⚓ anchorr-engine

A high-performance Rust-backed video processing engine for Python. Built to handle heavy-duty FFmpeg tasks with parallel execution and Scene-compliant defaults.

Features

• Parallel Processing: Uses Rust/Rayon to run batch encodes concurrently across all CPU threads.
• Scene Compliance: Automates Mod-16 scaling, metadata stripping, and 10-bit HEVC profiles.
• Smart Crop: Integrated cropdetect logic to automatically identify and strip black bars.
• Low Overhead: Direct binding via PyO3—no shell overhead for metadata extraction.

Prerequisites

You must have ffmpeg and ffprobe installed and available in your system PATH.

Installation

uv pip install anchorr_engine

Quick Start

1. Probe Metadata

from anchorr import AnchEngine

anch = AnchEngine()
info = anch.probe("video.mkv")
print(f"{info.width}x{info.height} using {info.codec}")

2. Auto-Crop & Encode

# Detect black bars automatically
coords = anch.get_blackbar_coords("input.mkv")

# Encode to HEVC 10-bit (mod-16 compliant)
anch.transform(
    input="input.mkv",
    output="output.mkv",
    codec="libx265",
    res="1080",
    crop=coords,
    flags="-crf 18 --preset slow"
)

3. Batch Encode

# List of (input, output, codec, resolution, crop, flags)
tasks = [
    ("v1.mkv", "v1_out.mkv", "libx265", "1080", "1920:800:0:140", ""),
    ("v2.mkv", "v2_out.mkv", "libx265", "1080", "1920:1080:0:0", "")
]

# Runs all encodes in parallel
results = anch.batch(tasks)

Development

To build from source:

1. Install Rust/Cargo.
2. Run uv run maturin develop to compile the Rust extension into your local Python environment.

License

MIT