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

UI for visualizing and comparing HTR output for a page #4327

Open
benwbrum opened this issue Sep 11, 2024 · 2 comments · May be fixed by #4392
Open

UI for visualizing and comparing HTR output for a page #4327

benwbrum opened this issue Sep 11, 2024 · 2 comments · May be fixed by #4392
Assignees

Comments

@benwbrum
Copy link
Owner

benwbrum commented Sep 11, 2024

Here is sample data to use, pulled from this page:

Plaintext for version

Miss Mildred Cameron
Raleigh
N. Karohna
To Miss Mildred Cameron
of Raleigh. N.C.
Philadelphia Sep 4/59
Dear Madame
I beg leave to introduce
myself for the sake of my subject, which I
know will interest you, and which I hope
will elect your warmest sympathies.
I have been lately touched to the heart
with a case of heart breaking distress which
you have it entirely in your power I find
to cure. And I know by my own mother
my sister and my wife that there can
be no surer confidence placed in any thing
on earth that that which a man instruct
ively places upon the delight with which
 saamen find they can alleviate or remove
distress. In this case the misery has been
so Comtracted and is so deep that if I can
2.
only present it clearly I feel sure that
you will conteht out your hand to it to 

Ground truth/alternate version for comparison (below)

To Miss Mildred Cameron
of Raleigh, N. C.
Philadelphia Sep 4/59

Dear Madame
I beg leave to introduce
myself for the sake of my subject, which I
know will interest you, and which I hope
will elicit your warmest sympathies.

I have been lately touched to the heart
with a case of heartbreaking distress which
you have it entirely in your power I find
to cure. And I know by my own mother
my sister and my wife that there can
be no surer confidence placed in any thing
on earth than that which a man instinctively
places upon the delight with which
women find they can alleviate or remove
distress. In this case the misery has been
so protracted and is so deep that if I can
only present it clearly I feel sure that
you will stretch out your hand to it to
@saracarl
Copy link
Collaborator

We need to build a UI for viewing and comparing results of machine generated text ("HTR", Handwritten Text Recognition). It will look something like this:

AIComparison

  1. A new "HTR" tab when HTR data is present. (@benwbrum to give a way to determine if HTR data is present.)
  2. A view that is similar to the versions tab, but allows users to select a variant for the left and the right side of the screen. @benwbrum to define how to list the variants for this dropdown.
  3. The text view is a diff view; we should be recycling the versions screen here.
  4. Underneath the text view is the image overlay view, but two columns with data from the selected variants. Recycle the one from the transcription screen. (I'm assuming it's easiest to use the same controls, but if it's not we definitely need zoom and AI Assist on/off.)
  5. If we only have one variant -- i.e. just one service's HTR data -- then we want to show the image overlay on the left and the text on the right.

@benwbrum
Copy link
Owner Author

You should be able to ask @page.has_alto? and have that respond. Currently that method just looks for a file on the filesystem matching the ID of the page. We plan to build something more sophisticated, later. The relevant code is here

You can get the alto_xml and ai_plaintext and page image for version A from this Google Cloud Vision example

You can get the alto_xml and ai_plaintext for the same image for version B from this Amazon Textract example, which was run on the same image.

@benwbrum benwbrum linked a pull request Nov 6, 2024 that will close this issue
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

3 participants