Description #

Video keyframe extraction processor that supports dynamic detection using multiple algorithms.

Key features #

• Multi-algorithm support:
  • Pixel difference method (difference)
  • Optical flow method (optical_flow)
  • Histogram method (histogram)
  • I-frame identification (I_frame)
• Supports custom threshold and quantity control
• Provides timestamp positioning functionality
• Supports multiple output formats and storage options

Format support #

• Input: MP4, AVI, MOV, and other common video formats
• Output: JPG, PNG image formats

Performance recommendations #

• The I_frame method is the most efficient and is recommended as the preferred method

Operator parameters #

Input #

Output #

The processed array, with struct fields including:

• keyframes: list[list[list[list[int]]]], array format of keyframe images
• base64: list[str], base64 encoding of keyframe images
• timestamps: list[float], timestamps corresponding to keyframes (unit: seconds)
• tos_paths: list[str], storage paths of keyframes on TOS

Parameters #

If a parameter does not have a default value, it is required

Examples #

The following code demonstrates how to use Daft in a distributed environment to run the operator and extract video keyframes.

from __future__ import annotations

import os

import daft
from daft import col
from daft.las.functions.udf import las_udf
from daft.las.functions.video import VideoKeyframes

if __name__ == "__main__":

    # After extraction, keyframes will be saved to the specified TOS path. Therefore, environment variables must be set to ensure write permissions to TOS, including: ACCESS_KEY, SECRET_KEY, TOS_ENDPOINT, TOS_REGION, TOS_TEST_DIR
    TOS_DIR = os.getenv("TOS_TEST_DIR", "tos_bucket")
    output_tos_dir = f"tos://{TOS_DIR}/video/video_keyframes"
    
    if os.getenv("DAFT_RUNNER", "native") == "ray":
        import logging

        import ray

        def configure_logging():
            logging.basicConfig(
                level=logging.INFO,
                format="%(asctime)s - %(name)s - %(levelname)s - %(message)s",
                datefmt="%Y-%m-%d %H:%M:%S.%s".format(),
            )
            logging.getLogger("tracing.span").setLevel(logging.WARNING)
            logging.getLogger("daft_io.stats").setLevel(logging.WARNING)
            logging.getLogger("DaftStatisticsManager").setLevel(logging.WARNING)
            logging.getLogger("DaftFlotillaScheduler").setLevel(logging.WARNING)
            logging.getLogger("DaftFlotillaDispatcher").setLevel(logging.WARNING)

        ray.init(dashboard_host="0.0.0.0", runtime_env={"worker_process_setup_hook": configure_logging})
        daft.set_runner_ray()
    daft.set_execution_config(actor_udf_ready_timeout=600)
    daft.set_execution_config(min_cpu_per_task=0)

    # Construct URL using environment variables
    tos_dir_url = os.getenv("TOS_DIR_URL", "las-cn-beijing-public-online.tos-cn-beijing.volces.com")
    samples = {
        "video_path": [
            f"https://{tos_dir_url}/public/shared_video_dataset/sample.mp4"
        ]
    }

    ds = daft.from_pydict(samples)

    extractor = las_udf(
        VideoKeyframes,
        construct_args={
            "method": "I_frame",
            "keyframes_cnt": 5,
            "output_tos_dir": output_tos_dir,
        },
    )

    # Use Daft for distributed processing
    ds = ds.with_column("results", extractor(col("video_path")))

    ds.show()
    # ╭────────────────────────────────┬────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
    # │ video_path                     ┆ results                                                                                                    │
    # │ ---                            ┆ ---                                                                                                        │
    # │ Utf8                           ┆ Struct[keyframes: List[List[List[List[Int64]]]], base64: List[Utf8], timestamps: List[Float64], tos_paths: │
    # │                                ┆ List[Utf8]]                                                                                                │
    # ╞════════════════════════════════╪════════════════════════════════════════════════════════════════════════════════════════════════════════════╡
    # │ https://las-cn-beijing-publi-…  ┆ {keyframes: [[[[9, 17, 16], […                                                                             │
    # ╰────────────────────────────────┴────────────────────────────────────────────────────────────────────────────────────────────────────────────╯