-
Notifications
You must be signed in to change notification settings - Fork 7
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Setup guide: better recommendations for Docker usage #98
Comments
thx for reporting @bertsky ! In the meantime this was already solved by adding the Docker calls in the respective sections of the user guide, wasn't it? https://ocr-d.de/en/user_guide#calling-several-processors |
No, that's not the same thing! (But notice that IMO it does not make much sense (practically) to recommend/explain the Docker option by showing how to repeat each individual command/step with a But then you should also explain how to "log in" to a Docker container – spinning it up only once – and there do all the wonderful things explained everywhere in the "native" installation context, including |
So to sum up:
|
Am I the only one unsatisfied with documentation of the Docker option? |
No, I second your recommendation of reworking the documentation in this regard. Users who are not familiar with docker will be surprised to see how quickly their drives will be cluttered after executing |
I hear you and we will improve docker documentation.
You can mitigate that problem by using the |
the user guide already shows this each time, but not the setup guide.
Nice, didn't know this (used to help myself with |
@kba: Thanks for replying! I myself prefer not to fire up a new container every time but instead naming and Probably it’s easier not to include all docker hints on the individual pages but instead have a special page about the possible usage scenarios with docker. This could include mentioning |
The current formulation of the setup guide recommends running the docker image individually for the individual processor CLIs (translating native commands to docker calls). This is one possibility, but I would not recommend it first/exclusively.
The current (fat) Docker configuration is tailored specifically for usage of more than 1 processor within the same container, which is not only more efficient than spinning up multiple containers for multiple tools, but can also be much more convenient to set up a workflow. You can "login" to the image interactively, and (interactive or not) to run an entire workflow with multiple tools on your dataset, you currently have 3 options:
ocrd process
workflow-configuration
(ocrd-make)The latter already provides a number of sensible preconfigured workflows.
This is what I would document in the guide. (And the section
Running a small workflow
also does not cover this, or the Docker option, yet.)The text was updated successfully, but these errors were encountered: