Improved article selector with titles and structure

The previous code could not correctly handle nested XML elements in the TOC text entries.

frontend/package-lock.json

@@ -10,6 +10,7 @@
       "dependencies": {
         "@tanstack/react-query": "^5.74.4",
         "@tanstack/react-query-devtools": "^5.74.6",
+        "dompurify": "^3.2.5",
         "react": "^19.0.0",
         "react-dom": "^19.0.0",
         "react-router-dom": "^7.5.1",
@@ -1881,6 +1882,13 @@
         "@types/react": "^19.0.0"
       }
     },
+    "node_modules/@types/trusted-types": {
+      "version": "2.0.7",
+      "resolved": "https://registry.npmjs.org/@types/trusted-types/-/trusted-types-2.0.7.tgz",
+      "integrity": "sha512-ScaPdn1dQczgbl0QFTeTOmVHFULt394XJgOQNoyVhZ6r2vLnMLJfBPd53SB52T/3G36VI1/g2MZaX0cwDuXsfw==",
+      "license": "MIT",
+      "optional": true
+    },
     "node_modules/@typescript-eslint/eslint-plugin": {
       "version": "8.31.0",
       "resolved": "https://registry.npmjs.org/@typescript-eslint/eslint-plugin/-/eslint-plugin-8.31.0.tgz",
@@ -2414,6 +2422,15 @@
         "node": ">=0.10"
       }
     },
+    "node_modules/dompurify": {
+      "version": "3.2.5",
+      "resolved": "https://registry.npmjs.org/dompurify/-/dompurify-3.2.5.tgz",
+      "integrity": "sha512-mLPd29uoRe9HpvwP2TxClGQBzGXeEC/we/q+bFlmPPmj2p2Ugl3r6ATu/UU1v77DXNcehiBg9zsr1dREyA/dJQ==",
+      "license": "(MPL-2.0 OR Apache-2.0)",
+      "optionalDependencies": {
+        "@types/trusted-types": "^2.0.7"
+      }
+    },
     "node_modules/dotenv": {
       "version": "16.5.0",
       "resolved": "https://registry.npmjs.org/dotenv/-/dotenv-16.5.0.tgz",

frontend/package.json

@@ -12,6 +12,7 @@
   "dependencies": {
     "@tanstack/react-query": "^5.74.4",
     "@tanstack/react-query-devtools": "^5.74.6",
+    "dompurify": "^3.2.5",
     "react": "^19.0.0",
     "react-dom": "^19.0.0",
     "react-router-dom": "^7.5.1",

frontend/src/App.tsx

@@ -1,6 +1,3 @@
-import { useQueries } from "@tanstack/react-query";
-
-import { getArticleIds, getToc } from "./lib/api";
 import { Language } from "./lib/types";
 
 import ArticleSelector from "./components/ArticleSelector/ArticleSelector";
@@ -12,45 +9,29 @@ import useUIStore from "./store/uiStore";
 
 import styles from "./App.module.css";
 import CelexSelector from "./components/CelexSelector/CelexSelector";
+import { useTOC } from "./hooks/toc";
 
 function App() {
   const { numPanels, addPanel } = useUIStore();
   const { celexId, articleId } = useNavState();
-
+  const { data: toc, isPending, error } = useTOC();
-  const results = useQueries({
-    queries: [
-      {
-        queryKey: ["articleIds", celexId],
-        queryFn: () => getArticleIds(celexId!),
-        enabled: !!celexId,
-      },
-      {
-        queryKey: ["toc", celexId],
-        queryFn: () => getToc(celexId!, Language.ENG),
-        enabled: !!celexId,
-      },
-    ],
-  });
-
-  const isPending = results.some((result) => result.isPending);
-  const error = results.find((result) => result.isError);
 
   if (isPending) {
     return <div>Loading...</div>;
   }
   if (error) {
-    return <div>Error: {error.error?.message}</div>;
+    return <div>Error: {error.message}</div>;
   }
 
   return (
     <div className={styles.App}>
       <div className={styles.controls}>
         <CelexSelector />
-        <ArticleSelector articleIds={results[0].data!} />
+        <ArticleSelector toc={toc} />
         <button onClick={addPanel}>Add Panel</button>
       </div>
       <div className={styles.panelContainer}>
-        <TOC toc={results[1].data!} />
+        <TOC toc={toc} />
         {Array.from({ length: numPanels }, (_, index) => (
           <Panel
             key={index}

frontend/src/components/ArticleSelector/ArticleSelector.module.css

@@ -0,0 +1,4 @@
+.articleSelector {
+  width: 40ch;
+  text-overflow: ellipsis;
+}

frontend/src/components/ArticleSelector/ArticleSelector.tsx

@@ -1,32 +1,54 @@
+import { JSX } from "react";
+import type { Division } from "../../lib/types";
 import useNavState from "../../store/navStore";
+import styles from "./ArticleSelector.module.css";
 
 type ArticleSelectorProps = {
-  articleIds: number[];
+  toc: Division[];
 };
 
-function ArticleSelector({ articleIds }: ArticleSelectorProps) {
+function ArticleSelector({ toc }: ArticleSelectorProps) {
   const { articleId, setArticleId } = useNavState();
+
+  function renderDivision(div: Division): JSX.Element {
+    const contents = div.content.map((c) => {
+      if (c.type === "division") {
+        return renderDivision(c);
+      } else {
+        return (
+          <option key={c.id} value={c.id}>
+            {c.title}: {c.subtitle}
+          </option>
+        );
+      }
+    });
+
+    if (div.level === 0) {
+      const title = `${div.title}: ${div.subtitle}`;
+      return (
+        // For top-level divisions, we can use optgroup
+        <optgroup key={title} label={title}>
+          {contents}
+        </optgroup>
+      );
+    } else {
+      // HTML does not support nested optgroups, so we need to flatten the structure
+      return <>{contents}</>;
+    }
+  }
+
   return (
     <>
-      {articleId && articleId > 1 && (
-        <button onClick={() => setArticleId(articleId - 1)}>prev</button>
-      )}
       <select
         value={articleId || undefined}
+        className={styles.articleSelector}
         onChange={(e) => {
-          const id = parseInt(e.currentTarget.value);
+          const id = parseInt(e.target.value);
           setArticleId(id);
         }}
       >
-        {articleIds.map((id) => (
+        {toc.map((div) => renderDivision(div))}
-          <option key={id} value={id}>
-            Article {id}
-          </option>
-        ))}
       </select>
-      {articleId && articleId < articleIds[articleIds.length - 1] && (
-        <button onClick={() => setArticleId(articleId + 1)}>next</button>
-      )}
     </>
   );
 }

frontend/src/components/Panel/Panel.tsx

@@ -1,4 +1,5 @@
 import { useQuery } from "@tanstack/react-query";
+import DOMPurify from "dompurify";
 import { useEffect, useRef, useState } from "react";
 
 import { getArticle } from "../../lib/api";
@@ -84,7 +85,7 @@ function Panel({ celexId, language, articleId }: PanelProps) {
       <div
         ref={articleRef}
         lang={lang.substring(0, 2)}
-        dangerouslySetInnerHTML={{ __html: data || "" }}
+        dangerouslySetInnerHTML={{ __html: DOMPurify.sanitize(data) || "" }}
       />
     </div>
   );

frontend/src/components/TOC/TOC.tsx

@@ -3,10 +3,8 @@ import { Division } from "../../lib/types";
 import useNavState from "../../store/navStore";
 import styles from "./TOC.module.css";
 
-type TOC = Division[];
-
 type TOCProps = {
-  toc: TOC;
+  toc: Division[];
 };
 
 function TOC({ toc }: TOCProps) {

frontend/src/hooks/toc.ts

@@ -0,0 +1,14 @@
+import { useQuery } from "@tanstack/react-query";
+import { getToc } from "../lib/api";
+import { Language } from "../lib/types";
+import useNavState from "../store/navStore";
+
+export const useTOC = () => {
+  const celexId = useNavState((state) => state.celexId);
+  const query = useQuery({
+    queryKey: ["toc", celexId],
+    queryFn: () => getToc(celexId!, Language.ENG),
+    enabled: !!celexId,
+  });
+  return query;
+};

frontend/src/lib/api.ts

@@ -1,5 +1,4 @@
-import TOC from "../components/TOC/TOC";
+import { Division, Language } from "./types";
-import { Language } from "./types";
 
 const API_URL = import.meta.env.VITE_API_URL;
 
@@ -23,10 +22,13 @@ async function getArticleIds(celexId: string): Promise<number[]> {
   return await response.json();
 }
 
-async function getToc(celexId: string, language: Language): Promise<TOC> {
+async function getToc(
+  celexId: string,
+  language: Language
+): Promise<Division[]> {
   console.debug(`Fetching TOC for CELEX ID ${celexId}`);
   const response = await fetch(`${API_URL}/${celexId}/toc/${language}`);
-  return await response.json();
+  return (await response.json()) as Division[];
 }
 
 export { getArticle, getArticleIds, getToc };

src/formex_viewer/server.py

@@ -68,9 +68,16 @@ def article_ids(celex_id: str, language: Language = Language.ENG):
 
 @api_router.get("/{celex_id}/toc/{language}")
 def toc(celex_id: str, language: Language = Language.ENG):
+    def _extract_text(root: ET.Element, tag: str) -> str:
+        """
+        Extract text from the given tag in the XML element.
+        """
+        text = root.xpath(f"{tag}//text()")
+        return "".join(text) if text else ""
+
     def _handle_division(division: ET.Element, level: int):
-        title = ti_el[0] if (ti_el := division.xpath("TITLE/TI//text()")) else ""
+        title = _extract_text(division, "TITLE/TI")
-        subtitle = sti_el[0] if (sti_el := division.xpath("TITLE/STI//text()")) else ""
+        subtitle = _extract_text(division, "TITLE/STI")
 
         subdivisions = []
         for subdivision in division.xpath("DIVISION") or []:
@@ -81,10 +88,9 @@ def toc(celex_id: str, language: Language = Language.ENG):
             art_id = article.get("IDENTIFIER")
             if not art_id:
                 continue
-            art_title = ti_el[0] if (ti_el := article.xpath("TI.ART//text()")) else ""
+
-            art_subtitle = (
+            art_title = _extract_text(article, "TI.ART")
-                sti_el[0] if (sti_el := article.xpath("STI.ART//text()")) else ""
+            art_subtitle = _extract_text(article, "STI.ART")
-            )
             articles.append(
                 {
                     "id": int(art_id.lstrip("0")),