18 lines
1.1 KiB
Markdown
18 lines
1.1 KiB
Markdown
|
<!-- snippet:skip reason="The Java binding generates IDocumentExtractor + DocumentExtractorBridge but the InternalDocument Java class referenced by the interface is not generated by the alef Java backend. Custom DocumentExtractor implementations cannot construct return values until the alef-generated Panama type for InternalDocument lands." -->
|
||
|
|
```java title="Java"
|
||
|
|
import dev.kreuzberg.IDocumentExtractor;
|
||
|
|
import dev.kreuzberg.DocumentExtractorBridge;
|
||
|
|
|
||
|
|
// Java's Panama FFM binding exposes the IDocumentExtractor interface and the
|
||
|
|
// DocumentExtractorBridge.registerDocumentExtractor / unregisterDocumentExtractor
|
||
|
|
// helpers, but the InternalDocument return type is referenced from the
|
||
|
|
// interface signature without a corresponding generated Java class. Until the
|
||
|
|
// alef Java backend emits dev.kreuzberg.InternalDocument, custom Java
|
||
|
|
// DocumentExtractor implementations cannot return a value from extract_bytes /
|
||
|
|
// extract_file.
|
||
|
|
//
|
||
|
|
// Implement the extractor in Rust as `Plugin + DocumentExtractor` and register
|
||
|
|
// it via `register_document_extractor` in a Rust shim crate that links
|
||
|
|
// kreuzberg before the JVM loads the native library.
|
||
|
|
```
|