Search bar: erratic results

[samuel.g...@inserm.fr]

Hi,
Thanks for the latest improvements. I got some strange results when using the search bar in order to find some nodes. Please do find a simple network attached. Here are the testing queries and corresponding results:

• A => could not find any match
• A* => could not find any match
• B => find all correct match
• B- => find the same set as B, ie report A+B+ that should not IMHO
• B\- => find the same set as B, ie report A+B+ that should not IMHO
• B+ => find the same set as B, but should report A+B+ IMHO
• B\- => find the same set as B, but should report A+B+ IMHO

May be those erratic results are due to my install, and let me know. I didn't find any clue in the doc.

The network is very simple:
source    target
A+B+    A+B-
A+B+    A-B-
B-C+    A+C+

Best.

[Barry Demchak]

Interesting report, Samuel  (cytostaff: please have a look) –

 

From what I can see, the problem is that Cytoscape won’t find any nodes that start with A.

 

I have a simplified network below … a nice experiment is to load this network in and then search for “name:*”, which should select all nodes. Instead, it selects all nodes except A. All other queries give similar results.

 

I went back to Cytsocape 3.3, and the problem is still the same.

 

We’ll investigate further … sorry for the trouble.

 

source  target

X             Z

X             W

Y              A

B             C

--
You received this message because you are subscribed to the Google Groups "cytoscape-helpdesk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cytoscape-helpd...@googlegroups.com.
To post to this group, send email to cytoscape...@googlegroups.com.
Visit this group at https://groups.google.com/group/cytoscape-helpdesk.
For more options, visit https://groups.google.com/d/optout.

[samuel.g...@inserm.fr]

Hi,
Thanks for your attention. I find the problem in 3.4.0 then switch to 3.5.0 in order to check if the problem is still there.
I also tried "name:" and other stuff, but I get back to a simple report.
Let me know,
Sam

[Barry Demchak]

Hi, Sam --

I think we have a good answer. 

The issue is that the Search bar uses the Lucene search for indexing text. Lucene is normally used for prose, and is configured to drop common, meaningless words like "a", "the", "I" and others. So, you would have problems finding nodes named with those words. 

We concluded that Lucene is misconfigured for use in Cytoscape, as a node name is not ordinary prose.

We now have that on the bug list, with hopes to fix it in the next Cytoscape release, which won't be for some time.

For now, can you avoid naming nodes as filler words, you can avoid this problem.

Additionally, though, please note that there are a number of special characters Lucene gives meaning to, and won't find in a node name unless you escape them. The list is + - & | ! ( ) { } [ ] ^ " ~ * ? : \ ... you can read more here: http://manual.cytoscape.org/en/stable/Finding_and_Filtering_Nodes_and_Edges.html?highlight=search#search-bar

[samuel.g...@inserm.fr]

Hi Barry,

Many thanks for your job. I will wait for the next release.
I tried to escape the plus and minus characters, but it does not improve the results.

Best regards.

[Barry Demchak]

Right … the real problem is that common words are being eliminated from the search result.

 

Maybe try replacing “A” with “AA”?? (Not right, I realize, but it’s a cheap workaround.)

--

[tremblayr...@gmail.com]

Le samedi 1 avril 2017 07:28:20 UTC-4, samuel.g...@inserm.fr a écrit :
> Hi,
> Thanks for the latest improvements. I got some strange results when using the search bar in order to find some nodes. Please do find a simple network attached. Here are the testing queries and corresponding results:

> A => could not find any matchA* => could not find any matchB => find all correct matchB- => find the same set as B, ie report A+B+ that should not IMHOB\- => find the same set as B, ie report A+B+ that should not IMHOB+ => find the same set as B, but should report A+B+ IMHOB\- => find the same set as B, but should report A+B+ IMHOMay be those erratic results are due to my install, and let me know. I didn't find any clue in the doc.

[Barry Demchak]

Hi --

I see that this message is dated April 1. Since this was posted, we fixed these problems and released the result in v3.6, now available for download.

If you have more feedback, would you mind re-posting?? I think this post has somehow become garbled.