Description #

Video cropping processor that supports multiple cropping modes and parameter configurations.

Key features: #

• Multiple cropping modes:
  • Precise cropping based on bounding box
  • Resolution-based cropping
  • Aspect ratio-based cropping
  • Multiple cropping positions (center, top, bottom, left, right)
• Encoding mode support:
  • CPU encoding (libx264): quality prioritized
  • GPU encoding (h264_nvenc): speed prioritized
• Automatic audio stream processing: keeps audio unaffected
• Multiple input and output methods: supports path input, binary input, and TOS output
• Fine quality control: adjustable encoding parameters and quality settings

Format support: #

• Input: MP4, AVI, MOV, and other mainstream video formats
• Output: consistent with input format

Parameter priority: #

• Bounding box (bbox) > target resolution (target_width/target_height) > aspect ratio (aspect_ratio)

Operator parameters #

Input #

Output #

An array containing the paths of cropping results. Returns the output path on success, returns an empty string on failure.

Parameters #

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

Examples #

The following code demonstrates how to use Daft (for distributed scenarios) to run the operator for video cropping.

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.video_crop import VideoCrop

if __name__ == "__main__":
    # The cropped video will be saved to the specified TOS path. Therefore, you need to set environment variables to ensure you have write permissions for TOS, including: ACCESS_KEY, SECRET_KEY, TOS_ENDPOINT, TOS_REGION, TOS_TEST_DIR
    TOS_TEST_DIR = os.getenv("TOS_TEST_DIR", "your-bucket")
    output_tos_dir = f"tos://{TOS_TEST_DIR}/video_crop"
    
    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",
            )
            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 the 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)

    cropper = las_udf(
        VideoCrop,
        construct_args={
            "output_tos_dir": output_tos_dir,
            "target_width": 1280,
            "target_height": 720,
            "crop_mode": "center",
        },
        num_cpus=1,
        concurrency=1,
        batch_size=1,
    )

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

    ds.show()
    # ╭────────────────────────────────┬────────────────────────────────╮
    # │ video_path                     ┆ crop_result                    │
    # │ ---                            ┆ ---                            │
    # │ Utf8                           ┆ Utf8                           │
    # ╞════════════════════════════════╪════════════════════════════════╡
    # │ https://las-cn-beijing-publi-… ┆ tos://your-bucket/video_crop/… │
    # ╰────────────────────────────────┴────────────────────────────────╯