tesseract.js
tesseract.js copied to clipboard
Allow for setting parameters for single recognize job when using scheduler
At present, using a scheduler assumes that all workers are fungible, so any job can be sent to any worker. This means that any tasks that require 2+ "jobs" being sent to the same worker cannot be accomplished when using a scheduler. For example, in Issue #488 users point out that PDFs cannot be created using a scheduler (as this requires a recognize
and getPDF
job sent to the same worker).
Along with getPDF
, the one other function where this seems unnecessarily restrictive is setParameters
. It is useful to be able to run different recognition jobs with different settings (e.g. perhaps different types of images should be treated differently, or images with poor OCR results can be re-tried with different parameters). We should figure out how to let users set options when using schedulers (without changing the settings for all workers/jobs).