PDF Parsing

{
  "parsers": [
    {
      "pdf-parser": {
        "markedContent": { "strategy": "NONE" },
        "extractBookmarksText": false,
        "extractAnnotationText": false
      }
    },
    { "default-parser": {} }
  ]
}

java PdfToXhtml catalogue.pdf tika-config.json

<body><div class="page"><p />
<p>Autumn Catalogue 2026
</p>
<p>New fiction
</p>
<p>The Silent Shelf by A. Moreno: a librarian finds a book that rewrites itself.
</p>
<p>Paper Lanterns by K. Ito: twelve stories set in one Kyoto bookshop.
</p>
<p>Bookshelf - page 1</p>
<p />
</div>
<div class="page"><p />
<p>Order form
</p>
<p>Non-fiction
</p>
<p>Ink and Iron: a history of the printing press, 480 pages.
</p>
<p>Fill in the member number below and return this page to the shop.
</p>
<p>Bookshelf - page 2</p>
<p />
</div>
<div source="attachment" class="embedded" id="errata.pdf" />
<div class="page"><p />

<p>Errata: The Silent Shelf, 2nd printing, page 112.</p>

<p />

</div>

<div class="acroform"><ol>	<li altName="Member number">memberNumber: BK-20417</li>
</ol>
</div>
</body>

import java.nio.file.Path;
import org.apache.tika.io.TikaInputStream;
import org.apache.tika.metadata.Metadata;
import org.apache.tika.parser.AutoDetectParser;
import org.apache.tika.parser.ParseContext;
import org.apache.tika.parser.pages.PagesConfig;
import org.apache.tika.parser.pages.TextPolicy;
import org.apache.tika.parser.pdf.PDFParserConfig;
import org.apache.tika.sax.BodyContentHandler;

public class PdfIntake {

    public static void main(String[] args) throws Exception {
        AutoDetectParser parser = new AutoDetectParser();

        // Per-request PDF options: a typed object in the ParseContext.
        PDFParserConfig pdf = new PDFParserConfig();
        pdf.setExtractBookmarksText(false);   // the outline repeats the headings
        pdf.setExtractAnnotationText(true);   // keep reviewer comments
        pdf.setMaxPages(50);                  // a catalogue never needs more

        // Per-request pages policy: never OCR in this request.
        PagesConfig pages = new PagesConfig();
        pages.setText(TextPolicy.EXTRACT);

        ParseContext context = new ParseContext();
        context.set(PDFParserConfig.class, pdf);
        context.set(PagesConfig.class, pages);

        for (String file : new String[] {"files/catalogue.pdf", "files/scanned-loan-form.pdf"}) {
            Metadata metadata = new Metadata();
            BodyContentHandler handler = new BodyContentHandler(-1);
            try (TikaInputStream tis = TikaInputStream.get(Path.of(file), metadata)) {
                parser.parse(tis, handler, metadata, context);
            }
            System.out.println("== " + file);
            System.out.println(handler.toString().strip().replaceAll("\\s*\\n\\s*", "\n"));
            System.out.println("pdf:chars-per-page = " + String.join(", ", metadata.getValues("pdf:chars-per-page")));
            System.out.println("pdf:ocr-page-count = " + metadata.get("pdf:ocr-page-count"));
        }
    }
}

== files/catalogue.pdf
Autumn Catalogue 2026
New fiction
The Silent Shelf by A. Moreno: a librarian finds a book that rewrites itself.
Paper Lanterns by K. Ito: twelve stories set in one Kyoto bookshop.
Bookshelf - page 1
Check the price of Paper Lanterns before printing.
Order form
Non-fiction
Ink and Iron: a history of the printing press, 480 pages.
Fill in the member number below and return this page to the shop.
Bookshelf - page 2
Errata: The Silent Shelf, 2nd printing, page 112.
memberNumber: BK-20417
pdf:chars-per-page = 194, 161
pdf:ocr-page-count = 0
== files/scanned-loan-form.pdf
Cover note: request scanned at the front desk on 2026-10-02.
Staff: process within five working days.
pdf:chars-per-page = 0, 100
pdf:ocr-page-count = 0

<!-- Tika 3.x only: tika-config.xml -->
<properties>
  <parsers>
    <parser class="org.apache.tika.parser.DefaultParser">
      <parser-exclude class="org.apache.tika.parser.pdf.PDFParser"/>
    </parser>
    <parser class="org.apache.tika.parser.pdf.PDFParser">
      <params>
        <param name="extractBookmarksText" type="bool">false</param>
        <param name="extractMarkedContent" type="bool">false</param>
        <param name="ocrStrategy" type="string">auto</param>
        <param name="ocrDPI" type="int">300</param>
      </params>
    </parser>
  </parsers>
</properties>

{
  "parsers": [
    { "default-parser": {} },
    { "pdf-parser": { "pages": { "render": { "dpi": 200 },
                                 "emit": { "render": { "imageType": "RGB" } } } } }
  ]
}

{
  "engines": {
    "tesseract": { "tesseract-ocr-parser": { "language": "eng" } }
  },
  "text-recognizers": [
    { "engine": "tesseract" }
  ],
  "parsers": [
    { "default-parser": {} },
    { "pdf-parser": { "pages": { "ocr": { "maxPages": 20 } } } }
  ]
}

org.apache.tika.exception.TikaConfigException: Text recognizer engine "tesseract" advertises no media types (a _mime-include list does not count). Is the engine unavailable (missing native binary, unreachable inference server) or configured to skip enrichment?

{
  "parsers": [
    { "default-parser": {} },
    {
      "pdf-parser": {
        "pages": {
          "text": "AUTO",
          "emit": {
            "enabled": true,
            "maxPages": 1,
            "render": { "maxWidth": 256, "maxHeight": 256, "imageType": "RGB" }
          }
        }
      }
    }
  ]
}

import java.nio.file.Path;
import java.util.List;
import org.apache.tika.config.loader.TikaLoader;
import org.apache.tika.io.TikaInputStream;
import org.apache.tika.metadata.Metadata;
import org.apache.tika.parser.ParseContext;
import org.apache.tika.parser.RecursiveParserWrapper;
import org.apache.tika.sax.BasicContentHandlerFactory;
import org.apache.tika.sax.RecursiveParserWrapperHandler;

public class LoanFormPages {

    public static void main(String[] args) throws Exception {
        TikaLoader loader = TikaLoader.load(Path.of(args[0]));
        RecursiveParserWrapper wrapper = new RecursiveParserWrapper(loader.loadAutoDetectParser());
        ParseContext context = loader.loadParseContext();
        RecursiveParserWrapperHandler handler = new RecursiveParserWrapperHandler(
                new BasicContentHandlerFactory(BasicContentHandlerFactory.HANDLER_TYPE.TEXT, -1));

        Metadata metadata = new Metadata();
        try (TikaInputStream tis = TikaInputStream.get(Path.of("files/scanned-loan-form.pdf"), metadata)) {
            wrapper.parse(tis, handler, metadata, context);
        }

        List<Metadata> list = handler.getMetadataList();
        Metadata pdf = list.get(0);
        System.out.println("chars per page : " + String.join(", ", pdf.getValues("pdf:chars-per-page")));
        System.out.println("pages for OCR  : " + pdf.get("pdf:ocr-page-count"));
        System.out.println("text           : " + pdf.get("tk:content").strip().replaceAll("\\s*\\n\\s*", " | "));
        for (Metadata child : list.subList(1, list.size())) {
            System.out.println("embedded       : " + child.get("tk:embedded-resource-type")
                    + " " + child.get("Content-Type")
                    + " page " + child.get("tk:page:number")
                    + " " + child.get("tiff:ImageWidth") + "x" + child.get("tiff:ImageLength"));
        }
    }
}

java LoanFormPages pages-config.json

chars per page : 0, 100
pages for OCR  : 1
text           : Inter-library loan form | Cover note: request scanned at the front desk on 2026-10-02. | Staff: process within five working days.
embedded       : RENDERING image/png page 1 181x256

org.apache.tika.exception.TikaException: Unable to extract PDF content
Caused by: java.io.IOException: org.apache.tika.exception.TikaException: I regret that I couldn't find an OCR engine to handle image/png. Name one that covers it in "text-recognizers" (a configured list is authoritative), add one to the classpath when no list is configured, or set "pages": {"text": "EXTRACT"}.

import java.nio.file.Path;
import org.apache.tika.exception.EncryptedDocumentException;
import org.apache.tika.io.TikaInputStream;
import org.apache.tika.metadata.Metadata;
import org.apache.tika.parser.AutoDetectParser;
import org.apache.tika.parser.ParseContext;
import org.apache.tika.parser.PasswordProvider;
import org.apache.tika.sax.BodyContentHandler;

public class MembersCatalogue {

    static void parse(AutoDetectParser parser, ParseContext context, String label) throws Exception {
        Metadata metadata = new Metadata();
        BodyContentHandler handler = new BodyContentHandler(-1);
        try (TikaInputStream tis = TikaInputStream.get(Path.of("files/members-catalogue.pdf"), metadata)) {
            parser.parse(tis, handler, metadata, context);
            String text = handler.toString().strip();
            System.out.println(label + ": " + text.substring(0, text.indexOf('\n')));
        } catch (EncryptedDocumentException e) {
            System.out.println(label + ": " + e.getMessage() + " (cause: " + e.getCause().getMessage() + ")");
        }
        System.out.println("   pdf:encrypted = " + metadata.get("pdf:encrypted")
                + ", access-permission:extract-content = " + metadata.get("access-permission:extract-content"));
    }

    public static void main(String[] args) throws Exception {
        AutoDetectParser parser = new AutoDetectParser();

        // 1. No PasswordProvider in the context.
        parse(parser, new ParseContext(), "no password");

        // 2. A PasswordProvider that reads the secret from the environment.
        ParseContext context = new ParseContext();
        context.set(PasswordProvider.class, metadata -> System.getenv("BOOKSHELF_MEMBERS_PDF_PASSWORD"));
        parse(parser, context, "with password");
    }
}

# The value comes from your secret store; it is never written in source code.
export BOOKSHELF_MEMBERS_PDF_PASSWORD="$(your-secret-tool read bookshelf/members-pdf)"
java MembersCatalogue

no password: Unable to process: document is encrypted (cause: Cannot decrypt PDF, the password is incorrect)
   pdf:encrypted = true, access-permission:extract-content = null
with password: Autumn Catalogue 2026
   pdf:encrypted = true, access-permission:extract-content = false

{
  "parsers": [
    { "pdf-parser": { "accessCheckMode": "IGNORE_ACCESSIBILITY_ALLOWANCE" } },
    { "default-parser": {} }
  ]
}