Executing Preview from a command line and use the find instruction

The AppleScript dictionary support in Preview excludes any means of finding, or highlighting the PDF content. That is something you would need to do interactively in Preview once the PDF is opened.

Is the PDF display a necessary part of this, or incidental to locating the text?

Spotlight can query a PDF file, or there’s a pdftotext tool in the Xpdf open source package. (related)

On the AppleScript-based path you’re presently on, here’s an earlier discussion:

… Applescript to find a text string in a PDF - Apple Community

Thommy_ wrote:

@MrHoffmann: Thank you. But Spotlight only finds metadata. I need them in the text itself.

Spotlight indexes both data and metadata.

If you couldn’t query for text located within a file, then Spotlight wouldn’t be all that useful, no?

The limitation here being how fast the data can get indexed and added to the index, assuming a very-recently-arrived file. That indexing can be encouraged with the mdimport command, and the mdls command can be useful for seeing what metadata (as differentiated from data) is collected for a particular file.

General syntax:

mdfind -onlyin /path/to/file “query”

In this case, look at kMDItemTextContent possibly combined with other keywords for the text content.

Syntax is king.

I have some PDFs on my Desktop and a few of these contain the same text content. I want to see which PDFs contain the case-insensitive word "graecis." The only modifiers supported are c for case-insensitive and d for insensitive to diacritical marks. There is no w modifier.

odin: ~/Desktop % mdfind -onlyin ~/Desktop 'kMDItemTextContent == "graecis"cd' *.pdf
/Users/viking/Desktop/test.pdf
/Users/viking/Desktop/test_pdfx3.pdf
/Users/viking/Desktop/text_svg_13_1.pages
odin: ~/Desktop %



If I have changed "graecis" to "*aeci*", I would have obtained the same result.

I just tested that syntax on my macOS Sonoma 14.2.1 installation and it works just like it has for several releases of macOS before it. When you use -onlyin ".", mdfind will be recursive below that filesystem location, and *.pdf would restrict it to only that file type. That is why I expressly limited the scope to my Desktop.

Use my syntax exactly as shown and be specific about the -onlyin argument.

The File Metadata programming guide only offers c and d as modifiers. Nothing there about w, even though the mdfind(1) man page shows an example for -interpret that uses an undocumented w modifier in how Spotlight tool would form the search — except you are on the command line and I never use the -interpret switch.

If you are passing this mdfind against PDFs that were scanned, and not OCRd, they are PDF wrappers around an image and the kMDItemTextContent cannot match text against an image.

I see what you mean now trying my own ~/Downloads folder. The following restricts your results to PDF. Remember the c modifier allows for case-insensitive searches which is why hannover appears below. Remember, it is recursive, so it will drill into sub-folders…

mdfind -onlyin ~/Downloads 'kMDItemTextContent == "hannover"cd' * | grep pdf

or:

mdfind -onlyin ~/Downloads 'kMDItemKind == "*pdf*"cd && kMDItemTextContent == "hannover"cd' *

And if you do not want recursion, there is the AppleScript approach that you can copy/paste into Apple's Script Editor… 🧐 and run.

use framework "Cocoa"
use AppleScript version "2.4" -- Yosemite or later
use scripting additions

property NSString : a reference to current application's NSString
property NSWorkspace : a reference to current application's NSWorkspace
property NSURL : a reference to current application's NSURL

set query to "kind:pdf AND \"Hannover\""

set thisFolder to POSIX path of (choose folder) as text
set basefolder to ((NSString's stringWithString:thisFolder)'s lastPathComponent()) as text
(NSWorkspace's sharedWorkspace)'s openURL:(NSURL's fileURLWithPath:thisFolder)
set result_code to (NSWorkspace's sharedWorkspace)'s showSearchResultsForQueryString:query
-- close the opened folder but not the search results folder
tell application "Finder" to close Finder window basefolder
return



And a pleasant weekend and Frohe Weihnachten to you.

I forgot about the German locale and understandably, kind:pdf would not work. Just as a Spotlight search for a Finder tag cannot be tag:xxx but rather kMDItemUserTags:xxx.

Spotlight info:

man mdfind