Read the inline text, and decode the escapes and references CommonMark spells (#33)

corpus/normalization/entity-references.json

@@ -0,0 +1,37 @@
+{
+  "content": [
+    {
+      "content": [
+        {
+          "text": "© 2026 — the № you asked for, & nothing else.",
+          "type": "text"
+        }
+      ],
+      "type": "paragraph"
+    },
+    {
+      "content": [
+        {
+          "text": "Escaped, the reference stays literal: © and ",
+          "type": "text"
+        },
+        {
+          "marks": [
+            {
+              "type": "code"
+            }
+          ],
+          "text": "©",
+          "type": "text"
+        },
+        {
+          "text": " alike.",
+          "type": "text"
+        }
+      ],
+      "type": "paragraph"
+    }
+  ],
+  "type": "doc",
+  "version": 1
+}

corpus/normalization/entity-references.md

@@ -0,0 +1,3 @@
+© 2026 — the № you asked for, & nothing else.
+
+Escaped, the reference stays literal: \© and `©` alike.

corpus/normalization/soft-wraps.json

@@ -0,0 +1,22 @@
+{
+  "content": [
+    {
+      "content": [
+        {
+          "text": "A paragraph the author soft-wrapped across three lines, holding a hard break",
+          "type": "text"
+        },
+        {
+          "type": "hardBreak"
+        },
+        {
+          "text": "the backslash spells.",
+          "type": "text"
+        }
+      ],
+      "type": "paragraph"
+    }
+  ],
+  "type": "doc",
+  "version": 1
+}

corpus/normalization/soft-wraps.md

@@ -0,0 +1,3 @@
+A paragraph the author soft-wrapped
+across three lines, holding a hard break\
+the backslash spells.