276 lines
13 KiB
Java
Generated
276 lines
13 KiB
Java
Generated
// This file is auto-generated by alef — DO NOT EDIT.
|
|
// alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
|
|
// To regenerate: alef generate
|
|
// To verify freshness: alef verify --exit-code
|
|
// Issues & docs: https://github.com/kreuzberg-dev/alef
|
|
package dev.kreuzberg;
|
|
|
|
import com.fasterxml.jackson.databind.deser.std.StdDeserializer;
|
|
import com.fasterxml.jackson.databind.ser.std.StdSerializer;
|
|
import com.fasterxml.jackson.core.JsonParser;
|
|
import com.fasterxml.jackson.core.JsonGenerator;
|
|
import com.fasterxml.jackson.databind.DeserializationContext;
|
|
import com.fasterxml.jackson.databind.SerializerProvider;
|
|
import com.fasterxml.jackson.databind.node.ObjectNode;
|
|
import com.fasterxml.jackson.databind.annotation.JsonDeserialize;
|
|
import com.fasterxml.jackson.databind.annotation.JsonSerialize;
|
|
import org.jspecify.annotations.Nullable;
|
|
|
|
/**
|
|
* Format-specific metadata (discriminated union).
|
|
*
|
|
* Only one format type can exist per extraction result. This provides
|
|
* type-safe, clean metadata without nested optionals.
|
|
*/
|
|
@com.fasterxml.jackson.annotation.JsonIgnoreProperties(ignoreUnknown = true)
|
|
@JsonDeserialize(using = FormatMetadataDeserializer.class)
|
|
@JsonSerialize(using = FormatMetadataSerializer.class)
|
|
public sealed interface FormatMetadata {
|
|
|
|
record Pdf(PdfMetadata value) implements FormatMetadata { }
|
|
|
|
record Docx(DocxMetadata value) implements FormatMetadata { }
|
|
|
|
record Excel(ExcelMetadata value) implements FormatMetadata { }
|
|
|
|
record Email(EmailMetadata value) implements FormatMetadata { }
|
|
|
|
record Pptx(PptxMetadata value) implements FormatMetadata { }
|
|
|
|
record Archive(ArchiveMetadata value) implements FormatMetadata { }
|
|
|
|
record Image(ImageMetadata value) implements FormatMetadata { }
|
|
|
|
record Xml(XmlMetadata value) implements FormatMetadata { }
|
|
|
|
record Text(TextMetadata value) implements FormatMetadata { }
|
|
|
|
record Html(HtmlMetadata value) implements FormatMetadata { }
|
|
|
|
record Ocr(OcrMetadata value) implements FormatMetadata { }
|
|
|
|
record Csv(CsvMetadata value) implements FormatMetadata { }
|
|
|
|
record Bibtex(BibtexMetadata value) implements FormatMetadata { }
|
|
|
|
record Citation(CitationMetadata value) implements FormatMetadata { }
|
|
|
|
record FictionBook(FictionBookMetadata value) implements FormatMetadata { }
|
|
|
|
record Dbf(DbfMetadata value) implements FormatMetadata { }
|
|
|
|
record Jats(JatsMetadata value) implements FormatMetadata { }
|
|
|
|
record Epub(EpubMetadata value) implements FormatMetadata { }
|
|
|
|
record Pst(PstMetadata value) implements FormatMetadata { }
|
|
|
|
record Code(String value) implements FormatMetadata { }
|
|
|
|
/** Returns the Pdf data if this is a Pdf variant, otherwise null. */
|
|
default @Nullable PdfMetadata pdf() {
|
|
return this instanceof Pdf e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Docx data if this is a Docx variant, otherwise null. */
|
|
default @Nullable DocxMetadata docx() {
|
|
return this instanceof Docx e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Excel data if this is a Excel variant, otherwise null. */
|
|
default @Nullable ExcelMetadata excel() {
|
|
return this instanceof Excel e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Email data if this is a Email variant, otherwise null. */
|
|
default @Nullable EmailMetadata email() {
|
|
return this instanceof Email e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Pptx data if this is a Pptx variant, otherwise null. */
|
|
default @Nullable PptxMetadata pptx() {
|
|
return this instanceof Pptx e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Archive data if this is a Archive variant, otherwise null. */
|
|
default @Nullable ArchiveMetadata archive() {
|
|
return this instanceof Archive e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Image data if this is a Image variant, otherwise null. */
|
|
default @Nullable ImageMetadata image() {
|
|
return this instanceof Image e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Xml data if this is a Xml variant, otherwise null. */
|
|
default @Nullable XmlMetadata xml() {
|
|
return this instanceof Xml e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Text data if this is a Text variant, otherwise null. */
|
|
default @Nullable TextMetadata text() {
|
|
return this instanceof Text e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Html data if this is a Html variant, otherwise null. */
|
|
default @Nullable HtmlMetadata html() {
|
|
return this instanceof Html e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Ocr data if this is a Ocr variant, otherwise null. */
|
|
default @Nullable OcrMetadata ocr() {
|
|
return this instanceof Ocr e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Csv data if this is a Csv variant, otherwise null. */
|
|
default @Nullable CsvMetadata csv() {
|
|
return this instanceof Csv e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Bibtex data if this is a Bibtex variant, otherwise null. */
|
|
default @Nullable BibtexMetadata bibtex() {
|
|
return this instanceof Bibtex e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Citation data if this is a Citation variant, otherwise null. */
|
|
default @Nullable CitationMetadata citation() {
|
|
return this instanceof Citation e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the FictionBook data if this is a FictionBook variant, otherwise null. */
|
|
default @Nullable FictionBookMetadata fictionBook() {
|
|
return this instanceof FictionBook e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Dbf data if this is a Dbf variant, otherwise null. */
|
|
default @Nullable DbfMetadata dbf() {
|
|
return this instanceof Dbf e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Jats data if this is a Jats variant, otherwise null. */
|
|
default @Nullable JatsMetadata jats() {
|
|
return this instanceof Jats e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Epub data if this is a Epub variant, otherwise null. */
|
|
default @Nullable EpubMetadata epub() {
|
|
return this instanceof Epub e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Pst data if this is a Pst variant, otherwise null. */
|
|
default @Nullable PstMetadata pst() {
|
|
return this instanceof Pst e ? e.value() : null;
|
|
}
|
|
|
|
/** Returns the Code data if this is a Code variant, otherwise null. */
|
|
default @Nullable String code() {
|
|
return this instanceof Code e ? e.value() : null;
|
|
}
|
|
|
|
}
|
|
|
|
// Custom deserializer for sealed interface with unwrapped variants
|
|
class FormatMetadataDeserializer extends StdDeserializer<FormatMetadata> {
|
|
FormatMetadataDeserializer() {
|
|
super(FormatMetadata.class);
|
|
}
|
|
|
|
@Override
|
|
public FormatMetadata deserialize(JsonParser parser, DeserializationContext ctx)
|
|
throws java.io.IOException {
|
|
ObjectNode node = parser.getCodec().readTree(parser);
|
|
com.fasterxml.jackson.databind.JsonNode tagNode = node.get("format_type");
|
|
if (tagNode == null || tagNode.isNull()) {
|
|
throw new com.fasterxml.jackson.databind.JsonMappingException(
|
|
parser, "Missing discriminator field: format_type");
|
|
}
|
|
String tagValue = tagNode.asText();
|
|
node.remove("format_type");
|
|
|
|
return switch (tagValue) {
|
|
case "pdf" -> new FormatMetadata.Pdf(ctx.readTreeAsValue(node, PdfMetadata.class));
|
|
case "docx" -> new FormatMetadata.Docx(ctx.readTreeAsValue(node, DocxMetadata.class));
|
|
case "excel" -> new FormatMetadata.Excel(ctx.readTreeAsValue(node, ExcelMetadata.class));
|
|
case "email" -> new FormatMetadata.Email(ctx.readTreeAsValue(node, EmailMetadata.class));
|
|
case "pptx" -> new FormatMetadata.Pptx(ctx.readTreeAsValue(node, PptxMetadata.class));
|
|
case "archive" -> new FormatMetadata.Archive(ctx.readTreeAsValue(node, ArchiveMetadata.class));
|
|
case "image" -> new FormatMetadata.Image(ctx.readTreeAsValue(node, ImageMetadata.class));
|
|
case "xml" -> new FormatMetadata.Xml(ctx.readTreeAsValue(node, XmlMetadata.class));
|
|
case "text" -> new FormatMetadata.Text(ctx.readTreeAsValue(node, TextMetadata.class));
|
|
case "html" -> new FormatMetadata.Html(ctx.readTreeAsValue(node, HtmlMetadata.class));
|
|
case "ocr" -> new FormatMetadata.Ocr(ctx.readTreeAsValue(node, OcrMetadata.class));
|
|
case "csv" -> new FormatMetadata.Csv(ctx.readTreeAsValue(node, CsvMetadata.class));
|
|
case "bibtex" -> new FormatMetadata.Bibtex(ctx.readTreeAsValue(node, BibtexMetadata.class));
|
|
case "citation" -> new FormatMetadata.Citation(ctx.readTreeAsValue(node, CitationMetadata.class));
|
|
case "fiction_book" -> new FormatMetadata.FictionBook(ctx.readTreeAsValue(node, FictionBookMetadata.class));
|
|
case "dbf" -> new FormatMetadata.Dbf(ctx.readTreeAsValue(node, DbfMetadata.class));
|
|
case "jats" -> new FormatMetadata.Jats(ctx.readTreeAsValue(node, JatsMetadata.class));
|
|
case "epub" -> new FormatMetadata.Epub(ctx.readTreeAsValue(node, EpubMetadata.class));
|
|
case "pst" -> new FormatMetadata.Pst(ctx.readTreeAsValue(node, PstMetadata.class));
|
|
case "code" -> new FormatMetadata.Code(node.toString());
|
|
default -> throw new com.fasterxml.jackson.databind.JsonMappingException(
|
|
parser, "Unknown FormatMetadata discriminator: " + tagValue);
|
|
};
|
|
}
|
|
}
|
|
|
|
// Custom serializer for sealed interface with unwrapped variants — emits
|
|
// the discriminator tag alongside the inner record's fields (flat object).
|
|
class FormatMetadataSerializer extends StdSerializer<FormatMetadata> {
|
|
private static final com.fasterxml.jackson.databind.ObjectMapper MAPPER =
|
|
new com.fasterxml.jackson.databind.ObjectMapper()
|
|
.registerModule(new com.fasterxml.jackson.datatype.jdk8.Jdk8Module())
|
|
.setPropertyNamingStrategy(com.fasterxml.jackson.databind.PropertyNamingStrategies.SNAKE_CASE)
|
|
.setSerializationInclusion(com.fasterxml.jackson.annotation.JsonInclude.Include.NON_NULL);
|
|
|
|
FormatMetadataSerializer() {
|
|
super(FormatMetadata.class);
|
|
}
|
|
|
|
@Override
|
|
public void serialize(FormatMetadata value, JsonGenerator gen, SerializerProvider provider)
|
|
throws java.io.IOException {
|
|
String tag;
|
|
Object inner;if (value instanceof FormatMetadata.Pdf v) {
|
|
tag = "pdf"; inner = v.value(); }else if (value instanceof FormatMetadata.Docx v) {
|
|
tag = "docx"; inner = v.value(); }else if (value instanceof FormatMetadata.Excel v) {
|
|
tag = "excel"; inner = v.value(); }else if (value instanceof FormatMetadata.Email v) {
|
|
tag = "email"; inner = v.value(); }else if (value instanceof FormatMetadata.Pptx v) {
|
|
tag = "pptx"; inner = v.value(); }else if (value instanceof FormatMetadata.Archive v) {
|
|
tag = "archive"; inner = v.value(); }else if (value instanceof FormatMetadata.Image v) {
|
|
tag = "image"; inner = v.value(); }else if (value instanceof FormatMetadata.Xml v) {
|
|
tag = "xml"; inner = v.value(); }else if (value instanceof FormatMetadata.Text v) {
|
|
tag = "text"; inner = v.value(); }else if (value instanceof FormatMetadata.Html v) {
|
|
tag = "html"; inner = v.value(); }else if (value instanceof FormatMetadata.Ocr v) {
|
|
tag = "ocr"; inner = v.value(); }else if (value instanceof FormatMetadata.Csv v) {
|
|
tag = "csv"; inner = v.value(); }else if (value instanceof FormatMetadata.Bibtex v) {
|
|
tag = "bibtex"; inner = v.value(); }else if (value instanceof FormatMetadata.Citation v) {
|
|
tag = "citation"; inner = v.value(); }else if (value instanceof FormatMetadata.FictionBook v) {
|
|
tag = "fiction_book"; inner = v.value(); }else if (value instanceof FormatMetadata.Dbf v) {
|
|
tag = "dbf"; inner = v.value(); }else if (value instanceof FormatMetadata.Jats v) {
|
|
tag = "jats"; inner = v.value(); }else if (value instanceof FormatMetadata.Epub v) {
|
|
tag = "epub"; inner = v.value(); }else if (value instanceof FormatMetadata.Pst v) {
|
|
tag = "pst"; inner = v.value(); }else if (value instanceof FormatMetadata.Code v) {
|
|
tag = "code"; inner = v.value(); } else {
|
|
throw new com.fasterxml.jackson.databind.JsonMappingException(gen,
|
|
"Unknown FormatMetadata variant: " + value.getClass().getName());
|
|
}
|
|
|
|
gen.writeStartObject();
|
|
gen.writeStringField("format_type", tag);
|
|
if (inner != null) {
|
|
com.fasterxml.jackson.databind.JsonNode tree = MAPPER.valueToTree(inner);
|
|
if (tree.isObject()) {
|
|
java.util.Iterator<java.util.Map.Entry<String, com.fasterxml.jackson.databind.JsonNode>> it =
|
|
tree.fields();
|
|
while (it.hasNext()) {
|
|
java.util.Map.Entry<String, com.fasterxml.jackson.databind.JsonNode> e = it.next();
|
|
gen.writeFieldName(e.getKey());
|
|
gen.writeTree(e.getValue());
|
|
}
|
|
}
|
|
}
|
|
gen.writeEndObject();
|
|
}
|
|
}
|