Word Count

[Dynamedix]

How do I count words in an English PDF document?

[Dave Merchant]

There's no built-in word count tool. In Adobe Acrobat you can use a console JavaScript:

var cnt=0;

for (var p = 0; p < this.numPages; p++) cnt += getPageNumWords(p);

console.println("There are " + cnt + " words in this file.");

[tobywilmington]

When i use the script :

var cnt=0;

for (var p = 0; p < this.numPages; p++) cnt += getPageNumWords(p);

app.alert("There are " + cnt + " words in this file.");

I keep getting word count 0 on all PDFs?

Any ideas?

[try67]

The document is probably scanned and was not OCR-ed, so it doesn't contain

any actual words in it, just images with text on them...

[tobywilmington]

The PDF is full text, when i copy and paste to word the count is 1,052 words - Im just wondering whether i need to edit the script at all?

Script i am using -

var cnt=0;

for (var p = 0; p < this.numPages; p++) cnt += getPageNumWords(p);

app.alert("There are " + cnt + " words in this file.");

[try67]

No, that script comes from the JS API Reference and should work. Can you

share the file?

On Tue, Nov 19, 2013 at 10:15 AM, tobywilmington

[tobywilmington]

Hi Gilad,

It is happeneing with any PDF, ive just tried the ipad mannual - http://manuals.info.apple.com/MANUALS/1000/MA1595/en_US/ipad_user_guide.pdf

I have this now set up so when i open a pdf it runs the Java automatically but just doesnt seem to pick up words?

Cheers

Toby

[try67]

What do you mean, exactly? From where are you running this code?

[tobywilmington]

In Adobe Acrobat pro - im using this JS as a saved script, i think it runs through the debugger?

[try67]

So you're running it from the console directly?

[try67]

Also, are you selecting all of the code when you run it?

[Guest]

I'm attempting to use the same JS code in Adobe Pro version 10 and receive the "no words found" response. What and how did you finally resolve your issues? If you don't mind sharing....

Thanx

[try67]

That usually means that your file contains only images, no real text...

[try67]

I ran it on the iPad Manual file and the result was:

[tobywilmington]

When i run it directly in the console - the console replies as below

Sorry to be a pain!

[try67]

Yes, that's what I thought... You have to select all of the code (with the mouse) before running it.

[tobywilmington]

So sorry! Rookie!

Massive thanks

Out of interest - does this script have the potential to count images? that would be a huge help to myself

[try67]

It's a common mistake to make...

No, JS has no access to the images in the file, at all. Only to the textual content, and even that's limited.

[David Cardillo, PRH]

anyone know why the Acrobat console reports such wildly different word counts than other tools (e.g. Word)?

What is this script counting in addition to word breaks? I get differences of a couple hundred to over a thousand extra "words."

[try67]

For example, Acrobat splits hyphenated words into two, so "right-handed" will count as two words, while Word counts it as one. If you can share a sample file that demonstrates this issue we can look more closely into it, but if Word behaves the way you're looking for, just export the PDF to Word and do the count there. That's the easiest solution.

[David Cardillo, PRH]

try this PDF. 

40,911 words per Acrobat

39,735 copy/pasted into Word

I'm very interested to find out where the discrepancy comes from, because then it can be corrected against.

[David Cardillo, PRH]

initially I thought it might be the metadata. The metadata of the above PDF, saved from Acrobat, comes to 37 "words" in Word. It's 157 copy/pasted from Word into Acrobat.

Big difference, but nowhere near the total difference.

[try67]

Metadata info is not included in the word count.

[try67]

Not the easiest file to work with... More than 12K words just on page 1! Can you find a smaller file that demonstrates this issue? Also, notice the page is cut off at the end of the page. The last line is duplicated on both pages, which might help explain the differences you're getting in the counts.

[David Cardillo, PRH]

We work with full-length novels. The word count is not unusual. I grabbed that file because it's in the public domain. I cannot share the full text of the books we work with.

I obtained that file from the Project Guttenberg website. It was HTML text that I saved as a PDF. I took off the headers and footers as they weren't part of the actual text. Whatever text is visible in the PDF, I did Select All > Copy, then Paste into Word, so the text is the same (or should be, let me know if you want the Word file that was created).

To your point, you could take _any_ PDF upwards of 1,000 words, run the word count in Acrobat, then paste the text into Word and run a word count there.

In my two examples — 1,000 words vs 46,000 words — the higher word count title resulted in larger number of "extra" words, though it was a smaller percentage of the total; while a smaller word count had a fewer number of "extra" words (naturally) though a higher percentage difference of the total.

Specifically: 1,000 words in Word vs 1,200 in Acrobat = 200 extra words or 20% variation; vs 46,000/47,000 or 1,000 extra words with only 2% variation.

The only reason I included the stats about the metadata is because the results are so wildly different: 37 vs 157! That's not a lot of text for such a large variation.

Perhaps that could be a clue as to what it is about the text that's causing it to be read so differently.