Demo processing server/web api deployment

[joschrew]

Current situation The processing server is integrated into core. The web-api implementation is available here: https://github.com/OCR-D/ocrd-webapi-implementation.

How it should be Both should be deployed on a vm to be able to upload workspaces, workflows and run processors and workflows. Deployment should be reusable and transparent through using ansible.

Steps

• create venv on VM to make processors available
• start processing server and webapi in docker
• reverse proxy for routing to endpoints /workspace /workflow /processor and /docs for swagger
• create/debug/improve nextflow-workflow-scripts to run workflows (re)using the processors available at /processor

current Tasks

• install ocrd-venv with processors with ocrd-all instead of install every processor individually
• publish the used nextflow-scripts somewhere. Maybe here: https://github.com/MehmedGIT/wf_server_nf_script
• test with a more realistic workspace then just one page
• try / Change: processing server does not start mongodb / rabbitmq itself but connects to existing one
• try to deploy a processors on other vms (Atm the processing-workers just run on the processing-server's own host)