Fix Elasticsearch ecm:binarytext mapping

The mapping for the binary fulltext :

      "ecm:binarytext*" : {
         "type" : "string",
         "analyzer" : "fulltext"
      },

is not taken in account, because wildcard in field definition is not supported.
As a result the ecm:binarytext is taken as a Token, which has size limitation (see NXP-16807).
This field should indeed not be indexed as all by default, just being part of _all is enough.