-- Tag PDFs that Need OCR
-- https://c-command.com/scripts/eaglefiler/tag-pdfs-that-need-ocr
-- Summary: Adds the “NeedsOCR” tag to the selected PDF files that do not have any text.
-- Requires: EagleFiler
-- Install Location: ~/Library/Scripts/Applications/EagleFiler/
-- Last Modified: 2019-05-24



property pMinimumTextLengthThatCounts : 1
property pIgnoreWhitespace : true

tell application "EagleFiler"
    set _records to selected records of browser window 1
    repeat with _record in _records
        if _record's universal type identifier is "com.adobe.pdf" then
            with timeout of 5 * 60 seconds
                set _string to _record's text content
            end timeout
            if pIgnoreWhitespace then
                set _string to my removeWhitespace(_string)
            end if
            if length of _string < pMinimumTextLengthThatCounts then
                set _oldTagNames to _record's assigned tag names
                set _record's assigned tag names to _oldTagNames & {"NeedsOCR"}
            end if
        end if
    end repeat
end tell

on removeWhitespace(_string)
    set _string to my replace(_string, " ", "")
    set _string to my replace(_string, tab, "")
    set _string to my replace(_string, return, "")
    return _string
end removeWhitespace

on replace(_string, _source, _replacement)
    set AppleScript's text item delimiters to _source
    set _items to every text item of _string
    set AppleScript's text item delimiters to _replacement
    return _items as Unicode text
end replace