Retrieve thumbnails

To help users quickly preview videos, the platform allows you to retrieve thumbnails of your videos that you can display in your application.

You can retrieve thumbnails when:

  • Your application displays previews for one or more videos. For example, users may want to see thumbnails of all the videos in an index before deciding which video they wish to delete.
  • Your application displays previews for a set of search results. For example, when users perform a search, your application could use thumbnails to give users a small view of each matching video clip.

Note the following about retrieving thumbnails:

  • By default, thumbnail generation is disabled. Before you can retrieve thumbnails, you must enable this feature on a per-index basis when creating a new index.
  • The platform returns a unique URL for each thumbnail, and the URLs expire in one hour.
  • You cannot disable this feature once an index has been created.

The steps for retrieving thumbnails are as follows:

  1. Enable thumbnail generation for a new index.
  2. Upload videos to your index.
  3. (Optional) To extract thumbnails from a specific video, you must retrieve the unique identifier of that video. Otherwise, you can skip this step.
  4. Depending on your use case, retrieve thumbnails from a specific video or a set of search results.

Prerequisites

  • You’re familiar with the concepts that are described on the Platform overview page.
  • Your video must meet the following requirements:
    • Video resolution: must be greater or equal than 360p and less than 1080p (FHD)
    • Duration: must be between 10 seconds and 2 hours (7,200s)
      If you require different options, send us an email at [email protected].

📘

Note

For consistent results, Twelve Labs recommends you upload 360p videos.

Procedure

Follow the steps in the sections below to retrieve thumbnails.

1. Enable the thumbnail generation feature

By default, the thumbnail generation feature is disabled, and you can enable it on a per-index basis when creating a new index. To do so, call the POST method of the /indexes endpoint, ensuring that the addons array contains the thumbnail value.

In the example code below, replace the placeholder surrounded by <> with the name you want to give to your index.

INDEX_NAME = "<YOUR_INDEX_NAME>" # Use a descriptive name for your index 
INDEXES_URL = f"{API_URL}/indexes"
data = {
    "engines": [
      {
        "engine_name": "marengo2.5",
        "engine_options":  ["visual", "conversation", "text_in_video", "logo"],
      }
    ],
    "index_name": INDEX_NAME,
    "addons": ["thumbnail"] # Enable thumbnail generation
}
response = requests.post(INDEXES_URL, headers=headers, json=data)
INDEX_ID = response.json().get("_id")
print (f"Status code: {response.status_code}")
pprint (response.json())
const INDEX_NAME = '<YOUR_INDEX_NAME>' // Use a descriptive name for your index
const INDEXES_URL = `${API_URL}/indexes`
let data = JSON.stringify({
  engines: [
    {
      engine_name: "marengo2.5",
      engine_options: ["visual", "conversation", "text_in_video"]
    }
  ],
  index_name: INDEX_NAME,
  addons: ["thumbnail"], // Enable thumbnail generation
})
let config = {
  method: 'post',
  url: INDEXES_URL,
  headers: headers,
  data: data
}
let resp = await axios(config)
let response = await resp.data
const INDEX_ID = response._id
console.log(`Status code: ${resp.status}`)
console.log(response)

The output should look similar to the following one:

Status code: 200
{
  "_id": "63976041117eacc7c7398a66"
}

2. Upload videos

Once you've created an index and enabled thumbnail generation for it, you can upload videos to it.

The following example code shows how you can upload a video. Make sure to replace the placeholders surrounded by <> with your values.

TASKS_URL = f"{API_URL}/tasks"
file_name = "<FILE_NAME>" # Example: "test.mp4"
file_path = "<FILE_PATH>" # Example: "/Downloads/test.mp4"
file_stream = open(file_path,"rb")
data = {
    "index_id": INDEX_ID, 
    "language": "en"
}
file_param=[
    ("video_file", (file_name, file_stream, "application/octet-stream")),]
response = requests.post(TASKS_URL, headers=headers, data=data, files=file_param)
TASK_ID = response.json().get("_id")

const TASKS_URL = `${API_URL}/tasks`
const file_path = '<FILE_PATH>' // Example: "/Downloads/test.mp4"
const file_stream = fs.createReadStream(file_path)
let formData = new FormData()
formData.append('INDEX_ID', INDEX_ID)
formData.append('language', 'en')
formData.append('video_file', file_stream)
config = {
      method: 'post',
      url: TASKS_URL,
      headers: headers,
      data : formData
};
resp = await axios(config)
response = await resp.data
const TASK_ID = response._id
console.log(`Status code: ${resp.status}`)
console.log(response)

The output should look similar to the following one:

Status code: 200
{
  "_id": "639760f4117eacc7c7398a69"
}

3. (Optional) Retrieve the unique identifier of a video

You can skip the steps in this section if you want to retrieve thumbnails for a set of search results. The platform must have finished indexing your video before you can retrieve its unique identifier. Use the GET method of the /tasks/{_id} endpoint to monitor the indexing process.

  1. Construct the URL for retrieving the status of your video indexing task based on the TASK_ID variable you’ve declared in the previous section, and wait until the status shows as ready:

    TASK_STATUS_URL = f"{API_URL}/tasks/{TASK_ID}"
    while True:
        response = requests.get(TASK_STATUS_URL, headers=headers)
        STATUS = response.json().get("status")
        if STATUS == "ready":
            break
        time.sleep(10)
    
    const TASK_STATUS_URL = `${API_URL}/tasks/${TASK_ID}`
    const uploadResp = await new Promise((res) => {
    const interval = setInterval(async () => {
          const { data: response } = await axios.get(
                TASK_STATUS_URL,
                {
                    headers: {
                        'x-api-key': API_KEY
                    }
                }
            );
            if (response.status == 'ready') {
              clearInterval(interval)
                res(response)
            }
        }, 1000)
    });
    
  2. Store the unique identifier of your video in a variable named VIDEO_ID and print it:

    VIDEO_ID = response.json().get('video_id')
    print(f"VIDEO ID: {VIDEO_ID}")
    
    const VIDEO_ID = uploadResp.video_id
    console.log(`VIDEO_ID: ${VIDEO_ID}`)
    

    The output should look similar to the following one:

    VIDEO ID: 6391c8a669ff3402ec515aba
    

    Note that you can also use webhooks to monitor the status of the indexing process. For details, see the Using webhooks section.

4. Retrieve thumbnails

There are two ways to retrieve thumbnails, depending on your use case:

  • To retrieve a thumbnail of a video, you can call the GET method of the /indexes/{index-id}/videos/{video_id}/thumbnail endpoint.
  • To retrieve thumbnails for a set of search results, you can use the thumbnail_url field returned by the POST method of the /search endpoint.

Note that the platform returns a unique URL for each thumbnail, and the URLs expire in one hour.

Retrieve a thumbnail of a video

The GET method of the /indexes/{index-id}/videos/{video_id}/thumbnail endpoint retrieves a thumbnail of an individual video. By default, when calling this method, the platform retrieves a thumbnail from the middle of the video. Optionally, you can pass the time query parameter to specify the time, in seconds, at which the platform must retrieve the thumbnail.

Retrieve a thumbnail from the middle of the video

The following example code retrieves a thumbnail from the middle of the video by calling the GET method of the /indexes/{index-id}/videos/{video_id}/thumbnail endpoint and passing it the following arguments:

  • The unique identifier of your index
  • The unique identifier of your video.
THUMBNAILS_URL = f"{API_URL}/indexes/{INDEX_ID}/videos/{VIDEO_ID}/thumbnail"
response = requests.get(THUMBNAILS_URL, headers=headers)
print (f'Status code: {response.status_code}')
pprint (response.json())
THUMBNAILS_URL = `${API_URL}/indexes/${INDEX_ID}/videos/${VIDEO_ID}/thumbnail`
config = {
  method: 'get',
  url: THUMBNAILS_URL,
  headers: headers,
}
resp = await axios(config)
response = await resp.data
console.log(`Status code: ${resp.status}`)
console.log(response)

The output should look similar to the following one:

Status code: 200
{
  "thumbnail": "https://project-one-thumbnail.s3.us-west-2.amazonaws.com/6320748774fed2f3511f2a92/10.jpeg?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAYRWJPOVHXE5SJ77T%2F20220914%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Date=20220914T075213Z&X-Amz-Expires=86400&X-Amz-SignedHeaders=host&x-id=GetObject&X-Amz-Signature=46b3bac13fb8b022313244780c238cb6bfc53eddcbee3be23a8d2dd22d99543e"
}

Retrieve a thumbnail from a specific time

The following example code retrieves a thumbnail from a specific time by calling the GET method of the /indexes/{index-id}/videos/{video_id}/thumbnail endpoint, and passing it the following parameters:

  • The unique identifier of your index
  • The unique identifier of your video
  • The time, in seconds, at which the platform must retrieve a thumbnail

Make sure to replace the placeholder surrounded by <> with your value.

TIME = <SPECIFY_VALUE> #Example: 50
THUMBNAILS_URL = f"{API_URL}/indexes/{YOUR_INDEX_ID}/videos/{YOUR_VIDEO_ID}/thumbnail?time={TIME}"
response = requests.get(THUMBNAILS_URL, headers=headers)
print (f'Status code: {response.status_code}')
pprint (response.json())
const TIME = <SPECIFY_VALUE> //Example: 50
THUMBNAILS_URL = `${API_URL}/indexes/${INDEX_ID}/videos/${VIDEO_ID}/thumbnail?time=${TIME}`
config = {
  method: 'get',
  url: THUMBNAILS_URL,
  headers: headers,
}
resp = await axios(config)
response = await resp.data
console.log(`Status code: ${resp.status}`)
console.log(response)

Retrieve thumbnails for a set of search results

When you perform a search on an index for which thumbnail generation is enabled, the platform returns a field named thumbnail_url that contains a thumbnail from the middle of each matching video clip.

The following example code performs a search on an index for which thumbnail generation is enabled. Make sure to replace the placeholder surrounded by <> with your search query.

SEARCH_URL = f"{API_URL}/search"
data = {
    "query": "<YOUR_SEARCH_QUERY>",
    "index_id": INDEX_ID,
    "search_options": ["conversation", "visual", "text_in_video"],
    "operator": "or"
}
response = requests.post(SEARCH_URL, headers=headers, json=data)
print (f'Status code: {response.status_code}')
pprint (response.json())
const SEARCH_URL = `${API_URL}/search`
data = JSON.stringify(
  {
  'query': '<YOUR_SEARCH_QUERY>',
  'index_id': INDEX_ID,
  'search_options': ['conversation', 'visual', 'text_in_video'],
  'operator': 'or'
})
config = {
  method: 'post',
  url: SEARCH_URL,
  headers: headers,
  data: data
}
resp = await axios(config)
response = await resp.data
console.log(`Status code: ${resp.status}`)
console.log(response)

The following example output was truncated for brevity:

Status code: 200
{
  "conversation_option": "semantic",
  "data": [
    {
      "confidence": "high",
      "end": 426,
      "metadata": [{ "type": "visual" }],
      "score": 84.77,
      "start": 420,
      "thumbnail_url": "https://project-one-thumbnail.s3.us-west-2.amazonaws.com/63284a7ff8fcd9232826c11e/421.jpeg?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAYRWJPOVHXE5SJ77T%2F20220919%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Date=20220919T161325Z&X-Amz-Expires=3600&X-Amz-SignedHeaders=host&x-id=GetObject&X-Amz-Signature=5f37a929ea92e6c0a98707c55c9f2d73087ac889d486e3a3d1223fcc19bcd047",
      "video_id": "63284a7ff8fcd9232826c11e"
    },
    {
      "confidence": "high",
      "end": 502,
      "metadata": [{ "type": "visual" }],
      "score": 83.98,
      "start": 484,
      "thumbnail_url": "https://project-one-thumbnail.s3.us-west-2.amazonaws.com/63284a7ff8fcd9232826c11e/485.jpeg?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAYRWJPOVHXE5SJ77T%2F20220919%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Date=20220919T161325Z&X-Amz-Expires=3600&X-Amz-SignedHeaders=host&x-id=GetObject&X-Amz-Signature=b4b741b06a44c2a206903fc65cfd7932a4186e66e8f473fa83317710c59df8f1",
      "video_id": "63284a7ff8fcd9232826c11e"
    }
  ],
  "page_info": {
    "limit_per_page": 10,
    "next_page_token": "eb87df6f-c32e-48e0-93ac-4346020288f5-1",
    "page_expired_at": "2022-09-19T16:18:25Z",
    "total_results": 35
  },
  "query": "bear chasing a man",
  "response_type": "clip",
  "search_options": ["conversation", "visual", "text_in_video"],
  "search_pool": {
    "index_id": "63284a57e0a250d0763d12a3",
    "total_count": 1,
    "total_duration": 532
  }
}

In this example output, the data array contains two objects, and each object has a property named thumbnail_url containing the URL of the thumbnail for the matching video clip.