Noodlesoft Forums

Guidance would be appreciated here. I'll do my homework once I have some bread crumbs to lead me down the right path.

I download 12 monthly bank statements, and then append them into one sequential pdf file. Total is about 50-60 pages per year.

Each month includes several repetitive and unnecessary pages. For example, there is always a template page for balancing the checkbook (which I never use), and depending on the number of pages in the statement there will often be a blank page containing the single sentence, "This page intentionally left blank."

I am able to use content-matching tokens successfully. For example, a match for "intentionally" easily identifies the blank pages.

Once I have a match, is there any way (via Hazel, CLI, Applescript, Automator) to delete the 8-12 matched pages interspersed throughout the longer pdf file?

* Scan contents of page 1.
* IF page contains "intentionally", THEN delete page.
* Next page ...
* Continue UNTIL last page

This is a relatively trivial example because it doesn't take too long to do this manually in Preview. OTOH, we're always looking for ways to identify and eliminate repetitive (and error-prone) tasks, and this process would be invaluable for more complex situations, or much larger files.

TIA

Ed

After my original question, an inelegant, but probably workable solution occurred to me.

1. (Automator workflow) Burst the monthly PDF (4-5 pages per month x 12 months =~ 50-60 pages) into individual pages, thus creating 50-60 one-page pdf documents.

2A. (Hazel): IF THE FOLLOWING... USE content matching to identify those files that includes the word "intentionally"

2B. (Hazel) ACTION: Move matching files to 'discard folder'.

3. (Automator workflow) Append remaining pdf files into one sequential file.

Hmmm -- that doesn't seem so hard... I'll try it tomorrow and report back.

Ed

OK - I've given it the old college try, and more, over the past two days. I can make all the pieces work separately, but for the life of me cannot make them play nicely together.

Here's what works:

1. I place twelve (12) monthly bank statements into a folder. Each file is 2 to 4 pages.
2. A Hazel rule runs an Automator workflow to split each statement into separate pages.
3. The next Hazel rule matches contents which are found only on 'junk' pages. Once identified, these pages are moved into a 'discard' folder.

PROGRESS TO DATE:

The above workflow retains all the 'good' pages in the previous folder, labeled numerically. ALL (?) that needs to be done is to merge them into a single PDF file, thus collecting all the statements from one year into a single document.

Easy, right? I wish!

This works: Create an Automator.app (Append-PDF.app) that combines the files. Open (manually) the folder, select all, drag to the Append-PDF.app icon on the desktop, and voila! I'm done!

HOWEVER... AFAIK, Hazel cannot process an Automator APP; it requires an Automator WORKFLOW. Simple enough...

I have converted the app to a workflow, which again, works fine when I run it manually. But when I incorporate the workflow into a Hazel rule, it executes 12 times , once for each file in the folder. I have tried tuning up the Automator workflow to 'get folder', 'get folder contents' , merge pdf, which again, works fine (albeit, 12 times). I cannot seem to create a Hazel rule which sends ALL of the files to Automator, ONCE, and then stops.

It would take me another two hours to document all the things I've tried -- rather than do that, it's time to wave the white flag and see if anyone has the answer. (It's probably embarrassingly simple, but I can't figure it out.)

Suggestions? Comments? Help... !

Ed

OK.… !!!

I got it working -- combination of Hazel and Automator. Whew!

I notice that several people have asked the same, or similar questions in the past. Whether or not they got it working, I have never seen the workflow posted anywhere that would solve this problem.

I found several gotcha's that were (much) less than obvious. One appears to be an apparent bug in Automator. If not a bug, then the 'obvious' workflow for identifying the path to save items in a given folder is very inconsistent and misleading.

I'm tired -- it's late, so I won't write up the details tonight, but I promise I'll do this over the weekend. I hope others find it useful.

ejg

ejgallaher wrote:I won't write up the details tonight, but I promise I'll do this over the weekend.

That was written in 2014

Any chance anyone figured out how to make this work smoothly?