483812 - Search terms should appear bolded in search results

Status: VERIFIED FIXED

(support.mozilla.org :: General, defect, P1)

Description

[Stephen Donner [:stephend] Not actively reading bugmail]

Steps to Reproduce:

1. Load http://support-stage.mozilla.org/tiki-newsearch.php?where=d&locale=en-US&q=bookmark&sa=
2. Look for "bookmark" in the search results

Expected Results:

"bookmark" should appear bolded

Comment 1

[David Tenser [:djst]]

One of those search switch bugs that would be really really nice to have. Eric? :)

Comment 2

[E. Cooper [:ecooper]]

How fancy are we getting here?

Exact matches only? Plurals?...transitive verbs...gerunds?

Comment 3

[E. Cooper [:ecooper]]

Err...I meant verb tenses when I said transitive verbs.

...sigh.

Comment 4

[David Tenser [:djst]]

Ideally like Google is doing it, which is to highlight the individual words: http://www.google.se/search?q=firefox+support+help&ie=utf-8&oe=utf-8&aq=t&rls=org.mozilla:sv-SE:official&client=firefox-a

This is how our current search engine does it on SUMO prod, too.

Comment 5

[E. Cooper [:ecooper]]

The current SUMO uses google, doesn't it?

At the very least, it will bold the individual terms. If including plurals, gerunds, and verb tenses doesn't make things buggy, I'll include that too.

Comment 6

[Paul Craciunoiu [:paulc]]

Works for individual words.
r23761 / r23762


Other things, like plurals, gerunds, and so on are language-based and so ... I could maybe give it a try for English only. Either way, it'll be a hassle unless there's some large database out there that we can use.

Comment 7

[David Tenser [:djst]]

Yeah that sounds challenging to say the least. Wasn't sure what Eric meant when he said he could include that too. I have no idea how to bend e.g. a Japanese noun. :)

Comment 8

[E. Cooper [:ecooper]]

As I told Paul, I was just going to "wing it."

Comment 9

[Paul Craciunoiu [:paulc]]

Just so we're clear, this bolds all parts of words that match the search.
For example, on:
https://support-stage.mozilla.org/tiki-newsearch.php?locale=en&q=it+feed&where=all&sa=&filter_lang=1&l=en&en_too=1&lastmodif=0&type=0&author=

Both "it" and the "it" part of "item" are bolded in the 2nd result. I could potentially add an exception for languages that have words separated by spaces (or punctuation). I think japanese is not one of them, for example.

So, yeah, let me know how far we need to go and I can try to see :)

Comment 10

[krupa raj[:krupa]]

Search phrase highlighting is not case-sensitive.Reopening..

Comment 11

[krupa raj[:krupa]]

Attachment: highlighting not case-sensitive

Comment 12

[Paul Craciunoiu [:paulc]]

We need a better replace function. Anyone got more experience with preg_match?
Needs to replace case-sensitive with case-sensitive.
I need help! :)

Comment 13

[krupa raj[:krupa]]

Correction:Oops..my comments, 11 and 12-should say case insensitive.

Comment 14

[Paul Craciunoiu [:paulc]]

Updated. Now the search terms are bolded in title and summary as entire words. A word is considered anything that's not followed or preceded by punctuation or spaces.

This will need some good amount of testing once newsearch works again (seems that currently it doesn't on staging).

r24213 / r24215

Comment 15

[Stephen Donner [:stephend] Not actively reading bugmail]

Verified FIXED on https://support-stage.mozilla.org/tiki-newsearch.php?where=d&locale=en-US&q=bookmark&sa=; we'll need to keep testing, as you say, to look for regression and/or other lurking bugs.

Comment 16

[E. Cooper [:ecooper]]

Empty queries bold entire summaries.

https://support-stage.mozilla.org/tiki-newsearch.php?locale=&q=&where=all&sa=&filter_lang=1&l=ar&en_too=1&lastmodif=0&type=0&author=

Comment 17

[Paul Craciunoiu [:paulc]]

r24426 / r24427
Good catch Eric. Thanks.

Comment 18

[Stephen Donner [:stephend] Not actively reading bugmail]

Verified FIXED;

Search for bookmarks: https://support-stage.mozilla.org/tiki-newsearch.php?locale=en&q=bookmarks&where=all&sa=&filter_lang=1&l=en&en_too=1&lastmodif=0&type=0&author=

Empty search: https://support-stage.mozilla.org/tiki-newsearch.php?locale=&q=&where=all&sa=&filter_lang=1&l=ar&en_too=1&lastmodif=0&type=0&author

Comment 19

[E. Cooper [:ecooper]]

Attachment: Highlighting doesn't work with multibyte characters

Just as a heads up...

When/if the patch from bug 489277 lands, this will be broken for languages with  multibyte encodings, like Japanese.

Maybe we could move to something simpler (at line 635), like $preg_repl = "/(\b" . preg_quote($enc_q)."\b)/Ui";

Comment 20

[Paul Craciunoiu [:paulc]]

(In reply to comment #19)
> Maybe we could move to something simpler (at line 635), like $preg_repl =
> "/(\b" . preg_quote($enc_q)."\b)/Ui";
The idea of that was to bold the entire word, and yes, I must've missed the \b option. It's definitely better. I should wait for that patch to land first :)
Thanks again Eric!

Comment 21

[Paul Craciunoiu [:paulc]]

Committed Eric's solution to fix bug 489372.