Document the relevant score

[PerrineGilloteaux]

To be added somewhere on the documentation (maybe the result page it self). Also open to discussion for modification.
Relevance score is computed buy search api full text. In short it allows to search the filed below with some weight for some fields (all at 1.0 or 2.0 for now) .

Example: Searching "test" http://biii.eu/mevislab-xmarkershortestpath-module-dijsktra-shortest-path returns a score of 7:
because:

• Additional keywords contains twice "Shortest", score = 2*1.0

• Title contains twice shortest, score= 2*2.0

• Documentation link title contains also test score = 1*1.0
  TOTAL relevance score = 7
  Idem with http://biii.eu/test-java3d scoring:

• Additional keywords contains tester, score = 1*1.0

• Title contains Test, score= 1*2.0

• Documentation link and downlad page contains also test score = 2*1.0
  TOTAL relevance score = 5

IMPORTANT: WHILE the search should also be performed in synonyms for now it is not (see LSFM) , new issue for that

[PerrineGilloteaux]

ALSO when searching the following "processors" are applied :

• ignore case (apart for title and links)
• ignore accent or other non -ascii characters
• ignore .-_

[PerrineGilloteaux]

Added some short doc on the webpage itself

c631085 and c8c44a0

To be noted also some emty lines with relevance score, already mentionned in another issue

[PerrineGilloteaux]

Note also that the folliwng list of commit is related to the same config file and

• show some exceirpt (where the keywords were found)
• order the type of search results, first software, then dataset, then training material
• add some explanatory message about te results

https://github.com/NEUBIAS/bise/commits/b57fdb7e168b80fd8b6d817d7f823151a471129e/config/sync/views.view.testsearch.yml