[bt-devel] [ bibletime-Bugs-1466310 ] Search results problems

Wed Oct 18 22:42:51 MST 2006

Bugs item #1466310, was opened at 2006-04-07 15:06
Message generated for change (Comment added) made by ewst
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=100954&aid=1466310&group_id=954

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Frontend / Search dialog
Group: resolved
Status: Closed
Resolution: Fixed
Priority: 5
Submitted By: Wolfgang Stradner (ewst)
Assigned to: Jim Campbell (jim-campbell)
Summary: Search results problems

Initial Comment:
The search tools (Search in standard Bible/open
works/current bible) function in an unusual way to
display search results:
Searching for X (e.g. 'eat') in WEB gives the following
results:
- 1hit: it gives the references of the verse, which
contains X and shows in addition 2 verses before and 2
verses afterward
- if X is contained in more than one verse (within the
range of 5 verses), only one middle verse is given as
reference and some verses around are displayed, those
hits (shown by the sourounding verses) are counted as
one hit and show up only once.e.g.
-- 4 hits in Deut 14:9-12, the reference given is Deut
14:11 and displayed is14:9-13. (hits: the first 4 verses),
-- 3 hints in John 4:31-33, the reference given is John
4:31, displayed is John 4:29-33 (hits: last 3 verses),
- in both cases: not only the search for word X (e.g.
eat) is highlighted in the 5 verses displayed, but also
words containing the search word (e.g. Act 10:11-15:
great and eaten)
Testing out, what happens, when searching for words,
which appear more frequent than in five consecutive
verses (e.g. 'the' in Matthew) the Search is not working.

-> Allow the main search tools to
- highlightened only the words searched for
- search for frequent words in consucutive verses

----------------------------------------------------------------------

>Comment By: Wolfgang Stradner (ewst)
Date: 2006-10-19 07:42

Message:
Logged In: YES 
user_id=1385862

OK, I understand. Do you have any influence, which words are
exluded (for 'the' 'and' this is OK, for 'then' I am not so
sure).

As this feature (exclusion of some frequent English words)
is not common in bible programs (it is the first time, I saw
this) and as it is not working in other languages (e.g. German)

-> let BT give a message (when trying to search for such an
excluded word) that it does not look for this word,
-> put in a phrase in the handbook (like the one you wrote below
-> what other programs are doing, that they start a search
and having found a number of hits (result), it asks if it
should continue the search. As BT finds 3500 hits fast (e.g.
God in WEB), this number could be 3000.

----------------------------------------------------------------------

Comment By: Martin Gruner (mgruner)
Date: 2006-10-18 18:07

Message:
Logged In: YES 
user_id=169722

This is not a bug! The search engine we use currently 
removes a bunch of english stop words such as "the" 
or "and" that would produce a lot of hits. Maybe we should 
change this, however. Not sure.

----------------------------------------------------------------------

Comment By: Wolfgang Stradner (ewst)
Date: 2006-10-18 11:55

Message:
Logged In: YES 
user_id=1385862

Checking it again shows:

The problem with not showing up if it occurse a sequence of
more than 5 verses works OK now:

e.g. WEB/Matthew/5: Blessed occurs 9 times in a row and is
displayed correctly.

Then I tried to search in WEB for 
-'the' only in Matthew: no result
-'the' only in Tite: no result
-'they' only in Tite: no result
-'then' only in Tite: no result
-'therefore' only in Mattew: OK

Then I used: Settings / Bookshelf Manager / Manage Search
indices / selected the WEB index, deleted it and created it,
this does not make any change.

-> Check this problem out.

----------------------------------------------------------------------

Comment By: Martin Gruner (mgruner)
Date: 2006-10-16 20:16

Message:
Logged In: YES 
user_id=169722

Wrong bug report, sorry.

----------------------------------------------------------------------

Comment By: Martin Gruner (mgruner)
Date: 2006-10-16 20:15

Message:
Logged In: YES 
user_id=169722

It seems to work for me. In pilgrim, searching for "slough", I get hits in 

/CONTENTS
/PART I/THE FIRST STAGE
/PART I/THE SECOND STAGE
/PART I/THE FIFTH STAGE
/PART II/THE FIRST STAGE
/PART II/THE SIXTH STAGE
/PART II/THE EIGHTH STAGE

And I can also find all other words you listed.

Please try to delete your indices (using the bookshelf manager) and re-creating them (just execute a search).

----------------------------------------------------------------------

Comment By: Wolfgang Stradner (ewst)
Date: 2006-04-10 17:37

Message:
Logged In: YES 
user_id=1385862

Great. About the Tracker-markup: Is it right to mark it as:
- Status: Closed
- Group: In Progress?

----------------------------------------------------------------------

Comment By: Martin Gruner (mgruner)
Date: 2006-04-07 17:52

Message:
Logged In: YES 
user_id=169722

Wolfgang, I believe this is (at least partly) fixed in CVS.
We will try to release another beta soon so that you can
test it again, ok?

----------------------------------------------------------------------

Comment By: Jim Campbell (jim-campbell)
Date: 2006-04-07 16:13

Message:
Logged In: YES 
user_id=689979

I'm sorry for my confusion.  But if it is just the
highlighting that is not working, I believe this is fixed in
cvs.  I made several changes to cvs to enhance the
highlighting according to clucene rules.  

There are still some problems with it. Such as searching for
"kin*m", the regex parser does not pick this up properly (or
most likely I did not set-it-up correctly).  But the problem
stated should be fixed.  Searching for "eat" will not
highlight 'eat' within words like great and eaten.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=100954&aid=1466310&group_id=954