feat: add audioclip into the flow

[nan-wang]

todo:

• [x] Add AudioCLIP
• [x] Fix the AudioCLIP Hub executors https://github.com/jina-ai/executors/pull/266
• [x] Implement a Ranker to mix the results from both image and audio sources
• [x] Write the README

• [x] Segment audio tracks into chunks

As for the data, I suggest to use the following, please don't add these two data into the repo. They are quite large and it is better to write a script to download them

# video of different animal sounds with pictures
youtube-dl --recode-video mp4 -o 0GhwfyDcAoM.mp4 https://www.youtube.com/watch\?v\=0GhwfyDcAoM

# video of different housing sounds with pictures
youtube-dl --recode-video mp4 -o uTNL3WXEexA.mp4 https://www.youtube.com/watch\?v\=uTNL3WXEexA

[jakobkruse1]

The example is now running on my machine. Another tester would be nice to ensure that it is working on multiple machines. I suggest some follow-up tickets after merging this one:

• Add timestamps to the audio matches (right now it is not possible)
• Better dataset might be useful

[jakobkruse1]

For some reason, I can not add @nan-wang as reviewer.

[nan-wang]

Another thing is that I find the score from the audios is always lower than that from the images.