Closed andre-hohmann closed 4 years ago
Truncating the search query left and right is counterproductive in how the search engine works. Basically, it works, but it first becomes a large OR of all indexed strings that contain the substring, which is then searched for. The solution to this is to customize the tokenization. That is, for SrKMai_1438157967_2
, not only srkmai_1438157967_2
is indexed, but also srkmai
, srkmai_1438157967
, 1438157967
, 1438157967_2
and 2
, so that query e.g. after 1438157967
is possible. Adaptation of this is the task of the search engine preprocessing.
I am ok with that, too! For me, it is only important that the user finds the processes easily.
Problem
In the task list, processes can be found only, when the complete process title is entered. For example.
Practically, users search often only the identifier, especially, when several processes of a newspaper should be shown. For example:
The TSL_ATS and the suffix in the process title is often not known by the users and a search not possible.
Solution
In Kitodo.Production 2.x process titles were truncated left and right sided. That should be possible in Kitodo.Production 3.x, too.