From 1f783e5a415c211b05a84905e169021ad82b9ae0 Mon Sep 17 00:00:00 2001
From: Nicholas Ward <nicholaspward@outlook.com>
Date: Mon, 17 Aug 2026 18:25:28 -0500
Subject: [PATCH] feat(ingest): handle flavor variation inclusions and whole
 fruit zest count units in Streusel

---
 scripts/ingest-docling-book.mjs | 30 +++++++++++++++++++++++++++---
 1 file changed, 27 insertions(+), 3 deletions(-)

diff --git a/scripts/ingest-docling-book.mjs b/scripts/ingest-docling-book.mjs
index f0ce23e..bde9672 100644
--- a/scripts/ingest-docling-book.mjs
+++ b/scripts/ingest-docling-book.mjs
@@ -149,11 +149,14 @@ const KNOWN_INGREDIENT_MAP = {
   "olive oil": "olive_oil",
   "lemon juice": "lemon_juice",
   "lemon zest": "lemon_zest",
+  "lemon or lime zest": "lemon_zest",
+  "lime zest": "lime_zest",
   "lemons": "lemon",
   "orange juice": "orange_juice",
   "orange zest": "orange_zest",
   "lime juice": "lime_juice",
-  "lime zest": "lime_zest",
+  "loose tea": "tea_loose",
+  "chopped nuts": "walnut",
   "passion fruit puree": "passion_fruit_puree",
   "raspberry puree": "raspberry_puree",
   "strawberry puree": "strawberry_puree",
@@ -266,8 +269,29 @@ function cleanIngredientName(raw) {
 
   const parenMatch = cleaned.match(/^([^(]+)\s*\(([^)]+)\)$/);
   if (parenMatch) {
-    cleaned = parenMatch[1].trim();
-    notes = parenMatch[2].trim();
+    const baseName = parenMatch[1].trim();
+    const parenContent = parenMatch[2].trim();
+
+    if (/streusel/i.test(baseName)) {
+      if (/zest/i.test(parenContent)) {
+        const fruit = baseName.replace(/streusel/i, "").trim();
+        cleaned = `${fruit} Zest`;
+        notes = `Zest of whole fruit (for ${baseName})`;
+      } else if (/loose tea/i.test(parenContent)) {
+        cleaned = "Loose Tea";
+        notes = `For ${baseName}`;
+      } else if (/chopped/i.test(parenContent)) {
+        const nutType = baseName.replace(/streusel/i, "").trim();
+        cleaned = /nut/i.test(nutType) ? "Chopped Nuts" : (nutType || "Nuts");
+        notes = `${parenContent} (for ${baseName})`;
+      } else {
+        cleaned = baseName;
+        notes = parenContent;
+      }
+    } else {
+      cleaned = baseName;
+      notes = parenContent;
+    }
   }
 
   return { name: cleaned, notes };