Migrate fulltext (PDF) search from Lucene to PostgreSQL

[koppor]

Currently, JabRef employs two different search backends: PostgreSQL and Apache Lucene. PostgreSQL is used for the search within the library (.bib file) and Apache Lucene is used for the fulltext search of PDF files.

It turned out that it is really hard to get "contains" search working properly in Lucene.

the query must be tokenized using the same tokenizer as during indexing to extract words from the query and look into the index for these words.

This will fix #12685

Working on handling following variants for the same word is as hard in Postgres as it is in Lucene:

• Düsseldorf
• Duesseldorf
• D\"{u}sseldorf
• Dusseldorf

[GuilhermeRibeiroPereira]

Hi! @raquelgraos and I are looking for a feature to implement as part of a university project that will run for about 30 days. Do you think this feature would be feasible and appropriate?

[koppor]

Hi! @raquelgraos and I are looking for a feature to implement as part of a university project that will run for about 30 days. Do you think this feature would be feasible and appropriate?

I don't know about your background, experience and absolute time. If you are master students, having 40 a days each person, then it is easily doable.

All "medium" sized projects at Candidates for University Projects (view) should be doable within 30 days if you have some software engineering background.

[raquelgraos]

Thank you for the feedback!