User markings not displaying properly when concept ends with a space and a number

[JTFouquier]

Eg- actin-interacting protein 2 (PMID: 15351742) includes the '2' in the annotation table, but is not reflected in the highlights of the talk and feedback pages.
Here's the annotation in the annotation table, with the '2'.

https://bitbucket.org/repo/y7oq9p/images/2422505321-Annotation%20bug%201(2).PNG

And here's the annotation in the talk page for the same user:

https://bitbucket.org/repo/y7oq9p/images/2507544802-Annotation%20bug%202(2).PNG

We've also observed this in the machine pre-population of a T4 doc (Nitric Oxide Synthetase 2), where if you highlight the entire term, you will get just 'Nitric Oxide Synthetase' underlined in the feedback.

---

• Bitbucket: https://bitbucket.org/sulab/mark2cure/issue/133
• Originally reported by: gtsueng
• Originally created at: 2015-11-30T23:07:45.258

[JTFouquier]

This bug is causing confusion for the users, and may affect data quality if they learn from it.

---

Original comment by: gtsueng

[JTFouquier]

The talk page here, https://mark2cure.org/talk/15351742/, shows that user 128 didn't highlight the "2" after protein,

but here https://mark2cure.org/task/entity-recognition/2615/user/128/results.json in the data, it clearly shows that the user did highlight the correct information. The data used to pre populate the information is fine. The length of the word is also correct in the json. The offsets are correct here so this does not really appear to be related to the "offset" issue.

[gtsueng]

@JTFouquier Thanks. This doc confirms that it doesn't have to be a number to be skipped which is the case for 'DNAse I' which uses 'I' instead of '1'. Does the display of the abstract (without the annotations) rely on the pubtator file? If so, does it utilize the incorrect offset at all? Eg- if the text offset is by 126, while the annotations is 125, it would make sense how that last character gets dropped in the display of the highlighting, even if the length is correct.

[x0xMaximus]

Unclear what the source is, but will look into it