SpreadFlow is a metadata extraction and processing engine. Its main purpose is to support file based media production workflows by observing changes to text and media files in real-time.
Information gathered through this process can be used to transform and recombine content, e.g., in order to prepare it for different publishing channels.
- Robust and fast filesystem change monitor supporting multiple platforms via native APIs.
- Built-in support for Metadata extraction (via exiftool), XSLT transformations (via lxml), PDF manipulation (via pdfrw), Image conversion and face/feature recognition (via Thumbor and OpenCV)
- Built around a Dataflow engine which provides parallelisation out of the box and makes it easier to build reusable components.
- Based on Python and Twisted ➡ Batteries included.