refine logic for identifying transcription chunks that indicate new image

[rlskoeser]

dev notes

• add the Hebrew label for page two (ע“ב) ; can test with https://geniza.princeton.edu/en/documents/3573/
• check if label starts with ENA or Moss.; test with https://geniza.princeton.edu/en/documents/1546/

[richmanrachel]

Hebrew label for "margins" should NOT indicate new page (but noting for convenience here): שוליים

[rlskoeser]

@richmanrachel I looked at PGPID 3573 and noticed that ע“ב didn't look like it was tagged as a label. I checked the TEI to confirm, and then did a quick search and I found 66 xml documents total where this text is tagged as a line instead of a label, i.e.:
<l n="">ע"ב</l>

Is it appropriate to convert all of these to <label> tags ?

[richmanrachel]

@rlskoeser - yes, please convert those to labels! This will be a huge help :)

[richmanrachel]

@rlskoeser - it doesn't seem to be working :( https://test-geniza.cdh.princeton.edu/en/documents/3573/

[rlskoeser]

@richmanrachel oh no! I thought it worked properly when I ran it locally, once I changed the lines to labels; will have to investigate

[richmanrachel]

@rlskoeser - Bummer! But the 1546 labels with ENA, etc, are working :)

[rlskoeser]

@richmanrachel it's a smart quotes / straight quotes problem! The original version you gave me used a smart quote, and that's what I'm checking for in my code, but when I did my search and replace I must have converted to straight quotes. Which one do you want in the transcription label?

[richmanrachel]

@rlskoeser - oy vey, the smart quotes strike again! The instructions on PPA only showed how to override them on iPhones, not Macs, and so I didn't go deeper...

You might need the transcription label to have both versions, so we don't run into this problem - is that possible?

[rlskoeser]

@richmanrachel yes, I'll revise so it works with both variants

[richmanrachel]

@rlskoeser - fixed! Thank you!