1
@huey.db_task()
2
def do_video_indexing(video_tasks):
3
    print(f'Creating tasks: {video_tasks}')
4
5
    # Create a task for each video we want to index
6
    for video_task in video_tasks:
7
        task = TWELVE_LABS_CLIENT.task.create(
8
            TWELVE_LABS_INDEX_ID,
9
            url=default_storage.url(video_task['video']),
10
            disable_video_stream=True
11
        )
12
        print(f'Created task: {task}')
13
        video_task['task_id'] = task.id
14
15
    print(f'Created {len(video_tasks)} tasks')

1
    print(f'Polling TwelveLabs for {video_tasks}')
2
3
    # Do a single database query for all the videos we're interested in
4
    video_ids = [video_task['id'] for video_task in video_tasks]
5
    videos = Video.objects.filter(id__in=video_ids)
6
7
    while True:
8
        done = True
9
        videos_to_save = []
10
11
        # Retrieve status for each task we created
12
        for video_task in video_tasks:
13
            # What's our current state for this video?
14
            video = videos.get(video__exact=video_task['video'])
15
16
            # Do we still need to retrieve status for this task?
17
            if video.status != 'Ready':
18
                task = TWELVE_LABS_CLIENT.task.retrieve(video_task['task_id'])
19
                if task.status != 'ready':
20
                    # We'll need to go round the loop again
21
                    done = False
22
23
                # Do we need to write a new status to the DB?
24
                if video.status.lower() != task.status:
25
                    # We store the status in the DB in title case, so it's ready to render on the page
26
                    new_status = task.status.title()
27
                    print(f'Updating status for {video_task["video"]} from {video.status} to {new_status}')
28
                    video.status = new_status
29
                    if task.status == 'ready':
30
                        video.video_id = task.video_id
31
                        get_all_video_data(video)
32
                    videos_to_save.append(video)
33
34
        if len(videos_to_save) > 0:
35
            Video.objects.bulk_update(videos_to_save, ['status', 'video_id', THUMBNAILS_PATH, TRANSCRIPTS_PATH, TEXT_PATH, LOGOS_PATH])
36
37
        if done:
38
            break
39
40
        sleep(TWELVE_LABS_POLL_INTERVAL)
41
42
    print(f'Done polling {video_tasks}')

1
def get_queryset(self):
2
    """
3
    Search TwelveLabs for videos matching the query
4
    """
5
    query = self.request.GET.get("query", None)
6
7
    result = TWELVE_LABS_CLIENT.search.query(
8
        TWELVE_LABS_INDEX_ID,
9
        query,
10
        ["visual", "conversation", "text_in_video", "logo"],
11
        group_by="video",
12
        threshold="medium"
13
    )
14
15
    # Search results may be in multiple pages, so we need to loop until we're done retrieving them
16
    search_data = result.data
17
    print(f"First page's data: {search_data}")
18
19
    search_results = []
20
    while True:
21
        # Do a database query to get the videos for each page of results
22
        video_ids = [group.id for group in search_data]
23
        videos = Video.objects.filter(video_id__in=video_ids)
24
        for group in search_data:
25
            try:
26
                search_results.append(SearchResult(video=videos.get(video_id__exact=group.id),
27
                                                   clip_count=len(group.clips),
28
                                                   clips=group.clips.model_dump_json()))
29
            except self.model.DoesNotExist:
30
                # There is a video in TwelveLabs, but no corresponding row in the database.
31
                # Just report it and carry on.
32
                print(f'Can\'t find match for video_id {group.id}')
33
34
        # Is there another page?
35
        try:
36
            search_data = next(result)
37
            print(f"Next page's data: {search_data}")
38
        except StopIteration:
39
            print("There is no next page in search result")
40
            break
41
42
    return search_results