@peristyle/emdash-plugin-instagram-to-recipe 0.1.3 → 0.1.4

package/dist/index.js

@@ -2,7 +2,7 @@
 function instagramToRecipePlugin(options = {}) {
   return {
     id: "peristyle-instagram-to-recipe",
-    version: "0.1.3",
+    version: "0.1.4",
     format: "standard",
     entrypoint: "@peristyle/emdash-plugin-instagram-to-recipe/sandbox",
     options,

package/dist/sandbox-entry.js

@@ -349,26 +349,51 @@ function extractTitleFromCaption(caption) {
   }
   return titleCase(firstSentence.slice(0, 68).trim());
 }
+var _QTY_UNIT_SPLIT = new RegExp(
+  String.raw`(?<=\s)(?=(?:\d[\d/.\s]*|[¼½¾⅓⅔⅛]\s*)(?:lbs?|pounds?|cups?|tsp|tbsp|teaspoons?|tablespoons?|oz|ounces?|cloves?|ears?|heads?|cans?|bunch(?:es)?|sticks?|pinch|dash|g|kg|ml|l)\b)` + String.raw`|(?<=\s)(?=(?:zest|juice)\s+of\b)` + String.raw`|(?<=\s)(?<!\b(?:of|about|around|approx)\s)(?=[\d¼½¾⅓⅔⅛][\d/.¼½¾⅓⅔⅛]*\s+[a-zA-Z])`,
+  "giu"
+);
+function looksCollapsed(line) {
+  const starts = line.split(_QTY_UNIT_SPLIT);
+  return line.split(/\s+/).length > 8 && starts.length >= 3;
+}
+function resplitCollapsedIngredientLine(line) {
+  if (!looksCollapsed(line)) return [line];
+  return line.split(_QTY_UNIT_SPLIT).map((part) => part.trim()).filter(Boolean);
+}
 function parseIngredientsFromCaption(caption) {
   const match = caption.match(/what you['']ll need:?|ingredients\s*:/i);
   if (!match || match.index == null) return [];
   const afterHeader = caption.slice(match.index + match[0].length);
   const stepsMatch = afterHeader.match(/(?:^|\n)\s*\d+\.\s+/);
-  const ingredientsBlock = stepsMatch?.index ? afterHeader.slice(0, stepsMatch.index) : afterHeader;
+  let ingredientsBlock = stepsMatch?.index ? afterHeader.slice(0, stepsMatch.index) : afterHeader;
+  const instructionsHeader = ingredientsBlock.match(
+    /\b(?:instructions|directions|steps|method)\s*:/i
+  );
+  if (instructionsHeader?.index) {
+    ingredientsBlock = ingredientsBlock.slice(0, instructionsHeader.index);
+  }
   const lines = ingredientsBlock.split(/\n/);
   const ingredients = [];
   for (const raw of lines) {
-    const line = raw.trim();
-    if (!line) continue;
-    if (/^\[[^\]]+\]$/.test(line)) continue;
-    if (/^recipe below/i.test(line)) continue;
-    ingredients.push(line);
+    for (const piece of resplitCollapsedIngredientLine(raw)) {
+      const line = piece.trim();
+      if (!line) continue;
+      if (/^\[[^\]]+\]$/.test(line)) continue;
+      if (/^recipe below/i.test(line)) continue;
+      ingredients.push(line);
+    }
   }
   return ingredients;
 }
 function parseInstructionsFromCaption(caption) {
   const matches = [...caption.matchAll(/(?:^|\n)\s*(\d+)\.\s*([^\n]+)/g)];
-  return matches.map((m) => m[2]?.trim()).filter((s) => Boolean(s && s.length > 3));
+  const steps = matches.map((m) => m[2]?.trim()).filter((s) => Boolean(s && s.length > 3));
+  if (steps.length > 0) return steps;
+  const header = caption.match(/\b(?:instructions|directions|steps|method)\s*:/i);
+  if (!header || header.index == null) return [];
+  const block = caption.slice(header.index + header[0].length);
+  return block.split(/(?=\b\d{1,2}\.\s)/).map((s) => s.replace(/^\d{1,2}\.\s*/, "").trim()).filter((s) => s.length > 3);
 }
 function parseRecipeCaption(caption) {
   if (!isLikelyRecipeCaption(caption)) return null;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@peristyle/emdash-plugin-instagram-to-recipe",
-  "version": "0.1.3",
+  "version": "0.1.4",
   "type": "module",
   "description": "Import Instagram post URLs as recipe drafts in the EmDash admin",
   "publishConfig": {

package/src/instagram-recipes.ts

@@ -117,25 +117,69 @@ export function extractTitleFromCaption(caption: string): string {
   return titleCase(firstSentence.slice(0, 68).trim());
 }
 
+// A quantity token starting a new ingredient: "1 lb", "1/3 cup", "½ tsp", "3
+// cloves"... Used to re-split caption lines whose newlines were lost upstream
+// (Instagram's og:description meta tag flattens single line breaks, so a whole
+// ingredient section can arrive as one long line).
+const _QTY_UNIT_SPLIT = new RegExp(
+  String.raw`(?<=\s)(?=(?:\d[\d/.\s]*|[¼½¾⅓⅔⅛]\s*)(?:lbs?|pounds?|cups?|tsp|tbsp|teaspoons?|tablespoons?|oz|ounces?|cloves?|ears?|heads?|cans?|bunch(?:es)?|sticks?|pinch|dash|g|kg|ml|l)\b)` +
+    String.raw`|(?<=\s)(?=(?:zest|juice)\s+of\b)` +
+    String.raw`|(?<=\s)(?<!\b(?:of|about|around|approx)\s)(?=[\d¼½¾⅓⅔⅛][\d/.¼½¾⅓⅔⅛]*\s+[a-zA-Z])`,
+  "giu",
+);
+
+/** Count of quantity-like tokens — how many ingredients a line appears to hold. */
+function looksCollapsed(line: string): boolean {
+  const starts = line.split(_QTY_UNIT_SPLIT);
+  return line.split(/\s+/).length > 8 && starts.length >= 3;
+}
+
+/**
+ * Split a line that holds several ingredients back into one line each.
+ *
+ * Lossy-input recovery only: applied when a line clearly contains multiple
+ * quantity tokens (see ``looksCollapsed``), never to normal one-ingredient
+ * lines. A trailing "Section name:" prefix is kept as its own line so the
+ * downstream header filters can drop it.
+ */
+export function resplitCollapsedIngredientLine(line: string): string[] {
+  if (!looksCollapsed(line)) return [line];
+  return line
+    .split(_QTY_UNIT_SPLIT)
+    .map((part) => part.trim())
+    .filter(Boolean);
+}
+
 export function parseIngredientsFromCaption(caption: string): string[] {
   const match = caption.match(/what you['']ll need:?|ingredients\s*:/i);
   if (!match || match.index == null) return [];
 
   const afterHeader = caption.slice(match.index + match[0].length);
   const stepsMatch = afterHeader.match(/(?:^|\n)\s*\d+\.\s+/);
-  const ingredientsBlock = stepsMatch?.index
+  let ingredientsBlock = stepsMatch?.index
     ? afterHeader.slice(0, stepsMatch.index)
     : afterHeader;
+  // With a collapsed caption the numbered steps carry no leading newline, so
+  // the cut above misses. Cut again at an explicit instructions header so the
+  // steps can't leak into the ingredient list.
+  const instructionsHeader = ingredientsBlock.match(
+    /\b(?:instructions|directions|steps|method)\s*:/i,
+  );
+  if (instructionsHeader?.index) {
+    ingredientsBlock = ingredientsBlock.slice(0, instructionsHeader.index);
+  }
 
   const lines = ingredientsBlock.split(/\n/);
   const ingredients: string[] = [];
 
   for (const raw of lines) {
-    const line = raw.trim();
-    if (!line) continue;
-    if (/^\[[^\]]+\]$/.test(line)) continue;
-    if (/^recipe below/i.test(line)) continue;
-    ingredients.push(line);
+    for (const piece of resplitCollapsedIngredientLine(raw)) {
+      const line = piece.trim();
+      if (!line) continue;
+      if (/^\[[^\]]+\]$/.test(line)) continue;
+      if (/^recipe below/i.test(line)) continue;
+      ingredients.push(line);
+    }
   }
 
   return ingredients;
@@ -143,9 +187,19 @@ export function parseIngredientsFromCaption(caption: string): string[] {
 
 export function parseInstructionsFromCaption(caption: string): string[] {
   const matches = [...caption.matchAll(/(?:^|\n)\s*(\d+)\.\s*([^\n]+)/g)];
-  return matches
+  const steps = matches
     .map((m) => m[2]?.trim())
     .filter((s): s is string => Boolean(s && s.length > 3));
+  if (steps.length > 0) return steps;
+  // Collapsed caption (og:description strips newlines): fall back to slicing
+  // the text after an instructions header on the "1." "2." markers themselves.
+  const header = caption.match(/\b(?:instructions|directions|steps|method)\s*:/i);
+  if (!header || header.index == null) return [];
+  const block = caption.slice(header.index + header[0].length);
+  return block
+    .split(/(?=\b\d{1,2}\.\s)/)
+    .map((s) => s.replace(/^\d{1,2}\.\s*/, "").trim())
+    .filter((s) => s.length > 3);
 }
 
 export function parseRecipeCaption(caption: string): {