Optimize speech search and admin payloads

src/lib/admin-speeches.ts

@@ -0,0 +1,52 @@
+import type { ArchiveEntry } from "./archive";
+
+export type SpeechAdminEntry = {
+  id: string;
+  path: string;
+  title: string;
+  publishedDate: string;
+  speechYear: string;
+  speaker: "rev" | "mrs" | "other";
+  source: string;
+  categories: string[];
+  tags: string[];
+  bodyUrl: string;
+};
+
+function encodePath(value: string): string {
+  return value.split("/").map(encodeURIComponent).join("/");
+}
+
+export function speechAdminEntryFromArchiveEntry(entry: ArchiveEntry, fallbackDate: string): SpeechAdminEntry {
+  const categories = entry.data.categories ?? [];
+  const tags = entry.data.tags ?? [];
+  const publishedDate = entry.data.published?.slice(0, 10) ?? fallbackDate;
+  const speechYear = entry.data.speechYear ?? entry.id.match(/\b(1\d{3}|20\d{2})\b/)?.[1] ?? new Date().getFullYear().toString();
+  const categoryText = categories.join(" ").toLowerCase();
+  const collection = entry.data.speechCollection ?? "";
+  const speaker = categoryText.includes("hak ja han") || collection === "mrs-hak-ja-han-moon"
+    ? "mrs"
+    : categoryText.includes("rev sun myung moon") || /^\d{4}$/.test(collection) || /^\d{4}$/.test(speechYear)
+      ? "rev"
+      : "other";
+
+  return {
+    id: entry.id,
+    path: `content/${entry.id}.md`,
+    title: entry.data.title ?? entry.id,
+    publishedDate,
+    speechYear,
+    speaker,
+    source: entry.data.source ?? "",
+    categories,
+    tags,
+    bodyUrl: `/admin/content/${encodePath(entry.id)}.json`,
+  };
+}
+
+export function speechAdminEntryWithBody(entry: ArchiveEntry, fallbackDate: string) {
+  return {
+    ...speechAdminEntryFromArchiveEntry(entry, fallbackDate),
+    body: entry.body ?? "",
+  };
+}

src/lib/speeches.ts

@@ -32,6 +32,14 @@ export type SpeechArchiveItem = {
   searchText: string;
 };
 
+export type SpeechSearchIndexItem = Omit<SpeechArchiveItem, "searchText">;
+
+export type CompactSpeechSearchIndex = {
+  version: 2;
+  items: SpeechSearchIndexItem[];
+  index: Record<string, string>;
+};
+
 const entityMap: Record<string, string> = {
   amp: "&",
   apos: "'",
@@ -215,6 +223,39 @@ export function getSpeechArchive(): SpeechArchiveItem[] {
     });
 }
 
+function packDocumentIds(documentIds: number[]): string {
+  return documentIds.map((documentId) => documentId.toString(36)).join(",");
+}
+
+export function getCompactSpeechSearchIndex(): CompactSpeechSearchIndex {
+  const items = getSpeechArchive();
+  const index = new Map<string, number[]>();
+
+  items.forEach((item, documentId) => {
+    const terms = new Set(item.searchText.match(/[a-z0-9]+/g) ?? []);
+
+    terms.forEach((term) => {
+      if (term.length < 2) {
+        return;
+      }
+
+      const documentIds = index.get(term) ?? [];
+      documentIds.push(documentId);
+      index.set(term, documentIds);
+    });
+  });
+
+  return {
+    version: 2,
+    items: items.map(({ searchText, ...item }) => item),
+    index: Object.fromEntries(
+      Array.from(index)
+        .sort(([a], [b]) => a.localeCompare(b))
+        .map(([term, documentIds]) => [term, packDocumentIds(documentIds)]),
+    ),
+  };
+}
+
 export function uniqueSorted(values: string[], direction: "asc" | "desc" = "asc"): string[] {
   const unique = Array.from(new Set(values.filter(Boolean)));
   return unique.sort((a, b) => {