perf: fix 9 JS inefficiencies across /docs chapter, search, and bookmarks pages; add Devnagri Urdu (Maududi) translation

[Copilot]

Several hot-path JS routines in the generated /docs pages had algorithmic and DOM inefficiencies causing measurable jank on filter toggles, scroll, and search operations across all 114 chapter files.

Chapter pages (001.html–114.html)

• applyAllRows() — replaced O(n×m) nested Set+classList loop with Array.prototype.some.call(tr.classList, c => enabledTypes.has(c)), flipping the iteration to walk the small classList (2–3 items) and hit the Set in O(1)
• applyFs() — collapsed two querySelectorAll passes into one; also fixes a silent bug where td:not(.label) in the second pass overwrote .arabic-text cells with the wrong font size at non-default scale steps
• Scroll handler — gated the display toggle behind requestAnimationFrame + ticking flag to cap DOM writes at one per frame
• Permalink clicks — replaced up to 286 per-element addEventListener calls with a single delegated listener on document
• copyVerse — replaced tr.querySelector('.label') with direct tr.cells[0]

search.html

• highlight() — moved new RegExp(…) construction out of the per-result-per-field call into buildHighlightRegexes(), called once per search; previously created 200+ RegExp objects per page render
• doSearch() — pre-computes fieldArrays before the 6,236-row loop, eliminating repeated fieldCache[f[0]]||[] lookups per row
• renderPage() — same pre-cache pattern applied before the results render loop

bookmarks.html

• Delete button — replaced onclick="deleteBm(i)" (fragile stale-index binding) with data-key="s:a" and a single delegated listener; deletion now resolves by stable surah:ayah identity

Devnagri Urdu (Maududi) translation

Added the Devanagari-script Urdu translation by Abul Ala Maududi to the website:

• src/gentxtforquran.py — added a Devnagri-Urdu-HTML parser that strips HTML tags from data/quran_devnagri_urdu_maududi.html, normalizes split verse markers (e.g. [ \nS:A] → [S:A]), and extracts all [S:A] verse lines to produce output/quran_devnagri_urdu_maududi.txt (6,210 of 6,236 verses; 26 are absent in the upstream source HTML)
• src/gendocshtml.py — loads the new output file, adds a styled devnagri-urdu row (light blue-grey background, Devanagari font, indigo text) with dark mode support, adds it to the Content Filter checkbox list, writes docs/sd/devnagri-urdu.json for search, and registers it in search.html's FIELD_MAP

Approach

All JS and translation templates live in src/gendocshtml.py. Changes were made there and the full /docs directory regenerated — 114 chapter files + search.html + bookmarks.html + config.html + config.js + docs/sd/devnagri-urdu.json.

---

🔒 GitHub Advanced Security automatically protects Copilot coding agent pull requests. You can protect all pull requests by enabling Advanced Security for your repositories. Learn more about Advanced Security.

[druvx13]

Nice work; @copilot , i have added “quran_devnagri_urdu_maududi.html” in /data directory creates it's output in /output directory like in the format as others are and also include it in /docs website.

[Copilot]

Nice work; @copilot , i have added “quran_devnagri_urdu_maududi.html” in /data directory creates it's output in /output directory like in the format as others are and also include it in /docs webs...

Done in commit 069f726. Here's what was added:

• src/gentxtforquran.py: Added a Devnagri-Urdu-HTML parser that strips HTML tags, fixes split verse markers (e.g. [ \nS:A] → [S:A]), and extracts all [S:A] verse lines → output/quran_devnagri_urdu_maududi.txt (6,210 verses; 26 are absent in the source HTML).
• src/gendocshtml.py: Added the translation to all chapter pages (light blue-grey row, Devanagari font), the Content Filter checkbox, docs/sd/devnagri-urdu.json for search, and the FIELD_MAP in search.html.
• All 114 chapter HTMLs, search.html, config.html, config.js, and docs/sd/devnagri-urdu.json were regenerated.

The new "Devnagri Urdu (Maududi)" row now appears in each verse alongside the other translations and is searchable from search.html.