Nomad changes

packages/java/dev/kreuzberg/HtmlMetadata.java

@@ -0,0 +1,247 @@
+// 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 java.util.List;
+import java.util.Map;
+import com.fasterxml.jackson.annotation.JsonProperty;
+import com.fasterxml.jackson.annotation.JsonInclude;
+import com.fasterxml.jackson.databind.annotation.JsonDeserialize;
+import com.fasterxml.jackson.databind.annotation.JsonPOJOBuilder;
+import org.jspecify.annotations.Nullable;
+
+/**
+ * HTML metadata extracted from HTML documents.
+ *
+ * Includes document-level metadata, Open Graph data, Twitter Card metadata,
+ * and extracted structural elements (headers, links, images, structured data).
+ */
+@JsonInclude(JsonInclude.Include.NON_ABSENT)
+@JsonDeserialize(builder = HtmlMetadata.Builder.class)
+public record HtmlMetadata(
+    /**
+     * Document title from {@code <title>} tag
+     */
+    @Nullable @JsonProperty("title") String title,
+    /**
+     * Document description from {@code <meta name="description">} tag
+     */
+    @Nullable @JsonProperty("description") String description,
+    /**
+     * Document keywords from {@code <meta name="keywords">} tag, split on commas
+     */
+    @Nullable @JsonProperty("keywords") List<String> keywords,
+    /**
+     * Document author from {@code &lt;meta name="author"&gt;} tag
+     */
+    @Nullable @JsonProperty("author") String author,
+    /**
+     * Canonical URL from {@code &lt;link rel="canonical"&gt;} tag
+     */
+    @Nullable @JsonProperty("canonical_url") String canonicalUrl,
+    /**
+     * Base URL from {@code &lt;base href=""&gt;} tag for resolving relative URLs
+     */
+    @Nullable @JsonProperty("base_href") String baseHref,
+    /**
+     * Document language from {@code lang} attribute
+     */
+    @Nullable @JsonProperty("language") String language,
+    /**
+     * Document text direction from {@code dir} attribute
+     */
+    @Nullable @JsonProperty("text_direction") TextDirection textDirection,
+    /**
+     * Open Graph metadata (og:* properties) for social media
+     * Keys like "title", "description", "image", "url", etc.
+     */
+    @Nullable @JsonProperty("open_graph") Map<String, String> openGraph,
+    /**
+     * Twitter Card metadata (twitter:* properties)
+     * Keys like "card", "site", "creator", "title", "description", "image", etc.
+     */
+    @Nullable @JsonProperty("twitter_card") Map<String, String> twitterCard,
+    /**
+     * Additional meta tags not covered by specific fields
+     * Keys are meta name/property attributes, values are content
+     */
+    @Nullable @JsonProperty("meta_tags") Map<String, String> metaTags,
+    /**
+     * Extracted header elements with hierarchy
+     */
+    @Nullable @JsonProperty("headers") List<HeaderMetadata> headers,
+    /**
+     * Extracted hyperlinks with type classification
+     */
+    @Nullable @JsonProperty("links") List<LinkMetadata> links,
+    /**
+     * Extracted images with source and dimensions
+     */
+    @Nullable @JsonProperty("images") List<ImageMetadataType> images,
+    /**
+     * Extracted structured data blocks
+     */
+    @Nullable @JsonProperty("structured_data") List<StructuredData> structuredData
+) {
+    public static Builder builder() {
+        return new Builder();
+    }
+
+    // CPD-OFF
+    @JsonPOJOBuilder(withPrefix = "with", buildMethodName = "build")
+    public static final class Builder {
+
+private String title = null;
+private String description = null;
+private List<String> keywords = null;
+private String author = null;
+        @JsonProperty("canonical_url")
+private String canonicalUrl = null;
+        @JsonProperty("base_href")
+private String baseHref = null;
+private String language = null;
+        @JsonProperty("text_direction")
+private TextDirection textDirection = null;
+        @JsonProperty("open_graph")
+private Map<String, String> openGraph = null;
+        @JsonProperty("twitter_card")
+private Map<String, String> twitterCard = null;
+        @JsonProperty("meta_tags")
+private Map<String, String> metaTags = null;
+private List<HeaderMetadata> headers = null;
+private List<LinkMetadata> links = null;
+private List<ImageMetadataType> images = null;
+        @JsonProperty("structured_data")
+private List<StructuredData> structuredData = null;
+
+        /** Sets the title field. */
+        @JsonProperty("title")
+        public Builder withTitle(final @Nullable String value) {
+            this.title = value;
+            return this;
+        }
+
+        /** Sets the description field. */
+        @JsonProperty("description")
+        public Builder withDescription(final @Nullable String value) {
+            this.description = value;
+            return this;
+        }
+
+        /** Sets the keywords field. */
+        @JsonProperty("keywords")
+        public Builder withKeywords(final @Nullable List<String> value) {
+            this.keywords = value;
+            return this;
+        }
+
+        /** Sets the author field. */
+        @JsonProperty("author")
+        public Builder withAuthor(final @Nullable String value) {
+            this.author = value;
+            return this;
+        }
+
+        /** Sets the canonicalUrl field. */
+        @JsonProperty("canonical_url")
+        public Builder withCanonicalUrl(final @Nullable String value) {
+            this.canonicalUrl = value;
+            return this;
+        }
+
+        /** Sets the baseHref field. */
+        @JsonProperty("base_href")
+        public Builder withBaseHref(final @Nullable String value) {
+            this.baseHref = value;
+            return this;
+        }
+
+        /** Sets the language field. */
+        @JsonProperty("language")
+        public Builder withLanguage(final @Nullable String value) {
+            this.language = value;
+            return this;
+        }
+
+        /** Sets the textDirection field. */
+        @JsonProperty("text_direction")
+        public Builder withTextDirection(final @Nullable TextDirection value) {
+            this.textDirection = value;
+            return this;
+        }
+
+        /** Sets the openGraph field. */
+        @JsonProperty("open_graph")
+        public Builder withOpenGraph(final @Nullable Map<String, String> value) {
+            this.openGraph = value;
+            return this;
+        }
+
+        /** Sets the twitterCard field. */
+        @JsonProperty("twitter_card")
+        public Builder withTwitterCard(final @Nullable Map<String, String> value) {
+            this.twitterCard = value;
+            return this;
+        }
+
+        /** Sets the metaTags field. */
+        @JsonProperty("meta_tags")
+        public Builder withMetaTags(final @Nullable Map<String, String> value) {
+            this.metaTags = value;
+            return this;
+        }
+
+        /** Sets the headers field. */
+        @JsonProperty("headers")
+        public Builder withHeaders(final @Nullable List<HeaderMetadata> value) {
+            this.headers = value;
+            return this;
+        }
+
+        /** Sets the links field. */
+        @JsonProperty("links")
+        public Builder withLinks(final @Nullable List<LinkMetadata> value) {
+            this.links = value;
+            return this;
+        }
+
+        /** Sets the images field. */
+        @JsonProperty("images")
+        public Builder withImages(final @Nullable List<ImageMetadataType> value) {
+            this.images = value;
+            return this;
+        }
+
+        /** Sets the structuredData field. */
+        @JsonProperty("structured_data")
+        public Builder withStructuredData(final @Nullable List<StructuredData> value) {
+            this.structuredData = value;
+            return this;
+        }
+
+        /** Builds the HtmlMetadata instance. */
+        public HtmlMetadata build() {
+            return new HtmlMetadata(
+                title,
+                description,
+                keywords,
+                author,
+                canonicalUrl,
+                baseHref,
+                language,
+                textDirection,
+                openGraph,
+                twitterCard,
+                metaTags,
+                headers,
+                links,
+                images,
+                structuredData
+            );
+        }
+    }
+    // CPD-ON
+}