Semantic Search

We are implementing semantic search from note 866498. When
execute the pyton script txt2xtm.py, we have this problem:
suarbulvp1p01:td0adm 76> ./TREXSettings.csh
suarbulvp1p01:td0adm 78> exe/extensions/xtm/txt2xtm.py --
input=/usr/sap/TD0/TRX10/exe/topicmaps/example/example.txt --
output=/usr/sap/TD0/TRX10/exe/topicmaps/example/example.xtm
from: Command not found.
import: Command not found.
Badly placed ()'s.
I can´t determine wich variable is wrong.
env:
SAPSYSTEMNAME=TD0
DIR_LIBRARY=/usr/sap/TD0/SYS/exe/run
LD_LIBRARY_PATH=/usr/sap/TD0/TRX10/exe:/usr/sap/TD0/TRX10/exe/Python/lib:/usr/sap/TD0/TRX10/exe/filter:/usr/sap/TD0/TRX10/exe:/usr/sap/TD0/TRX10/exe/Python/lib:/usr/sap/TD0/TRX10/exe/filter:/usr/sap/TD0/SYS/exe/run
JAVA_HOME=/opt/IBMJava2-amd64-142
DIR_INSTANCE=/usr/sap/TD0/TRX10
DIR_EXECUTABLE=/usr/sap/TD0/TRX10/exe
SAP_RETRIEVAL_PATH=/usr/sap/TD0/TRX10/suarbulvp1p01
SECUDIR=/usr/sap/TD0/TRX10/sec
SNC_LIB=/usr/sap/TD0/TRX10/exe/libsapcrypto.so
PYTHONPATH=/usr/sap/TD0/TRX10/suarbulvp1p01:/usr/sap/TD0/TRX10/exe:/usr/sap/TD0/TRX10/exe/python_support
PYTHONHOME=/usr/sap/TD0/TRX10/exe/Python
Could you please help me?.

Hi,
As far as I know, Semantic Search is not supported in SharePoint 2013 by OOTB. You need to consider the third party product to achieve your demand.
You can have a look at the thread:
https://social.technet.microsoft.com/Forums/office/en-US/85a26e35-e06c-4511-b68f-c91908506c52/sharepoint-searchfast-search-support-for-semantic-search?forum=sharepointgeneralprevious
Best Regards,
Eric
TechNet Community Support
Please remember to mark the replies as answers if they help, and unmark the answers if they provide no help. If you have feedback for TechNet Support, contact
[email protected]

Similar Messages

Problem with semantic search (topicmaps) and german umlauts (ä,ü,ö)

Hello,
i've implemented semantic search like it is described in "How to Enable semantic search.pdf". If i create a synonymlist with a word containing ä,ü or ö; for example "nuts;nüsse", and now i search for nuts, i get the following error message:
<<Error during search occurred - com.sapportals.wcm.WcmException: error parsing XML data (Unsupported character: 68 (:main:, row:37, col:4348)) (Errorcode 7260)>>
If i use "nuts;nuesse" everything works fine and trex finds both search terms.
Is there any possibility to solve this problem?
Thanks,
Sven

Hello Sven
You need to save the .txt-file with UTF-8 encoding. Try opening your synonym.txt-file with notepad and change the encoding to UTF-8. Then you convert the txt-file into the xtm-file and the xtm-file into .db-file.
Kind regards,
Martin

Fuzzy & Semantic Search in TREX

Hi Guys,
Does anyone know how we can enable Fuzzy & Semantic Search in the ISA B2B application?

Hi,
You can use them but you have to change the standard flow of thee application.
[Modify Product Catalog Search in E-Commerce|https://wiki.sdn.sap.com/wiki/display/Snippets/ModifyProductCatalogSearchin+E-Commerce]
Please follow the above code to use or you can modify the existing application.
Regards,
Sateesh Chandra

Semantic Search and Exact search

Hi,
We have implemented semantic search for our trex like it is described in the HowToGuide.
We defined a lot of synonyms and it works fine.
But,
we expected, that the semantic search is an extension to the linguistic search method, so that synonyms are not taken into account, if we use the exact search method. But trex always searches for synonyms.
When a user uses the exact search method he doesn't want to search for synonyms.
How can we implement, that trex only searches for synonyms, if the automatic or linguistic search method is used?
Maybe someone has an idea.
Thanks!
Regards,
Sven

Hi Sven
I have implemented the same synonym search as you mentions, and I noticed the same problem.
Unfortunately I have never stumpled across a workaround for this issue.
Lets hope that either Mikhail Sirotkine or Matthias Röbig-Landau from development sees this question and answers it. I believe they should know the answer.
Kind regards,
Martin

How to search for a variety of characters

I am working on a document wherein a script I ran has applied differential results to page indicators. I am trying to ensure that I have easy accessibility to indexing in an e-book format, and so I am doing the following:
1.) Inserting notes at physical page breaks in the print layout in InDesign
2.) making those notes visible
3.) applying the "page" style to them, which I have set to display hidden in my CSS but will allow me to insert anchor points for a hyperlink index
The problem is the note shows up differentially. Some say:
{~?~PG: @##@} while others show {~?~PG: %##%} where ## represents some page number. The @##@ received the page style properly, but not the %#%. What would be the GREP approach to searching for every instance of {~?~PG: %##%} and applying the page style to it?
I know how to make sure that the page style is applied. I know that I'd use (\d+) for the numbers. I don't know hwo to represent the brackets, tildes, question marks, or other characters. Is there a good reference for this somewhere or perhaps a tutorial? I hate to bug the forum community with it. I just don't know where to go to get the info I need.
Edit: To be clear, I'm trying to search for the string
{~?~PG: %##%}
All of the characters in the string are the same each time, but the numbers are sequentially higher and higher. There are hundreds of these in my document, so replacing each manually is taking a long time. Thanks for any feedback.
Edit #2: After a lot of trial and error, I nailed it. This worked for me:
\{(.+)\%(\d+)\%.
I left the "change to" blank and in the big Change Format window, I added character style "page", though of course this is going to be different for you depending on how you have the name of that style set up in your CSS file if you're making an e-book too. : )
Message was edited by: 1John5vs7

You should definitely look into full-text search. The idea that Kalman floated is doable, but it require a lot more work on your part. Full-text does a lot work for you, for instance handling inflections, so that a search on "goose" will get a
hit on "geese".
If you have never worked with full-text search before, I recommend to get your hands on this book:
http://www.sqlservermvpdeepdives.com/
http://www.amazon.com/SQL-Server-MVP-Deep-Dives/dp/1935182048/ref=sr_1_1?ie=UTF8&qid=1400851023&sr=8-1&keywords=sql+server+mvp+deep+dives
This book is a collection of chapters written by a number of SQL Server MVPs, and all our royalties goes to War Child International, so you are supporting a good aim if you buy this book.
Chapter 13 by Robert Cain is an excellent introduction to full-text search, although it does not handle Semantic Search added in SQL 2012.
As it happens, my chapter, describes a solution of what Kalman had in mind, although it aimed for the case where you want to permit users to search arbitrary character sequence, and that is not want you want.
Erland Sommarskog, SQL Server MVP, [email protected]

How to search for a set of string in a column

Dear all,
From a client application I have a Text field where the user can enter a text or a sentence for instance like :
"Home for rent"
How to make an efficient search in my table for records containing the search patter ?
Knowing that it could be a set of words like "HOME RENT POOL"
Idea is to make a kind of semantic search based on the words and return records that match.
Any advise is welcome
regards

You should definitely look into full-text search. The idea that Kalman floated is doable, but it require a lot more work on your part. Full-text does a lot work for you, for instance handling inflections, so that a search on "goose" will get a
hit on "geese".
If you have never worked with full-text search before, I recommend to get your hands on this book:
http://www.sqlservermvpdeepdives.com/
http://www.amazon.com/SQL-Server-MVP-Deep-Dives/dp/1935182048/ref=sr_1_1?ie=UTF8&qid=1400851023&sr=8-1&keywords=sql+server+mvp+deep+dives
This book is a collection of chapters written by a number of SQL Server MVPs, and all our royalties goes to War Child International, so you are supporting a good aim if you buy this book.
Chapter 13 by Robert Cain is an excellent introduction to full-text search, although it does not handle Semantic Search added in SQL 2012.
As it happens, my chapter, describes a solution of what Kalman had in mind, although it aimed for the case where you want to permit users to search arbitrary character sequence, and that is not want you want.
Erland Sommarskog, SQL Server MVP, [email protected]

Trex "synonym search" and taxonomy "search from here" issues

Hi Trex experts
I have two minor issues regarding Trex and Taxonomies
Taxonomy
I have created a Taxonomy and need to add the search from here function so its possible to create a more specific search from within one are of the taxonomy.
When I add a layout set where this command is enabled then its not shown.
I guess that its my repository manager for the Trex repository that I need to add a service to, but I cant figure out which one, or is this feature not possible for taxonomies?
Trex
Is it possible to define synonym search entries? Like I need to create a dependency on a bunch of terms, like when a user for example search for vacation then search for holiday should be included automatically in the search result as well. Is there some kind of configuration file where I can add a list of dependencies?
Thanks in advance
Regards
John

Hi John,
Related to synomym you should read this material:
"How to Enable Semantic Search or Search for Synonyms in TREX"
Related to taxonomy, the answer is yes it is possible.
If "search from here" botton is not shown, is due to you don't create an index (search option).
Patricio.

Search Funtion

Dear Sir,
I need to do Search Function and specific some folder for search in EP6 SP14,
How should we do?
Thanks
Vimol

Dear Vimol,
Please refer to http://help.sap.com/saphelp_nw2004s/helpdata/en/40/83505303bd5616e10000000a114cbd/frameset.htm. This talks about the configuration of TREX and anything and everything you would like to know about TREX.
https://www.sdn.sap.com/irj/sdn/go/portal/prtroot/docs/library/uuid/1545e1bf-0d01-0010-a5ab-f80e574423bf - How ToConfigure TREX 6.1 for Efficient Indexing
https://www.sdn.sap.com/irj/sdn/go/portal/prtroot/docs/library/uuid/f0d3be0e-0401-0010-b780-ff7e4e103ea0 - How Toenable semantic search/search for synonyms
https://www.sdn.sap.com/irj/sdn/go/portal/prtroot/docs/library/uuid/d17d18d0-0a01-0010-0db1-f96a947e38a0 - How-to Guide: Searchable HTML Tags
https://www.sdn.sap.com/irj/sdn/go/portal/prtroot/docs/library/uuid/77f6aa90-0201-0010-b681-e013540efb3b - How toset up Web Repository and Crawling it for Indexing
You may also refer to : TREX and its application on SDN to know about the different kinds of search.
If you want to create your own search using KM APIs then refer : How to write a Search Application using the KM Indexmanagement API for TREX
Regards,
Sunil

Integration of Oracle RDF and Enterprise Search

Hi,
Has anyone tried integrating Enterprise Search with the new RDF feature to enable true semantic search?

It would be interesting to hear from people who have tried this in an application.
As an FYI, using RDF query capabilities in Oracle, simple PL/SQL code can be written to do this. Consider search based on a query term X. SDO_RDF_MATCH can be used to find all terms related to X in a given ontology/RDF data set. All these terms can then be used in the 'Contains' operator that is part of Oracle Text, to find all documents in a table that contain any of those terms. We have some sample code to do this, if you are interested write to me at melliyal <dot> annamalai <at> oracle <dot> com. Tighter integration is being planned in a future release as well.
RDF queries can be also combined with other relational queries. An example in Oracle of simple SQL to combine querying RDF with a relational query is as follows:
SELECT emp1.name, emp1.salary
FROM emp1, TABLE(SDO_RDF_MATCH(
'(?name :fatherOf :Matt)',
SDO_RDF_Models('family'),
null,
SDO_RDF_Aliases(SDO_RDF_Alias('','http://www.example.org/family/')),
null)) t where emp1.name=t.name;
The father of Matt is found by querying the family model (represented using RDF) using SDO_RDF_MATCH, and his salary is found out by doing a join with the table emp. Since the output of SDO_RDF_MATCH is a table, it can be easily joined with other relational data.

Release date for new TREX features (esp. guided nav, thesauri/topic maps

In a SAP presentation on TREX, the outlook for future improvements listed the following items :
- combine QBT and EBT
- Thesauri and XML Topic Maps ***
- Taxonomy Import
- Guided Navigation ***
- Additional languages
- Additional Formats
I am especially interested in the items annotated with ***
Is there any information available when these feature will become available.
Will they be part of one of the next Netweaver 2004 stack levels or are these features planned only for the next Netweaver release (2005 or later)
In addition there has been the announcement of TREX Enterprise Search which should allow searching across both structured and unstructured data sources. Any news on that ?
Thanks in advance for your information
Message was edited by: Theo Paesen
Message was edited by: Theo Paesen

Hi Theo,
in general, the features you list are TREX features and have hopefully been presented as such. Features of TREX are not necessarily fetaures of KM. Many SAP solutions use TREX in quite a few different ways. Some for document search (such as KM, PLM DMS and more) some for accelerated table data scans (such as BI, CRM and more).
Thus, you do not find all TREX features being used at the same time in the same solutiuon. And, to anticipate that question, sorry, the direct TREX APIs are still not public.
Point by point:
- combine QBT and EBT
Available in TREX engine
Planned for KM (no date)
- Thesauri and XML Topic Maps
XTM available through TREX Python Extensions
No plan for KM
More info under:
https://www.sdn.sap.com/irj/servlet/prt/portal/prtroot/com.sap.km.cm.docs/documents/a1-8-4/how to enable semantic search or search for synonyms in trex.pdf
- Taxonomy Import
Preliminary workaround available in KM
Full-featured with KM Transport in SPS15
- Guided Navigation
Available in TREX engine
(very simple example at http://help.sap.com "narrow by info class")
Different plans in different solutions, KM not before 2006
- Additional languages
Happening constantly, coming up:
Slovakian, Slovenian, Catalan, Croatian
(That is for TREX engine linguistics. Not necessarily all UIs translated.)
- Additional Formats
Happening constantly, coming up:
HWP2 Hangul 2002 (text only)
ICH new support for Ichitaro 13 and 2004
JBG2 new support for JBIG2 files embedded in PDF files
MIME Mime encoded mail messages
MPP new support for Project 2003 (text only)
PDF new support for JBIG2 files,
PP97 new support for PowerPoint 2004 and v.X for Mac
QP9 new support for Quattro Pro version 12
SHW3 new support for Presentations version 12
SO6 new support for paragraph- and character-level attributes
VCRD new support for vCalendar
VISO new support for Visio 2003
W97 new support for Word 2004 and v.X for Mac
WP6 new support for Word Perfect version 12
XL5 new support for Excel 2004 and v.X for Mac
Enterprise Search...
...is not a TREX feature, technically speaking.
...is a meta search framework.
...will connect to any solution providing a search functionality.
...will connect to as well TREX-using is completely non-SAP solutions.
...will use e.g. KM, BI and MDM as some of its default sources.
...will ship in 2006.
Hope that clarifies,
Karsten

Stored procedure to Insert any type of file in to sql server from the path ?

Hi,
I have a table "FileMGT" and storing files in to disks.
FileMGT Columns :
FileName
FilePath
020324.doc
f:\cfs\Claims Document
5013.tif
f:\Addendums
2790.msg
t:\Treatycfs
Now I would like to store all files from above table in to sql sever by adding content column ( VARBINARY ) datatype.
Please help me do solve this.
Thank you.

Why do you want to store files in SQL Server? If your version is SQL Server 2012, then you can try with FILETable.
Please refer the below link for the same
http://technet.microsoft.com/en-us/library/ff929144.aspx
The FileTable feature brings support for the Windows file namespace and compatibility with Windows applications to the
file data stored in SQL Server. FileTable lets an application integrate its storage and data management components, and provides integrated SQL Server services -
including full-text search and semantic search - over unstructured data and metadata.

Image Files - Within Database, or Use FileStream?

I have millions of photos tied to transactions. There are several photos tied to one transaction. My app pages and displays hundreds of transactions with the photos. I would also like a user to be able to download all of the transaction
information including the photo to a spreadsheet. Currently, the images are stored within the database - is this the best way to go?
Considerations -
I am currently using transactional replication to replicate this SQL database to another location and the database for the images is 25GBs and will continue to grow.
The images do not change but daily hundreds more transaction and images are added to the system.
I know there is a newer method - FileStream. Should I be using this method and FileTable? I know I can't use Database mirroring with this option (is that a concern) and am currently using transactional based replication.
What backup method should I use with the recommended database deign - replication or ?

Consider FileTable for your application.
BOL: "The FileTable feature brings support for the Windows file namespace and compatibility with Windows applications to the file data stored in SQL Server. FileTable lets an application integrate its storage and data management components, and provides
integrated SQL Server services - including full-text search and semantic search - over unstructured data and metadata.
In other words, you can store files and documents in special tables in SQL Server called FileTables, but access them from Windows applications as if they were stored in the file system, without making any changes to your client applications."
LINK: http://msdn.microsoft.com/en-us/library/ff929144.aspx
Kalman Toth Database & OLAP Architect
SQL Server 2014 Database Design
New Book / Kindle: Beginner Database Design & SQL Programming Using Microsoft SQL Server 2014

No module named collections

Hi Experts,
I've installed TREX 7.0(standalone, win32). I'm trying to enable semantic search by following the document at 'https://www.sdn.sap.com/irj/servlet/prt/portal/prtroot/docs/library/uuid/f0d3be0e-0401-0010-b780-ff7e4e103ea0'.
In the step 3., when i try to create a .xtm file using
"txt2xtm.py --input=
<TREX_DIR> opicmapsexampleexample.txt -- output=<TREX_DIR> opicmapsexampleexample.xtm"
(in my case, txt2xtm.py --input=C:usrsapTRXTRX00exe opicmapsexampleexample.txt --output=C:usrsapTRXTRX00exe opicmapsexampleexample.xtm)
I get the import error, 'No module named collections'.
I searched for 'collections.py' file in the entire directory where trex was installed(C:usrsap), but found result is zero.
It requires to activate the python activation as given in,
'http://help.sap.com/saphelp_nw04s/helpdata/en/fe/72f11aae4f754d90f44680283a3634/frameset.htm'
Here, i followed 'Activate the Python extension handler' and in 'Registering the Python extensions' part , I changed "if '0' to '1'" for 'Topic Map' and 'Semantic Search' extensions.
Here're the commands executed in command prompt,
C:Documents and SettingsC5119720>cd C:usrsapTRXTRX00
C:usrsapTRXTRX00>TREXSettings.bat
C:usrsapTRXTRX00>cd C:usrsapTRXTRX00exeextensionsxtm
C:usrsapTRXTRX00exeextensionsxtm>txt2xtm.py --input=C:usrsapTRXTRX00exe opicmapsexampl
eexample.txt --output=C:usrsapTRXTRX00exe opicmapsexampleexample.xtm
Traceback (most recent call last):
File "C:usrsapTRXTRX00exeextensionsxtm xt2xtm.py", line 4, in <module>
 from testbase import TestBase, TestError
File "C:usrsapTRXTRX00exeextensionsxtm estbase.py", line 2, in <module>
 import os, sys, string, thread, threading, time, fuzzypy
File "C:usrsapTRXTRX00exePythonlib hreading.py", line 13, in <module>
 from collections import deque
ImportError: No module named collections
C:usrsapTRXTRX00exeextensionsxtm>
Please suggest how to get this 'collections' module.
Thanks,
Arun

Hi All,
I got the solution for this. I've been executing the scripts txt2xtm.py without using 'python' keyword. The screenshots given in the document doesn't contain this information and it's relevant for a release lower than 7.0. After adding 'python' keyword before the script name, the error has gone.
Best Regards,
Arun

Google calendar in akonadi

How can I use google calendar as an akonadi resource? On the net I find there should be support for it in kde 4.3 using extragear packages, but I cannot find such a package.
Thanks in advance!

toad wrote:At the risk of appearing totally dumb and getting told off for hijacking this thread - will akonadi then automatically update both google and korganizer calendars?
That's the idea. You edit one, and it gets sync'ed in the other.
toad wrote:If so, how do you do it?
Theoretically, you install the resource and select it in the KCM. I am not sure how well it works right now, though.
toad wrote:Just curious and interested in the power and scope of akonadi...
The concept is great. Basically, Akonadi will offer simple APIs for all kinds of personal information management programs to hook into data and syncronize changes. Akonadi plugins for IMAP, POP, Google data, addressbooks, etc. automatically enable all clients to use these mechanisms and provide for search and caching capabilities. Moreover, Nepomuk will be part of the whole thing and allow for semantic searches, automatic tagging, identification of different kinds of accounts (email, IM, ...) with identities, etc.
However, while the akonadi server itself is ready, there is a limited amount of plugins available right now and even less applications support it. KMail for example does not yet use it. Worse, data migration has not been fully worked out yet. This does not mean that you lose your data, but that certain Akonadi compatibility resources use e.g. the standard Maildir or Mbox files for your emails. Conversion to an Akonadi format (a database) is expected for KDE 4.4. Thus, if you use Akonadi now, you will have an extra server running, but few added features.

How to classify a document based on Microsoft's Naive Bayes algorithm?

I am quite unfamiliar with using SSAS as of the moment. I am aware of the classical implementation of Naive Bayes. I have learned about it from the
here. However what I am looking for is a complete walkthrough of how to use this particular algorithm with SSAS.
For simplicity let me assume we are supposed to classify a news write-up as having positive criticism or negative criticism. So for the positive articles we can observe words like
good, awesome, super, recommended, love, like, etc occuring frequently. For negative articles we can observe words like
bad, poor, unsatisfactory, unsatisfied, pathetic, etc mostly. There are only two possible outcomes (positive
or negative), hence, generalizing on patterns is fairly simple.
To start with we have a few write-ups with their corresponding outcomes, which are
mostly in accordance with patterns we've generalized above. If we were to do this without the help of a data mining tool, we would do the following:
Take the first write-up (assume this one is a positive article)
We'd first split the whole write-ups into words.
Remove the stopwords in them, like the, this, that, etc. (Words meant to provide a grammatical structure to the write-up but they occur frequently hence get rid of them). We get a corpus of words now.
This corpus is assigned to the outcome positive. We simply note the frequency of how many time positive appears, and also the frequency of the individual words tending to give outcome
positive.
The next write-up is taken. (Assume this one to be negative).
Steps 2-5 is repeated and the particular frequencies are updated each time.
So once we have looked into all documents, we can actually prepare test cases.
In accordance with the formula above, nc is the no.of times
good actually give the outcome positive. p is a prior estimate (=0.5 since only 2 outcomes), and
n is no.of time positive outcome appears in our corpus.
How can I use SSAS to and go about verifying these kind of test cases manually?
I am a bundle of mistakes intertwined together with good intentions

You posted such a nice question -- and no one answered it.
I would not use Naïve Bayes for this task. There are three other approaches:
1) Use either Association Rules or Decision Trees to determine associated words inside individual documents (treating each document as a transaction).
http://msdn.microsoft.com/en-us/library/ms175595(v=sql.120).aspx
2) Use Integration Services Term Extraction Transformation and Term Lookup Transformation to create this application in Integration Services (Data Flow).
http://technet.microsoft.com/en-us/library/ms141713.aspx
3) Use Semantic Search.
http://technet.microsoft.com/en-us/library/gg492075.aspx
Mark Tabladillo PhD (MVP, SAS Expert; MCT, MCITP, MCAD .NET) http://www.marktab.net

Semantic Search

Similar Messages

Maybe you are looking for