PDF Classifier (Document Classification Based on Classification Rules) – classify and sort incoming pdf documents and scans based on easily created classification rules in CSV based on the set of keywords with optional OR and AND logics

If you need to extract data from a lot of PDF documents coming from different sources then the best way is to first sort them by the vendor. We’ve created a PDF Classifier tool that is available in both cloud and on-premise versions of our PDF Extractor API.

1. How it Works
2. Create and Test Classification Rules
3. Test Classification Rules on Folders With PDF Documents
4. Test Classification Rules on Scanned Documents
5. Copy Rules as JSON
6. How to Use PDF Classifier in PDF.co

How it Works

• Create rules as CSV (comma-separated values) where every row has the following columns: classname, OR or AND logic (OR is used by default), keyword1 or phrase1, keyword2 or phrase2,…
• Test these rules on your sample PDF files
• Generate JSON request for use with PDF.co or just save rules as CSV and pass the link along with all requests
• Use pdf/classifier endpoint in PDF.co (cloud) or API Server (on-prem)
• pdf/classifier endpoint will return detected class for input PDF, JPG, PNG, or TIFF document

To make it easy to quickly test, maintain, update your classification rules we’ve created the classification rules testing tool that is available as a part of the PDF Multitool desktop app (download page is here).

Here are the short demos of what you can do with this tool (it works fast and no Internet is required because no files are uploaded).

Create and Test Classification Rules

Use the spreadsheet-like interface to define new classes with rules, use plain text, use regular expressions and quickly test rules to see how they work on your PDF documents.

Test Classification Rules on Folders With PDF Documents

As the ultimate goal is to sort PDF files in a batch, you can test classification rules on folders with PDF files to see which class every file will produce.

Test Classification Rules on Scanned Documents

You can also test classification rules on scanned documents as well.

Copy Rules as JSON

You can save classification rules into a CSV file or you can simply copy ready-to-use JSON request. You can use this request with PDF.co (cloud) and API Server (on-prem version of PDF.co).

How to Use PDF Classifier in PDF.co

In this tutorial, we will demonstrate how to use the PDF Classifier in PDF.co. To follow along, you can download the file here. We will use both PDF Multitool and PDF.co to showcase this functionality. If you haven’t yet, you can download the PDF Multitool here.

Step 1 – First, download and open the file in the PDF Multitool Classifier Test Tool.

Step 2 – Run the Test Rules.

Step 3 – Click on the Copy Rules for PDF.co or API Server.

Step 4 – Paste the JSON in your favorite text editor and copy the rulesCSV value.

Step 5 – Open the PDF.co Request Tester. This will require that you log in to your PDF.co account. Here’s the direct link https://app.pdf.co/request-tester.

Step 6 – Under PDF.co API endpoint field, select the /pdf/classifier.

Step 7 – Remove the rulescsv parameter’s default value and paste the rulesCSV value that we copied from the Classifier Test Tool.

Step 8 – Run the Request button to see the output.

Step 9 – Click on the output file link to preview.

Step 10 – Here’s the PDF.co PDF Classifier in action.

Get the PDF classification testing tool now from this page (you need to download the PDF Multitool app).