Renaming PDF files with RO

THE_DREAM_FACTORY wrote:

You seem to be on the brink of cracking the code! This will for sure help, it is however putting the name twice in certain files where there are two RO numbers. Many of these will contain a second page where the number is listed again with a /2 at the end of it. Is there anyway I can make a rule in the code where it only allows the code to be written once per file?

add -m1 to grep (-m num: Stop reading the file after num matches)

echo -n $(grep -m1 -oE '656[0-9]*[/ 0-9]')

We only see a screenshot, so can't test what you posted.

However, see if the Shortcut that I posted works

(It worked on one test file I tried (I converted your screenshot to a PDF to test)

>can someone please explain to me in layman's terms how I would select those operations?

>IE "Select Folders" , "Get contents of folder" "Repeat with each item" , etc.

>Then how to run the script shown below?

Open Applications->Shortcuts and search for the Action in the the App

For info on how to use Shortcuts, see: Shortcuts User Guide for Mac - Apple Support

This Shortcut may work (I only tested on the file you posted).

There is no file testing (assumes ALL files in ONE folder are PDF's)

TEST on a folder with a few copies of your PDF's

If it works, Duplicate your folder and run only on the copies.

In addition to checking if the file already exists, this also checks to see if grep finds 656[0-9]*

**Test this on copies, not live data**

Here's the link: https://www.icloud.com/shortcuts/0ba443c85f6d4598831cf8b794300b08

and here's what it looks like:

Still recovering from the dentist visit yesterday, but because these PDFs are generated from an IBM AS/400 report writer, there is no need use the Get Text from PDF action, as one can get it directly from the individual PDF. In production, one would omit the egrep -H flag which lists the files with the result:

The regular expression says capture a first seven-digit number match that may be optionally followed by a forward slash and another number sequence. If the latter exist, they will be captured too.

I also constructed a Zsh array containing the range of valid serial numbers (without the /1), and as I obtained the REGEX capture serial number, I stripped the /1 and then checked if it was in that valid serial range. I also converted the /1 to _1 for UNIX filesystem purposes when constructing the output filename. One would have to check that the intended serial number rename file does not exist before the rename though to avoid name space collisions, and choose some fallback naming convention when duplicates are detected. Or, the mv command has a -n option that prevents overwriting duplicates.

Tony,

The OP does not indicate if they are on Monterey or later, so Shortcuts may not be the drug of choice here.

There are some challenges to be acquainted with in this post, so not being critical of your submission:

If these PDFs were saved from an application, or even OCRd, your approach may work, but it is more complicated as that Get Text from PDF action is a line scanner, so any numerals that appear on the same line before the desired serial number will appear before it in a single column text list output. And what does

"the number varies in its location from time to time but is however always in the top right corner"

mean, and how hard is it to get it right without examples of those location variations? Is the nnnnn/n the entire serial number or only digits preceding that slash?

If these are all scanned PDFs then there will be no text to be retrieved from the PDF as it is a PDF wrapper around an image, though you can use the Extract Text From Image action which does work on obtaining the text from a scanned PDF. Tested that.

When I built that solution that you linked to earlier, I had several downloaded PDFs and corresponding images to work with to get the code accurate, and there were several variations in those PDFs/Images that made the work harder and the REGEX uglier.

So, I built a Shortcut solution that for a few hours was capturing the serial numbers just fine from the two images above converted to PDF that I put in a folder, but under no circumstances would a rename work, though it worked fine in an external shell script. Then, out of the blue, the unchanged Shortcut code stopped capturing the serial number from the OCR stream and ignored print statements in the Run Shell Script that also had been working.

With that, and after a reboot not fixing anything, I simple gave up as my time has a cost.

OK, so if you click this link, you can open the Shortcut in your Shortcut App:

https://www.icloud.com/shortcuts/d8a5566c7f4c45bb85c6499bf92eb4da

I did very limited testing. This Shortcut only works if the PDF contains OCR info.

Test on a few copies in a test Folder.

Best I can do, but at least worth a try.

Then my Shortcut doesn't work.

The only way I can think of getting it to work is if the reference number was always a fixed number of digits (like 12346567), with or without a slash and 1 digit (1234567/2), or at least in a range (like 6 or 8 digits , with or without a / with a fixed range of 1 or 2 digits)

THE_DREAM_FACTORY wrote:

Definitely not a big deal, but if you knew a way to remove the / from the files, just for the sake of cleaning up.

This will only include digits up to, and not including '/' (so any digits after / will also not be included)

echo -n $(grep -m1 -oE '656[0-9]*')

OP wants to exclude the ‘/‘, and digits always begin with 656 so '656[0-9]*' is what I used . While mv -n works to avoid a clobber, I always prefer to test for the file’s existence as it leaves open the possibility of editing the script to add to the name for any dups (i.e. 6561234-dup)