fix(videointelligence): revert #9440; make features a keyword parameter

[busunkim96]

Creating this for a rollback.

Python sample code code was passing the keyword argument input_uri positionally before this change. Adding a required parameter changed the parameter order and broke existing code.

from google.cloud import videointelligence
video_client = videointelligence.VideoIntelligenceServiceClient()
features = [videointelligence.enums.Feature.LABEL_DETECTION]
operation = video_client.annotate_video(
'gs://cloud-samples-data/video/cat.mp4', features=features)
print('\nProcessing video for label annotations:')
But I have a problem with the video annotation function, I get the following error message:
TypeError Traceback (most recent call last)
<ipython-input-24-5eb4216acaf7> in <module>
4 features = [videointelligence.enums.Feature.LABEL_DETECTION]
5 operation = video_client.annotate_video(
----> 6 'gs://cloud-samples-data/video/cat.mp4', features=features)
7 print('\nProcessing video for label annotations:')
TypeError: annotate_video() got multiple values for argument 'features'

Note: We're still debating what to do about this longer term. I may add a synth replace to make this semi-permanent until the transition to the microgenerator. Alternatively we may take this change but do a major version bump to indicate that it is breaking.