Hi Frank
Using SOLR for global search does have some known limitations around completeness of results and performance
However searching content in pdf files isn't available with other Totara features yet . Possibly the AI features such as summarise might be able to be developed further.
I can't find any documents in Totara around setting this up but it sounds like you have it setup and have indexed the files OK it is the search results that aren't as accurate as expected. A partner using this may help provide tips for better results
Are there any messages in the SOLR log outlining any problems with pdf indexing?
Does the indexing work with text based . simple, small files - as the indexing doesn't handle image based, , complex formatting or large documents well
Regards