Getting Started with AMP - AudiovisualMetadataPlatform/amp_documentation GitHub Wiki

What is AMP?

AMP is an open source tool built with the goal of enabling efficient creation of metadata about digitized and born digital audio and moving image collections. These collections may contain significant backlogs of unprocessed materials, which makes them virtually impossible to discover and use. The vision of AMP is to allow archivists and librarians to create workflows leveraging artificial intelligence and machine learning services in combination with human expertise to help generate metadata that supports identification, discovery, navigation, and rights determination for audio and moving image collections. These workflows can include tools that perform tasks such as speech-to-text, named entity recognition, applause detection, audio segmentation, video optical character recognition, and facial recognition.

• Watch a brief video demonstrating how AMP can be used to create and run workflows. 
• Learn more about the Metadata Generation Mechanisms (or MGMs) available in AMP
• Use the AMP User Guide to learn more about how to actually use AMP.
• View a demo of the MGM evaluation functionality within AMP.
• (coming in fall/winter 2023) Visit the demo site to try out AMP functionality. AMP components

AMP consists of a number of components that are packaged together (see system architecture diagram).

• AMP UI - a front-end component built in Vue which communicates with the back-end via API 
• AMP Rest back-end - the back-end Java component that provides API and logic for access control, file management, etc. and communicates with Galaxy for workflow related tasks.
• AMP Galaxy - AMP bundles a forked and customized version of Galaxy, an open source platform for data analysis which provides infrastructure for creating and running workflows (see Galaxy docs)
• AMP MGMs - AMP bundles in a series of MGMs, both provided by open source communities or locally developed; for both local and cloud MGMs, AMP includes a wrapper that allows them to plug into Galaxy workflows
• MGM Evaluation scripts - the code used by the MGM Evaluation component in AMP to evaluate accuracy of workflow outputs

All of these components are included in the AMP package and installed / configured as part of the scripted installation process (see below).

How do I install AMP locally?

In order to simplify installation of a complex package, the AMP team has created a scripted VM installation mechanism. 

Follow the instructions in the amp_boostrap repository readme to get AMP running locally. First, you'll need to get a bare VM set up with Rocky Linux and pull down the bootstrap repository. The bootstrap repository provides a control script for managing your AMP installation, and will automatically install the most recent released version. You can then run the amp_bootstrap control script to download, install, configure, and start the AMP system with a few commands.

System requirements are included in the documentation.

How do I configure AMP?

Initial AMP configuration is done before starting the server for the first time through the amp.yaml file. Specific instructions for initial configuration are available as part of the overall installation process (you'll find them when you work your way through the installation documentation!). 

Are there known issues?

See the list of known issues in the current and past releases.