From 6f9a42c7073579817733bf130b67c14caa3c0107 Mon Sep 17 00:00:00 2001
From: lila <beiweitemderbeste@protonmail.com>
Date: Tue, 5 May 2026 18:57:55 +0200
Subject: [PATCH] feat: add stage 2 reverse link sync script

---
 data-pipeline/package.json                    |   1 +
 .../scripts/reverse-link.ts                   | 109 ++++++++++++++++++
 2 files changed, 110 insertions(+)
 create mode 100644 data-pipeline/stage-2-reverse-link/scripts/reverse-link.ts

diff --git a/data-pipeline/package.json b/data-pipeline/package.json
index b25f26d..e47936a 100644
--- a/data-pipeline/package.json
+++ b/data-pipeline/package.json
@@ -5,6 +5,7 @@
   "type": "module",
   "scripts": {
     "extract": "tsx stage-1-extract/scripts/extract.ts",
+    "reverse-link": "tsx stage-2-reverse-link/scripts/reverse-link.ts",
     "db:import": "tsx db/import.ts",
     "db:init": "tsx db/init.ts",
     "annotate": "tsx stage-2-annotate/scripts/annotate.ts",
diff --git a/data-pipeline/stage-2-reverse-link/scripts/reverse-link.ts b/data-pipeline/stage-2-reverse-link/scripts/reverse-link.ts
new file mode 100644
index 0000000..da8c9b6
--- /dev/null
+++ b/data-pipeline/stage-2-reverse-link/scripts/reverse-link.ts
@@ -0,0 +1,109 @@
+import { openDb } from "../../db/index.js";
+
+// ── Types ─────────────────────────────────────────────────────────────────────
+
+type TranslationRow = {
+  translation_id: number;
+  entry_id: number;
+  entry_language: string;
+  entry_headword: string;
+  target_lang: string;
+  word: string;
+  sense_hint: string | null;
+};
+
+type EntryRow = { id: number };
+
+// ── Sync ──────────────────────────────────────────────────────────────────────
+
+export function reverseLink(): void {
+  const db = openDb();
+
+  // Find all translations and their source entry details
+  const translations = db
+    .prepare(
+      `SELECT
+        t.id          AS translation_id,
+        t.entry_id,
+        e.language    AS entry_language,
+        e.headword    AS entry_headword,
+        t.target_lang,
+        t.word,
+        t.sense_hint
+       FROM translations t
+       JOIN entries e ON e.id = t.entry_id`,
+    )
+    .all() as TranslationRow[];
+
+  console.log(
+    `  Found ${translations.length.toLocaleString()} translations to check`,
+  );
+
+  const findEntry = db.prepare(
+    `SELECT id FROM entries WHERE headword = ? AND language = ? LIMIT 1`,
+  );
+
+  const insertReverseLink = db.prepare(
+    `INSERT INTO translations (entry_id, target_lang, word, sense_hint, source)
+     VALUES (?, ?, ?, ?, 'reverse_link')
+     ON CONFLICT (entry_id, target_lang, word) DO NOTHING`,
+  );
+
+  const sync = db.transaction(() => {
+    let inserted = 0;
+    let skipped = 0;
+    let noEntry = 0;
+
+    for (const t of translations) {
+      // Look for an entry in the target language with the translation word as headword
+      const targetEntry = findEntry.get(t.word, t.target_lang) as
+        | EntryRow
+        | undefined;
+
+      if (!targetEntry) {
+        noEntry++;
+        continue;
+      }
+
+      // Insert reverse link: target entry → source language → source headword
+      const result = insertReverseLink.run(
+        targetEntry.id,
+        t.entry_language,
+        t.entry_headword,
+        t.sense_hint ?? null,
+      );
+
+      if (result.changes > 0) {
+        inserted++;
+      } else {
+        skipped++;
+      }
+    }
+
+    return { inserted, skipped, noEntry };
+  });
+
+  const counts = sync();
+
+  db.close();
+
+  console.log(`  Inserted: ${counts.inserted.toLocaleString()} reverse links`);
+  console.log(
+    `  Skipped:  ${counts.skipped.toLocaleString()} (already existed)`,
+  );
+  console.log(
+    `  No entry: ${counts.noEntry.toLocaleString()} (target word not in entries)`,
+  );
+}
+
+// ── Main ─────────────────────────────────────────────────────────────────────
+
+function main(): void {
+  console.log("Running reverse link sync...");
+  reverseLink();
+  console.log("\nReverse link sync complete.");
+}
+
+if (import.meta.url === `file://${process.argv[1]}`) {
+  main();
+}