Title Back Colour Keyoti Title Line Title Curve
Blue Box Top

shows records which doesn't match word in a files. - SearchUnit - Forum

Welcome Guest Search | Active Topics | Log In | Register

Options
prashant
#1 Posted : Wednesday, February 18, 2015 10:11:46 AM
Rank: Member

Groups: Registered

Joined: 2/3/2015
Posts: 12
Hello,

1. I am uploading a document with two lines as:

username:xyz@gmail.com

Password:Usil@123

At the time of searching, when i tried to search for "Usil", it shows some result.
But when i tried to search for "xyz", then it is not showing any result.

2.
another case is that, i'm uploading 3 different document which contains text like
document 1- test
document 2- tests
document 3- test's

when i tried to search "test", it list out all documents. it shows highlighting text "test" from document-1.

in document-2 it not highlighting text "tests" but its shows document in search result.

in document-3 it highlighting text "test" from word " test's".
Dan
#2 Posted : Wednesday, February 18, 2015 2:56:33 PM
Rank: Advanced Member

Groups: Administrators, Registered

Joined: 10/20/2004
Posts: 456
Hi and thanks for posting.

Regarding point 1. - it should find "xyz" in your example. Is there anyway you can send the document in question...I guess not if it contains sensitive information but if you can please email to support at keyoti.com


2. The results you describe in point 2 are as designed because Lemma Searches are enabled by default. This can be changed in the configuration. For example if you change "Lemma Search Enabled" to false in the Configuration, you will find searches for each of your scenarios shows only one result.

There are other settings relating to such properties in the Configuration so be sure to have a good look.

Let me know if you're unsure of anything.

Dan
Keyoti
Dan Wright
prashant
#3 Posted : Friday, February 20, 2015 12:51:30 PM
Rank: Member

Groups: Registered

Joined: 2/3/2015
Posts: 12
Hi,

thanks for previous reply.
for first point i used Wildcard search, and it works..


Now i'm facing another issue which is describe below.

i have uploaded the pdf document which has following content,

"Table's test application".

file name : 20feb15-1_managers-contract_create-new-contract_version_1_4cd80d1e-0e3b-4370-be6a-7e2db250a76e.pdf

it shows in a result, when i serach any word from above sentence.

but in ResultPreview section it shows document text, in addition of that it also contains text like

"20Feb15-1_Managers-Contract_Create-New-Contract_Version_1_4cd80d1e-0e3b-4370-be6a-7e2db250a76e 20Feb15-1_Managers-Contract_Create-New-Contract_Version_1_4cd80d1e-0e3b-4370-be6a-7e2db250a76e.docx 20Feb15 1_Managers Contract_Create New Contract_Version_1_4cd80d1e 0e3b 4370 be6a 7e2db250a76e 20Feb15-1 Managers-Contract Create-New-Contract Version 1 4cd80d1e-0e3b-4370-be6a-7e2db250a76e"
at the end of document.
Jim
#4 Posted : Friday, February 20, 2015 10:38:23 PM
Rank: Advanced Member

Groups: Administrators, Registered

Joined: 8/13/2004
Posts: 2,667
Location: Canada
Hi, you can customize the result preview output with http://keyoti.com/produc...iew%20Customization.htm


You could do it with the simple sew_OnResultPreviewTextLoaded function in JS, try to determine the filename footer at the end and cut it off.

We will have built in support for this in a future version
-your feedback is helpful to other users, thank you!


prashant
#5 Posted : Saturday, February 21, 2015 10:56:43 AM
Rank: Member

Groups: Registered

Joined: 2/3/2015
Posts: 12
Hello,

i have uploaded document which has following content,

"Our new company Table#name this is our company".

when indexing a document, in code file WordBreakingCharacters & WordNonBreakingCharacters property set as ,

configration.WordBreakingCharacters = @"*=()[]{}<>""\-@.'#";
configration.WordNonBreakingCharacters = ",!$%^&+|;:?~`’”_/";

'#' charachter is mentioned in WordBreakingCharacters. but when try to search Table or name word seperatly it doesn't shows any records. if i search Table#name then it shows that record.


2.
I have uploaded three documents.
Text present in document 2: (Name name name name name name name)
Text present in document 3: (My name my name my name)
Text present in document 1: (My name is ABC)

When i am searching for text (My name is ABC), then search result shows the maximum count first.

i.e. it shows document result in following sequence -
document 2,
document 3,
document 1

it is possible to show document contain exact text (My name is ABC) first ?
means show document 1 at top of the result.
Jim
#6 Posted : Saturday, February 21, 2015 7:43:52 PM
Rank: Advanced Member

Groups: Administrators, Registered

Joined: 8/13/2004
Posts: 2,667
Location: Canada
1. I can't say for sure that this is the cause, but you've used xml entities in the strings. Those entities are apt in the configuration.xml file, but not in code. So you should have

configration.WordBreakingCharacters = @"*=()[]{}<>""\-@.'#";
configration.WordNonBreakingCharacters = ",!$%^&+|;:?~`’”_/";

2. It wasn't designed that way, but you could achieve it I think, using a post search sort, and the ResultItem.KeywordHitMap field. The KeywordHitMap property was added after the last release, so you would need to uninstall and reinstall to get the new DLLs.

Here's the basic idea.

1. Using this http://keyoti.com/produc...ith%20Custom%20Data.htm as a guide (you don't need to do anything with CustomData)

Code:

protected void Page_Load(object sender, EventArgs e)
{
    Sr1.FilterLoadLevel = Keyoti.SearchEngine.Search.FilterLoadLevel.Everything;
    Sr1.Configuration.CentralEventDispatcher.Action += new Keyoti.SearchEngine.Events.ActionEventHandler(CentralEventDispatcher_Action);

}

void CentralEventDispatcher_Action(object sender, Keyoti.SearchEngine.Events.ActionEventArgs e)

{

    if (sortByDate && e.ActionData.Name == Keyoti.SearchEngine.Events.ActionName.ResultItemsFinalized)

    {

        Keyoti.SearchEngine.Utils.ResultItemList resultItems = e.ActionData.Data as Keyoti.SearchEngine.Utils.ResultItemList;

        resultItems.Sort(new DocumentDateComparer(sortDirection));

    }

}


class DocumentDateComparer : IComparer<Keyoti.SearchEngine.Search.ResultItem>
{
public int Compare(Keyoti.SearchEngine.Search.ResultItem x, Keyoti.SearchEngine.Search.ResultItem y)
    {

        //in this method you will need to access x.KeywordHitMap and y.KeywordHitMap and if one of them has more keywords than the other, give it a +1 return value.

    }

}





That would be the idea... do you want the new MSI with KeywordHitMap included?
Jim
-your feedback is helpful to other users, thank you!


prashant
#7 Posted : Monday, February 23, 2015 5:23:52 AM
Rank: Member

Groups: Registered

Joined: 2/3/2015
Posts: 12
Hello,

yes, please provide me a link from where i can download new setup with KeywordHitMap.

for 1st point, i have done the previous changes in code. but still words contain # character is not search when try to search word before OR after # character separately as mention example in my last post.
Jim
#8 Posted : Monday, February 23, 2015 7:11:44 PM
Rank: Advanced Member

Groups: Administrators, Registered

Joined: 8/13/2004
Posts: 2,667
Location: Canada
About the # char: it appears to be a bug specifically relating to the # char. I've created a new MSI for you that includes the KeywordhitMap and the bug fix.

https://www.dropbox.com/...Hash%2BHitCount.msi?dl=0
-your feedback is helpful to other users, thank you!


Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.




About | Contact | Site Map | Privacy Policy

Copyright © 2002- Keyoti Inc.