request, requests, timeout=10800)
print('\nReading response.')
# Retrieve results from the response generator.
for response in responses:
for frame in response.annotation_results.explicit_annotation.frames:
likelihood = enums.Likelihood(frame.pornography_likelihood)
frame_time = frame.time_offset.seconds + frame.time_offset.nanos / 1e9
print('Time: {}s'.format(frame_time))
print('\tpornography: {}'.format(likelihood.name))
def streaming_annotate(stream_file):
"""Annotate a local video file through streaming API."""
client = videointelligence.StreamingVideoIntelligenceServiceClient()
# Set the chunk size to 5MB (recommended less than 10MB).
chunk_size = 5 * 1024 * 1024
# Open file.
with open(stream_file) as video_file:
requests = (types.StreamingAnnotateVideoRequest(input_content=chunk)
for chunk in stream(video_file, chunk_size))
# Set streaming config.
config = types.StreamingVideoConfig(
feature=enums.StreamingFeature.STREAMING_OBJECT_TRACKING)
config_request = types.StreamingAnnotateVideoRequest(video_config=config)
# streaming_annotate_video returns a generator.
# timeout argument specifies the maximum allowable time duration between
# the time that the last packet is sent to Google video intelligence API
# and the time that an annotation result is returned from the API.
# timeout argument is represented in number of seconds.
responses = client.streaming_annotate_video(config_request,
requests,
timeout=10800)
print '\nReading response.'
# Retrieve results from the response generator.
for response in responses:
object_annotations = response.annotation_results.object_annotations
# When object_annotations is empty, no object is found.
if object_annotations:
for annotation in object_annotations:
description = annotation.entity.description
confidence = annotation.confidence
track_id = annotation.track_id
print 'Entity description: {}'.format(description)
print 'Track Id: {}'.format(track_id)
if annotation.entity.entity_id:
print 'Entity id: {}'.format(annotation.entity.entity_id)
print 'Confidence: {}'.format(confidence)
# In streaming mode, len(annotation.frames) is always 1, and the frames
# in the same response share the same time_offset.
frame = annotation.frames[0]
box = frame.normalized_bounding_box
print('Time: {}s'.format(frame.time_offset.seconds +
frame.time_offset.nanos / 1e9))
print 'Bounding box position:'
print '\tleft : {}'.format(box.left)
print '\ttop : {}'.format(box.top)
print '\tright : {}'.format(box.right)
print '\tbottom: {}'.format(box.bottom)
print '\n'
def streaming_annotate(stream_file):
"""Annotate a local video file through streaming API."""
client = videointelligence.StreamingVideoIntelligenceServiceClient()
# Set the chunk size to 5MB (recommended less than 10MB).
chunk_size = 5 * 1024 * 1024
# Open file.
with open(stream_file) as video_file:
requests = (types.StreamingAnnotateVideoRequest(input_content=chunk)
for chunk in stream(video_file, chunk_size))
# Set streaming config with stationary_camera option.
# stationary_camera flag can be set to False (default option) or True.
label_config = types.StreamingLabelDetectionConfig(
stationary_camera=False)
config = types.StreamingVideoConfig(
feature=enums.StreamingFeature.STREAMING_LABEL_DETECTION,
label_detection_config=label_config)
config_request = types.StreamingAnnotateVideoRequest(
video_config=config)
# streaming_annotate_video returns a generator.
# timeout argument specifies the maximum allowable time duration between
# the time that the last packet is sent to Google video intelligence API
# and the time that an annotation result is returned from the API.
# timeout argument is represented in number of seconds.
responses = client.streaming_annotate_video(config_request,
requests,
timeout=10800)
print '\nReading response.'
# Retrieve results from the response generator.
for response in responses:
for annotation in response.annotation_results.label_annotations:
description = annotation.entity.description
# The response of steaming_annotate_video has only one frame for each
# annotation.
time_offset = annotation.frames[0].time_offset.seconds + \
annotation.frames[0].time_offset.nanos / 1e9
confidence = annotation.frames[0].confidence
print '{}s: {}\t ({})'.format(
time_offset,
description.encode('utf-8').strip(), confidence)
def streaming_annotate(stream_file, output_uri):
"""Annotate a local video file through streaming API."""
client = videointelligence.StreamingVideoIntelligenceServiceClient()
# Set the chunk size to 5MB (recommended less than 10MB).
chunk_size = 5 * 1024 * 1024
# Open file.
with open(stream_file) as video_file:
requests = (types.StreamingAnnotateVideoRequest(input_content=chunk)
for chunk in stream(video_file, chunk_size))
# Use storage config option in the config request to enable storage.
storage_config = types.StreamingStorageConfig(
enable_storage_annotation_result=True,
annotation_result_storage_directory=output_uri)
# LABEL_DETECTION feature is used as an example. Storage works for all
# supported features.
label_config = types.StreamingLabelDetectionConfig()
config = types.StreamingVideoConfig(
feature=enums.StreamingFeature.STREAMING_LABEL_DETECTION,
label_detection_config=label_config,
storage_config=storage_config)
config_request = types.StreamingAnnotateVideoRequest(
video_config=config)
# streaming_annotate_video returns a generator.
# timeout argument specifies the maximum allowable time duration between
# the time that the last packet is sent to Google video intelligence API
# and the time that an annotation result is returned from the API.
# timeout argument is represented in number of seconds.
responses = client.streaming_annotate_video(config_request,
requests,
timeout=10800)
print('\nReading response.')
# Retrieve results from the response generator.
for response in responses:
print('Storage uri: {}'.format(response.annotation_results_uri))
def streaming_annotate(stream_file):
"""Annotate a local video file through streaming API."""
client = videointelligence.StreamingVideoIntelligenceServiceClient()
# Set the chunk size to 5MB (recommended less than 10MB).
chunk_size = 5 * 1024 * 1024
# Open file.
with open(stream_file) as video_file:
requests = (types.StreamingAnnotateVideoRequest(input_content=chunk)
for chunk in stream(video_file, chunk_size))
# Set streaming config.
config = types.StreamingVideoConfig(
feature=enums.StreamingFeature.STREAMING_SHOT_CHANGE_DETECTION)
config_request = types.StreamingAnnotateVideoRequest(
video_config=config)
# streaming_annotate_video returns a generator.
# timeout argument specifies the maximum allowable time duration between
# the time that the last packet is sent to Google video intelligence API
# and the time that an annotation result is returned from the API.
# timeout argument is represented in number of seconds.
responses = client.streaming_annotate_video(config_request,
requests,
timeout=10800)
print('\nReading response.')
# Retrieve results from the response generator.
for response in responses:
for annotation in response.annotation_results.shot_annotations:
print('Shot: {}s to {}s'.format(
annotation.start_time_offset.seconds +
annotation.start_time_offset.nanos / 1e9,
annotation.end_time_offset.seconds +
annotation.end_time_offset.nanos / 1e9))
def streaming_annotate(stream_file, out_file):
"""Annotate a local video file through streaming API."""
client = videointelligence.StreamingVideoIntelligenceServiceClient()
# Set the chunk size to 5MB (recommended less than 10MB).
chunk_size = 5 * 1024 * 1024
# Open file.
with open(stream_file) as video_file:
requests = (types.StreamingAnnotateVideoRequest(input_content=chunk)
for chunk in stream(video_file, chunk_size))
# Set streaming config.
config = types.StreamingVideoConfig(
feature=enums.StreamingFeature.STREAMING_OBJECT_TRACKING)
config_request = types.StreamingAnnotateVideoRequest(
video_config=config)
# streaming_annotate_video returns a generator.
# timeout argument specifies the maximum allowable time duration between
# the time that the last packet is sent to Google video intelligence API
# and the time that an annotation result is returned from the API.
# timeout argument is represented in number of seconds.
responses = client.streaming_annotate_video(config_request,
requests,
timeout=10800)
print('[ANALYSIS] Reading response.')
with open(out_file, "w", 0) as out_handle:
print("[ANALYSIS] Opened output analysis file at {}".format(
out_file))
# Retrieve results from the response generator.
for response in responses:
object_annotations = response.annotation_results.object_annotations
# When object_annotations is empty, no object is found.
if object_annotations:
to_write = []
for annotation in object_annotations:
description = annotation.entity.description
confidence = annotation.confidence
track_id = annotation.track_id
# print('[ANALYSIS] Entity description: {}'.format(description))
# print('[ANALYSIS] Track Id: {}'.format(track_id))
# if annotation.entity.entity_id:
# print('[ANALYSIS] Entity id: {}'.format(annotation.entity.entity_id))
# print('[ANALYSIS] Confidence: {}'.format(confidence))
# In streaming mode, len(annotation.frames) is always 1, and the frames
# in the same response share the same time_offset.
frame = annotation.frames[0]
box = frame.normalized_bounding_box
time = frame.time_offset.seconds + frame.time_offset.nanos / 1e9
# print('[ANALYSIS] Time: {}s'.format(time))
# print('[ANALYSIS] Bounding box position:')
# print('[ANALYSIS] \tleft : {}'.format(box.left))
# print('[ANALYSIS] \ttop : {}'.format(box.top))
# print('[ANALYSIS] \tright : {}'.format(box.right))
# print('[ANALYSIS] \tbottom: {}'.format(box.bottom))
# print('\n')
cols = [
description,
track_id,
confidence,
time,
box.left,
box.top,
box.right,
box.bottom,
]
line_to_write = ",".join(str(col)
for col in cols) + "\n"
# print("[ANALYSIS] line: ", line_to_write)
to_write.append(line_to_write)
# print("[ANALYSIS] Writing labels!")
out_handle.writelines(to_write)
out_handle.flush()
