[
https://issues.apache.org/jira/browse/NIFI-1115?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15091412#comment-15091412
]
Aldrin Piri commented on NIFI-1115:
-----------------------------------
Code and contrib both look good.
I realize the need to truncate descriptions, but don't necessarily care for the
search working against words that may or may not be available. As an example,
doing a search for "compress" will include TailFile in a result, but due to
length of description, the string "compress" will not be available anywhere
within the Processor selection UI. It is impossible to know if there is/is not
value within that part truncated by the ellipsis, but find myself not in favor
of showing results with no discernible mapping to the associated search.
Instinctually, I would imagine the key parts of the processor behavior would be
first in any description.
Additionally, would it make sense to have a prioritization of processors
something along the lines of Name -> Tags -> Description?
Overall, probably a +0/0.5 (if such a stance exists) given the above, but
certainly open to hear other folks' perspectives. Without the searching of
truncated description sections, +0.5/+1.
[~joewitt] Any thoughts on this being the issue reporter?
> Tokenize processor descriptions for searching for processors to add
> -------------------------------------------------------------------
>
> Key: NIFI-1115
> URL: https://issues.apache.org/jira/browse/NIFI-1115
> Project: Apache NiFi
> Issue Type: Improvement
> Components: Core Framework, Core UI
> Reporter: Joseph Witt
> Assignee: Matt Gilman
> Priority: Minor
> Fix For: 0.5.0
>
> Attachments: 0001-NIFI-1115.patch
>
>
> We currently use tokens from processor names and processor tags to aid in
> searching for processors in the add dialogue. A user brought up a great
> point that we should also tokenize the descriptions as they often have useful
> words that aren't reflected in the tags. Perhaps we should also consider
> stop/skip words here too.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
