Skip to content
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

[formrecognizer] adds samples for labeled tables and addresses sample/doc feedback #18409

Merged
merged 5 commits into from
Apr 30, 2021

Conversation

kristapratico
Copy link
Member

Resolves #18047

This also adds the training docs for the fixed/dynamic tables samples. We got feedback from the docs team that it was really helpful to have this for users.

Copy link
Member

@catalinaperalta catalinaperalta left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Adding all the OCR files is mainly for customers to be able to see what's needed for training labeled tables?

@kristapratico
Copy link
Member Author

Adding all the OCR files is mainly for customers to be able to see what's needed for training labeled tables?

Yes, and they can download the files and put them in a container to train on. Let's them skip the whole labeling process and run the samples more easily.

Copy link
Member

@catalinaperalta catalinaperalta left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🚢

class TestDifferentiateOutputLabeledTablesAsync(object):

async def test_recognize_tables_fixed_rows_async(self):
from azure.core.credentials import AzureKeyCredential
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is this sample runnable?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

yes

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

hmm are the model ids created every time the sample is run? or how does the magic work?
Not blocking your PR, just curious

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No, these are static model IDs that get passed in as environment variables. All samples are run from an existing FR resource, nothing created on the fly here.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I worry someone might delete the models and then all samples stop working, but it will be pretty clear, so easy to fix.
of course, if it ever becomes a problem the samples can be updated an in charge of creating the model too

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah still looking for a better way to do this without having to add the "train a model" code to every custom form sample. Trying to keep the samples as focused as possible. I guess one option could be to make the samples have the option of training a model or passing an existing model_id and have the CI always do the former. Will look into this. #18465

@Azure Azure deleted a comment from check-enforcer bot Apr 30, 2021
@kristapratico kristapratico merged commit 70406b4 into Azure:master Apr 30, 2021
@kristapratico kristapratico deleted the tables-samples branch April 30, 2021 21:15
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

[formrecognizer] sample/docs feedback
3 participants