TIKA-4861: Detect raw camera formats by content

[dschmidt]

Raw camera formats were detected by extension only; without a name they came back as image/tiff and the previews of TIKA-4824 never appeared.

RawTiffDetector (tika-parser-image-module) tells the TIFF-based formats apart by their image directory:

1. DNGVersion tag: DNG
2. vendor Compression code (ExifTool's rule): Nikon, Sony, Pentax, Samsung
3. sensor data (CFA/LinearRaw, or a full-resolution single deep sample without PhotometricInterpretation) plus Make: the uncompressed variants

A TIFF a camera wrote has RGB data and stays a TIFF. The third rule's shape-based part exists only for the Samsung NX1 and is the weakest; dropping it costs just the NX1 without a file name.

It reads the stream on demand up to 1 MiB (directories sit behind the embedded previews, up to ~460 KB in practice), a plain TIFF costs a few KB, and the stream is reset.

Also: magic entries for RAF, RW2, MRW and the other ORF byte orders; image/x-raw-samsung added and handled by RawTiffParser.

Verified without a file name on 21 files from raw.pixls.us (NEF D1 to Z 6, NRW, ARW old and new, SR2, SRW NX100 and NX1, PEF, DNG from Ricoh, Leica and Nokia, RAF, RW2, MRW, ORF, and a CR2 that keeps its type), 20 to 600 microseconds each. Not added to the repo because of their size; the unit tests build their own TIFFs instead.

https://issues.apache.org/jira/browse/TIKA-4861

[Copilot]

Pull request overview

This PR improves Apache Tika’s raw camera format detection so TIFF-based raw files (e.g., NEF/ARW/PEF/SRW/DNG) can be correctly identified by content even when a resource name/extension is missing, and extends mime magic coverage for several other raw formats (RAF/RW2/MRW/ORF).

Changes:

• Add RawTiffDetector to distinguish TIFF-based raw camera formats from plain TIFF by walking TIFF/BigTIFF IFDs (including SubIFDs) within a bounded prefix.
• Add image/x-raw-samsung mime type and ensure RawTiffParser supports parsing Samsung SRW.
• Expand mime magic signatures (Fuji RAF, Panasonic RW2, Minolta MRW, additional Olympus ORF byte orders) and add/extend integration/unit tests covering the new behavior.

Reviewed changes

Copilot reviewed 7 out of 8 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-image-module/src/test/java/org/apache/tika/detect/image/RawTiffDetectorTest.java	Adds focused unit tests for raw-vs-TIFF discrimination logic, BigTIFF handling, cycle safety, and stream reset behavior.
tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-image-module/src/main/java/org/apache/tika/parser/image/RawTiffParser.java	Extends supported types to include image/x-raw-samsung.
tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-image-module/src/main/java/org/apache/tika/detect/image/RawTiffDetector.java	Introduces content-based detector for TIFF-based raw camera formats with bounded read and safe IFD/SubIFD traversal.
tika-parsers/tika-parsers-standard/tika-parsers-standard-integration-tests/src/test/java/org/apache/tika/mime/TestMimeTypes.java	Adds assertions for new/expanded raw magic signatures and documents the separation between MimeTypes-only vs container-aware detection for TIFF-based raws.
tika-parsers/tika-parsers-standard/tika-parsers-standard-integration-tests/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java	Adds integration test ensuring container-aware detection resolves TIFF-based raw types by data without relying on file name.
tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml	Adds image/x-raw-samsung and adds/extends magic matches for RAF/RW2/MRW/ORF variants.
CHANGES.txt	Documents the new content-based detection and the new Samsung raw mime type.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[dschmidt]

Thanks @THausherr for helping out :)

[dschmidt]

Wooop wooop 🎉