Pulumi with Artifactory Demo
This repository shows how to use JFrog Artifactory in conjunction with Pulumi.
The Pulumi application uses Amazon Web Services to enable users to upload video files to S3. Those
videos are then analyzed using Rekognition to detect objects at certain key frames. Each object
that is detected is then extracted from the video, using
ffmpeg running on Fargate.
The code is based on some examples in the pulumi/examples repository. Though this repo separated out the components involved so they could be built and versioned independently, and served from Artifactory.
video-labeler directory contains an NPM package abstracting the use of AWS Rekognition for
analyzing images. The Pulumi program will consume the
video-labeler package so that it can be
blissfully unaware of how Rekognition, Lambda, SNS, and the IAM policy schenanigans that go on
under the hood.
thumbnail-extractor directory contains a
Dockerfile which contains a script for extracting
images from a video stored on S3. It uses
ffmpeg under the hood.
analyzer-service directory contains the Pulumi program that puts it all together.