Tovek

Tovek Server 5


current documentation download




version 5.30

NEW FEATURES AND CHANGES:

  • Categorization feature
    Documents are automatically sorted by given criteria into categories. Queries, field values and manual input can be used. See documentation for details.

  • Metadata feature
    Users can interactively add data to documents (notes, images, links, access counters etc.). Metadata structure can be configured by administrator (what data types, access to metadata etc.). See documentation for details.

  • User credentials cache
    User credentials are stored in cache to speed up authorization.

  • Caching authentication
    Information about the authenticated user can be stored in cache, so subsequent requests from same user can be processed without authentication.

  • Extended LDAP authentication
    LDAP server support was completely rewritten.

  • Anonymous and pre-authenticated user support
    Server can process requests in anonymous mode. Also pre-authenticated users can be used. In this case server does not authenticate users’ identity and trusts to sender.

  • Possibility to activate profiling from external scripts
    Profiling can be activated from external scripts by creating specified file.

  • Storing profiling results and profiles into SQL database
    Profiling results and profiles can be stored in SQL database. MS SQL, Oracle, MySQL and PostgreSQL databases are supported.

  • On-demand profiling
    Possibility to send document text from the client into profiler added.

  • Faster searching
    idField, sourceField and dbuniquekeys parameters allows more efficient source identification. This speeds up searching, especially if a lot of result lines are requested.

  • Document level security and integrated windows authentication
    Document level security can be combined with integrated windows authentication.


BUGS FIXED:

96 If source name is invalid, all sources are searched
98 Default for profile's locale changed from utf-8 to uni
103 New function GetSourceInfo added
103 Internal function GetInfo on all sources added (session not needed)
104 Document from existing sources can still be retrieved even if some other sources are missing
105 SaveHistory in ProfilerDocFeed fails with exception if document was not loaded
108 While retrieving documents by source and key, the key was not escaped properly. Now \ and ` are escaped
109 Single threaded GW support (Lotus Notes gateway)
110 Possibility to switch off dynamic highlighting (-dynamicHL = false)
111 Key array converted to source query before search (-keyArrayAsQuery = true)
113 New method GetUserInfo of the Searcher returns information about TS user account
119 <target> tag in PrfDocFeed config is now optional
121 Licensing for advanced security added
127 Inline highlighting if hlInline = TRUE specified in config
130 Support for id array in ResultNew (categorization changes) added
131 New configuration parameter added to enable start of the LDAP component without any connection to LDAP server
132 kvoop.exe processes stays in memory after document view
134 Wrong source name sent to plugin
135 Field names array in ResultNew allocated only if count > 0
140 Some characters causes syntax error in freetext query
147 Preauthenticated users support
148 Cannot get result list if both docid array and search handle are specified
169 Case insensitive user names are now supported
171 Highlight offsets if characters were not recomputed to correct byte value (for utf-8 or ucs-2 output)
172 Support for inline highlights in profile result if output stored "by user" (history = none)
176 Profile store file output format configuration added
178 User token has no TOKEN_DUPLICATE rights, DuplicateTokenEx used to fix this
179 Wrong handling of user token in TSP_UserCreate & bug in TSP_UserFree fixed
180 Buffer for message was too small during sending profile results via e-mail
181 Rank value fixed in ResultNew for list of DOCIDs
183 $TITLE is not replaced by value (insufficient buffer) during sending profile results via SMS
184 Sequence for GetDocumentRequest fixed (user was in wrong position) in RSS interceptor
185 ${user} can be used in templates for RSS, e-mail and SMS
189 Results with no highlights are also stored
196 Some session members are not reset to default in SessionFree (memory leaks)
199 Search is slow if sortspec is used
201 Search fails with exception if scoreprecision = 1bit and field condition used and zero results found
202 Sortspec limited by configuration also in VDK V6 plugin
204 Wrong number of working threads if license has limit 0 (= unlimited)
206 Using X500Names as alternative to normal names plus its configuration
222 Maximum number of wildcards in user permission is ignored
223 Plugin doesn't start if too many databases specified
228 List of hidden fields in database config
229 New member m_bSortable in info sent to clients (TSP_FieldInfoRec; true -> field can be used in sortspec)
230 If allowed sortspec not specified, all fields can be used in VDK V6 plugin
270 Windows login in preauthenticated mode with no credential caching causes exception
271 Profiles that could not be stored are automatically removed from the profiling engine
272 New method of retrieving document added - TSP_DocType_XML - serves for example to retrieve highlight info for Acrobat reader (viewing PDF files in original with highlights)



version 5.11

NEW FUNCTION AND CHANGES:

  • You can define conditions for detecting new documents for profiling in the configuration.
    The filter and sort tags can be used in the <collection> tag in the ts_prffeed.xml configuration file. The filter tag defines a condition in VQL format, which can contain links to field values as follows: ${filed_name=default_value}. This default value is used while first running. The sort tag contains sort rules (in VDK sortspec format). The initial settings after the installation are <filter>_DOCID > ${_DOCID=0}</filter> <sort>_DOCID asc</sort>.
    ATTENTION! As the configuration is stored in XML, you must use html entity instead of > and < characters!
  • MS Office 2007 support
    DOCX, PPTX and XLSX formats
  • The Czech locale was supplemented with stemming according to rules.
    For transformation to basic form are in addition to dictionary used rules, which mostly solves problem with words not present in the dictionary.
  • Indexing and searching are installed as separate NT services.
    In order to increase the system stability, the indexing was separated into a special service.
  • The profiler configuration was divided into more files.
    ts_profiler.xml – common configuration
    ts_prffeed.xml – profiled sources configuration
    ts_prfstore.xml – result store configuration
  • Demo collections were moved from <install_dir>/vdk/610/demodata folder into <install_dir>/data/demodata folder.
  • If the database initialise is not successful, the server initialise is not terminated.
    If the database tag contains the required_init="false" attribute, the server will start even if the relevant database is not initialised.
  • Parameters common to all databases using the same plugin, can be placed into the TS_Plugins.xml file.
    Any parameters stated in <dbparam> tags can be stated as <plgparam> tags in the TS_Plugins.xml file. These parameters are shared among all databases, which use the relevant plugin.
  • The groups are supported in the profiling.
    If the users are authenticated against Windows domain, a group can be used as a user in a profile.
  • Profile was supplemented with Locale and Userdata items.
    A profile can contain query locale. It is UNI locale by default. UserData are any binary data, which can client application store together with the profile and use subsequently on editing or viewing.
  • The source tree in the Java client was supplemented with a root knot with a server name.
  • Internal Java client field with source name was renamed from Source to TS_Source.
  • Updated developer and administrator documentation and examples.


SOLVED PROBLEMS:

  • An incorrect request for server containing an empty enumerative value (enum) caused memory exception.
  • Clustering did not take into consideration maximal number of retrieved documents setting (all documents meeting query were clustered).
  • A request for found words list passing occasionally retrieved stems instead of words
  • The JobNew function retrieved the job ID in different format, than requested in other functions.
  • A trailing space in a path to collection in TS_Databases.xml caused the collection authentication failure.
  • Text of some documents couldn't be withdrawn from profiling results.
  • If TS_ProfilerDocFeed module doesn't have access to folder with stored history, all documents were send to profiler repeatedly.
  • User names were case sensitive.
  • Conditions for fields in profiling queries did not work always correctly.
  • Highlights flag in a request for profiling results was ignored (the words are always retrieved).
  • The clusters in the Java client did not expand when clicking the [+] icon.
  • The Java client did not remember source tree "expanding status".
  • If a source in the Java client couldn't be opened, the whole client finished with error.
  • The Java administrative client ignores the Enter key in dialogs.
  • The User has no permissions for "TS_Profiler:Profiles:Admin" message was recorded to log even if this permission is not required for the requested function.
  • The wrong profiles job synchronisation during the profiler start up.
  • Some error messages from profiler were not send to client application.
  • A user name authenticated against Windows domain did not contain a domain name.
  • Automatically created folders were not established, if there were more than one level to establish.


expand all collapse all

nevyplnovat


© Copyright 2010 TOVEK, spol. s r.o. | private data protection