Fix underscore-prefixed column names being treated as LaTeX subscript prefix

Author: Edwinexd

src/ra-engine/raHighlight.test.ts

@@ -171,4 +171,34 @@ describe("RA preview", () => {
     const opMatches = html.match(/ra-prev-op/g);
     expect(opMatches!.length).toBeGreaterThanOrEqual(2);
   });
+
+  it("should not treat underscore-prefixed column names as LaTeX subscript", () => {
+    const html = renderRAPreview("PI _name, person_id Person");
+    // _name should not be split — the underscore is part of the column name
+    expect(html).toMatch(/<sub[^>]*>.*_name.*person_id.*<\/sub>/);
+    expect(html).not.toMatch(/opacity/); // no faint bracket styling for _
+  });
+});
+
+describe("RA highlighter character preservation with underscore columns", () => {
+  it("should preserve character count with underscore-prefixed columns", () => {
+    const input = "PI _name, person_id (Person)";
+    const html = highlightRA(input);
+    const textOnly = html
+      .replace(/<[^>]+>/g, "")
+      .replace(/&lt;/g, "<").replace(/&gt;/g, ">")
+      .replace(/&amp;/g, "&").replace(/&quot;/g, '"');
+    expect(textOnly).toBe(input);
+  });
+
+  it("should still handle real _{} LaTeX notation with Unicode symbol", () => {
+    const input = "σ_{age > 20}(Person)";
+    const html = highlightRA(input);
+    const textOnly = html
+      .replace(/<[^>]+>/g, "")
+      .replace(/&lt;/g, "<").replace(/&gt;/g, ">")
+      .replace(/&amp;/g, "&").replace(/&quot;/g, '"');
+    expect(textOnly).toBe(input);
+    expect(html).toContain("opacity: 0.5"); // _ rendered as faint bracket
+  });
 });

src/ra-engine/raHighlight.ts

@@ -197,11 +197,14 @@ function renderSubscript(code: string, i: number, result: string[]): number {
     return i;
   }
 
-  // Handle _{ or _[ (LaTeX-style)
+  // Handle _{ or _[ (LaTeX-style) — only if followed by a bracket
   if (code[i] === "_") {
-    result.push(`<span style="${S.bracket}">${esc("_")}</span>`);
-    i++;
-    i = skipWs(code, i);
+    const afterUnderscore = skipWs(code, i + 1);
+    if (afterUnderscore < code.length && (code[afterUnderscore] === "{" || code[afterUnderscore] === "[")) {
+      result.push(`<span style="${S.bracket}">${esc("_")}</span>`);
+      i++;
+      i = skipWs(code, i);
+    }
   }
 
   // Try bracket content

src/ra-engine/raShared.ts

@@ -152,10 +152,12 @@ export function extractBracketContent(code: string, startI: number): [number, st
   let i = skipWs(code, startI);
   if (i >= code.length) return [startI, null];
 
-  // Handle _{ or _[ prefix
+  // Handle _{ or _[ prefix — only if followed by a bracket
   if (code[i] === "_") {
-    i++;
-    i = skipWs(code, i);
+    const afterUnderscore = skipWs(code, i + 1);
+    if (afterUnderscore < code.length && (code[afterUnderscore] === "{" || code[afterUnderscore] === "[")) {
+      i = afterUnderscore;
+    }
   }
 
   if (code[i] === "[" || code[i] === "{") {