OCR Worker
OCR Worker is configured via environment variables. Note that database and redis url must be same as for webapp.
Note
Source code for OCR worker is on github. Its docker image is on dockerhub.
PAPERMERGE__DATABASE__URL
Database URL. For PostgreSQL the database URL is given in following format:
postgresql://USER:PASSWORD@HOST:PORT/NAME
Example:
postgresql://scott:tiger@db:5432/mydatabase
For MariaDB and MySQL the URL scheme is mysql.
Example:
mysql://myuser:mypass@db:3306/paperdb
Note
Both web_app and ocrworker must have same PAPERMERGE__DATABASE__URL
PAPERMERGE__REDIS__URL
For Redis the URL is given in following format:
redis://HOST:PORT/NUMBER
For example:
redis://redis:6379/0
Note
Both web_app and ocrworker must have same PAPERMERGE__REDIS__URL
OCR_WORKER_ARGS
Pass celery worker parameters via this environment variable.
In case PAPERMERGE__MAIN__PREFIX value is empty, use OCR_WORKER_ARGS="-Q ocr" to
instruct worker to use name "ocr" for the input queue.
For non-empty prefixes, use OCR_WORKER_ARG="-Q prefix_ocr".
For example, if PAPERMERGE__MAIN__PREFIX=demo, then OCR_WORKER_ARG should have -Q as OCR_WORKER_ARG="-Q demo_ocr".
Long story short, if you don't know what celery is and how to work with it, use either:
OCR_WORKER_ARGS="-Q ocr"
or
OCR_WORKER_ARGS="-Q ocr -c 2"
Note
Celery? Not the plant 
! But Python's Celery this one. The task distribution thingy 
!
Storage
OCR Worker supports two storage backends: local file system and S3.