The text search in pdf file is a quite slow process.
Normally readers do a page scan, char by char and build a word dictionary. Such a process is weird and sometime rise to wrong result: sometime glyph on the page are represented by diffrent unicode character or kerning between char is uncommon and the word parser read them as separated letter instead a single word.
1 - if possible parse the pdf file on a server, extract text, make the text corrected with a syntax checker, send the text to the phone aside the pdf file. You could use the RadaeePDF Server.
2 - do the same procedure on the mobile device, at least the text extraction, you could use RadaeePDF SDK.
3 - do a double loop, one on each document, one on each page from each document.... as you may imagine, it will be really long!
During next few days I'll evaluate a different approach with a pdf file digger... I'll be back to you soon.