Details
-
Type:
Improvement
-
Status: Closed (View Workflow)
-
Priority:
Major
-
Resolution: Done
-
Affects Version/s: None
-
Fix Version/s: None
-
Labels:
-
Story Points:4
-
Epic Link:
-
Sprint:Fall 4 : 30 Sep to 11 Oct, Fall 5 : 14 Oct to 25 Oct, Fall 6 : 28 Oct to 8 Nov, Spring 2 : 20 Jan to 31 Jan, Spring 3 : 3 Feb to 14 Feb, Spring 3 : 17 Feb to 28 Feb, Spring 4 : 2 Mar to 13 Mar, Spring 5 : 16 Mar to 27 Mar
Description
Sai Charan Reddy Vallapureddy has implemented a topic branch with major potential improvements to the search functionality in IGB.
See: https://jira.transvar.org/browse/IGBF-1096
We need to revisit that branch and determine how to proceed with incorporating the new code.
Attachments
Issue Links
- relates to
-
IGBF-1096 Searching two files only returns one hit
-
- Closed
-
Activity
Using advanced search for rs1122608 on diseaseMom and diseaseDad files:
Properties - finds both hits
Keyword - finds one hit
ID, Name, or Title - finds both hits
I'm not sure if the Keyword results are correct. I would presume that keyword should also find both hits. Charan and I talked, and there may be an additional method that needs to be altered or overridden for the keyword search (possibly KeyWordSearch.java search). We will take a further look on Wednesday.
Philip Badzuh Nowlan Freese - Should P.B. commence testing? If yes, please use installer from my fork (see above link).
Another odd behavior:
Dr. Loraine pointed out that in 9.0.2 using the advanced search with no input would lead to IGB searching forever. This same behavior is present in this branch (2032-2) that includes Charan's updates. However, in 9.1.0 master (as of November 22, 2019) searching in IGB with no input returns all hits (for example every gene in the annotation) very quickly (even with tens of thousands of results). This behavior in 9.1.0 in master could be useful for users, and it appears to work consistently.
If we merge in 2032-2, we will need to address no input searching forever by forcing the user to enter input before searching.
Won't merge for 9.1.
Will come back to this in Jan (Dec?) for re-factoring search using OSGi pattern demonstrated in KeyWordSearch bundle.
Plan new approach.
For next step, Sai Charan Reddy Vallapureddy
Investigate possibly overriding equals method on SeqSymmetry
as a possibly more light-weight fix.
Also, investigate what would happen if we were to change other search code to use arraylist instead of hashset.
Note: You need to check two different ways of searching: Using AdvancedSearch and using the "Quick search" box at the top left of the IGB window.