Select video

author: Chen Zhang

Description

This operator select input list by simply aggregating, sorting and then filtering it.

Code Example

For two input lists: video urls and scores, aggregate url by reduce_function, then sort aggregated scores, then select top k results.

from towhee import pipe, ops, DataCollection

p = (
    pipe.input('video_urls', 'scores') \
        .map(('video_urls', 'scores'), 'res', ops.video_copy_detection.select_video(top_k=2, reduce_function='mean', reverse=True)) \
        .output('video_urls', 'scores', 'res')
)

DataCollection(p(['a', 'a', 'c', 'a', 'b', 'b', 'c', 'c'], [2, 8, 9.3, 5, 2, 1, 0, -1])).show()

from towhee import pipe, ops, DataCollection

p = (
    pipe.input('video_urls', 'scores') \
        .map(('video_urls', 'scores'), 'res', ops.video_copy_detection.select_video(top_k=2, reduce_function='sum', reverse=True)) \
        .output('video_urls', 'scores', 'res')
)

DataCollection(p(['a', 'a', 'c', 'a', 'b', 'b', 'c', 'c'], [2, 8, 9.3, 5, 2, 1, 0, -1])).show()

from towhee import pipe, ops, DataCollection

p = (
    pipe.input('video_urls', 'scores') \
        .map(('video_urls', 'scores'), 'res', ops.video_copy_detection.select_video(top_k=2, reduce_function='max', reverse=True)) \
        .output('video_urls', 'scores', 'res')
)

DataCollection(p(['a', 'a', 'c', 'a', 'b', 'b', 'c', 'c'], [2, 8, 9.3, 5, 2, 1, 0, -1])).show()

from towhee import pipe, ops, DataCollection

p = (
    pipe.input('video_urls', 'scores') \
        .map(('video_urls', 'scores'), 'res', ops.video_copy_detection.select_video(top_k=2, reduce_function='min', reverse=True)) \
        .output('video_urls', 'scores', 'res')
)

DataCollection(p(['a', 'a', 'c', 'a', 'b', 'b', 'c', 'c'], [2, 8, 9.3, 5, 2, 1, 0, -1])).show()

Factory Constructor

Create the operator via the following factory method

select_video(top_k: int, reduce_function: str, reverse: bool)

Parameters:

​ top_k: int

​ Select top k result.

​ reduce_function: str

​ Aggregate function name, support name:

• sum
• mean
• max
• min

​ reverse: bool

​ Whether sorted result is reversed.

Interface

Parameters:

​ video_urls: List[str]

​ Video urls.

​ scores: List[float]

​ Every video scores.

Returns: List[str]

​ Selected video url results.

More Resources

• Extracting Event Highlights Using iYUNDONG Sports App - Zilliz blog: iYUNDONG continually seeks to support its users with real-time image retrieval services. When it decided to develop its App, it chose Milvus to build its core image search system. Discover how Milvus, an open-source vector database, can help build an intelligent image retrieval system that can extract sport event highlights.
• 4 Steps to Building a Video Search System - Zilliz blog: Searching for videos by image with Milvus
• Vector Database Use Cases: Video Similarity Search - Zilliz: Experience a 10x performance boost and unparalleled precision when your video similarity search system is powered by Zilliz Cloud.