-
Notifications
You must be signed in to change notification settings - Fork 8
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
Missing highlights for most relations #177
Comments
How are we going to proceed with these? This type of an issue is dependent on pubtator and it's impossible now for QA user to evaluate if these are a data source problem or a YPet high lighting problem. I suggest we need to come to an agreement for determining what are words that are expected to be highlighted, this could be a new UI list or some way for QA user to differenciate between the two |
I'm sure you are aware that the behavior has changed since I see label changes... but I noticed the highlights were extremely sparse right now. So here is some updated testing: Zero highlights: Gene only is highlighted: Disease only is highlighted: |
@x0xMaximus, For the highlights it looks like this line in ypet is very often returning empty arrays for "var annotations" which is causing the highlights to be so sparse. When you see highlights, the array contains items. So something is wrong with "passage.annotations" in ypet because the "if(passage)" is always being evaluated. If you test with the links I provided last time, you can see this behavior. I know I'm behind, so sorry if this is old news. :) |
Apparently the lack of highlighting in the relationship module has gotten so bad that users were offering to contribute highlights for them. Some missing highlights in the relationship task as reported by users: PMID: 10980531 1.5/2 missing highlights: 1st pair (CDG-1b to protein losing enteropathy) Only CDG-1b was highlighted & that was only in the title ... not in the text. 2nd pair asked about mannose-6-phosphate & protein losing enteropathy. Nothing was highlighted. PMID: 21384228 2/3 missing highlights: 1st RD asked about dolichol & O-glucosylation and had 3 "dolichol" words highlighted - on lines 3, 5, & 10 (but there were more in the title & abstract). The 1st RD also had "dolichol-P-mannose" highlighted on line 6 ... but that was not a concept to consider. There were no highlights for O-glycosylation. 2nd RD asked for dolichol-P-mannose & glycosylation disorders. Nothing was highlighted. 3rd RD asked for polyprenol & glycosylation disorders. Nothing was highlighted. PMID 10914684: 1.5/2 missing highlights: 1st pair asked about N-glycan and glucosyltransferase. Nothing was highlighted. 2nd RD asked about oligosaccharides and genetic diseases. Only "genetic diseases" was highlighted. |
Missing highlights plagues majority of relationships. Of the 25 abstracts for relationship work displayed on my dashboard:
|
Extremely correlated to #204 - we now treat highlights the same if they're the users's or opponents so they're even more related now than they were |
I'm making this a running list of highlight errors to be debugged (low priority) as time allows. Consider editing the comment when each example is addressed.
Just one example... this is low priority because it is happening very rarely, but the blue/disease is not appearing. Might want to check the Pubtator.
The text was updated successfully, but these errors were encountered: