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

Let the user get a link in the email when training a classifier is done, and then upload an unlabeled set #301

Open
monajalal opened this issue Jul 14, 2020 · 12 comments

Comments

@monajalal
Copy link
Collaborator

monajalal commented Jul 14, 2020

Right now, the user has to keep their tab open, to make sure that they can train and predict on a custom dataset of their own.

@dnaaun dnaaun pinned this issue Jul 14, 2020
@dnaaun dnaaun changed the title The problem right now is: the user has to keep their tab open, to make sure that they can train and predict on a custom dataset of their own. If instead, the user can get a link in the email when training is done, and then the user can upload an unlabeled set, and also get an email when inference is done, then that fixes a huge usability issue. Let the user get a link in the email when training a classifier is done, and then upload an unlabeled set Jul 14, 2020
@edward-edberg
Copy link
Collaborator

@monajalal @davidatbu this method contradicts with the current one. I still don't know how to do it subtly.

@edward-edberg
Copy link
Collaborator

I think we have to do it in the backend. Frontend modification is kind of limited.

@dnaaun
Copy link
Owner

dnaaun commented Jul 14, 2020

Hi @edward-edberg,

I don't see how modifying the backend will help with this. The backend already sends an email when training finishes, and then another one when predicting finishes.

I understand what you are saying in that, in the frontend, this is not possible if the user is going to upload training dataset and the unlabeled dataset at the same time. But, if we have different pages for the two, imagine this:

The user goes to www.openframing.org/train_a_classifier.html
He gets an email when training finishes at www.openframing.org/predict_with_classifier.html?classifier_id=<SOME_ID>

On the frontend, the JS inside predict_with_classifier.html will grab the <SOME_ID> from the URL(this is possible, and this is what we do for topic modeling previewing already), and then it will let the user upload testing data.

And then we send an email to the user with the results.

@edward-edberg
Copy link
Collaborator

Okay, but how about if the client wants to use default classifier? They will upload the test file in the first place, right?
If we want to implement a new mechanism that you have mentioned, then we will separate training from scratch page and testing-only page.

@dnaaun
Copy link
Owner

dnaaun commented Jul 15, 2020 via email

@edward-edberg
Copy link
Collaborator

@davidatbu I have a question, how do you provide unique link for each task to upload test file?

@dnaaun
Copy link
Owner

dnaaun commented Jul 15, 2020 via email

@edward-edberg
Copy link
Collaborator

@davidatbu ok, you can provide me topic_model_id, then how can I get that topic_model_id? since I have to make post request

@edward-edberg
Copy link
Collaborator

parse the url? or do you have any suggestion? because I have never done this before

@edward-edberg
Copy link
Collaborator

My code looks terrible compared to this lol.
Got it. I think I can build them.

@dnaaun
Copy link
Owner

dnaaun commented Jul 15, 2020 via email

@dnaaun dnaaun unpinned this issue Jul 20, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

3 participants