#255: BibMatch: convert to simple search method
--------------------------+----------------------
  Reporter:  jlavik       |      Owner:  jlavik
      Type:  enhancement  |     Status:  in_merge
  Priority:  major        |  Milestone:
 Component:  BibMatch     |    Version:
Resolution:               |   Keywords:
--------------------------+----------------------
Changes (by jlavik):

 * status:  in_work => in_merge


Comment:

 {{{
  BibMatch: convert to simple search method

     * BibMatch now defaults to simple search method instead of
       advanced search (fixes #255):

       ** -q, --query-string option now expects an Invenio search
          query or a query-template, it also supports old-style
          querystrings.

       ** When -m, --mode is specified the search will fall
          back to advanced search in given mode.

       ** All queries are performed with timeouts, as introduced
          in invenio_connector.

       ** Added support for cases where several datafields of the
          same type exists in the records. The different values
          for each field will generate unique search queries
          using the Cartesian product of all the values in the
          query.

       ** Introduced the concept of "query-completeness" which
          indicates if all fields specified in matching query can
          exists in input records. If a particular field cannot
          be found, the match is deemed "incomplete", which will
          result in a fuzzy/ambiguous/non match regardless.

       ** Fuzzy matching has been extended to perform set
          intersection of results for every field specified in
          matching query. Various options for the fuzzy matching
          can be tweaked in invenio.conf.

       ** If number of search results exceed 10 it will be
          treated as a non-match.

       ** Added option to sanitize final queries WRT double quotes

       ** The Querystring class have been completely revamped to
          conform with these new changes.

     * Batch output files suffixes are now named instead of numbered
       (i.e. filename.matched instead of filename.0)

     * invenio.conf has been populated with various configuration
       options to BibMatch. These options are mostly related to fuzzy
       matching, but query-templates are also maintained there.

     * Updated regression tests and documentation.

     * Several other minor additions and fixes.

 }}}

-- 
Ticket URL: <http://invenio-software.org/ticket/255#comment:3>
Invenio <http://invenio-software.org>

Reply via email to