Regenerate HTML docs from real zip data; fix Yusuf Ali sura|ayah| prefix leak

[Copilot]

LFS pointer zips have been replaced with real archives. This PR regenerates all output text files and HTML pages from the actual source data, and fixes a pre-existing bug uncovered in the process.

Bug fix — src/gentxtforquran.py

en.yusufali.txt shares the sura|ayah|text pipe format with en.qarai.txt and en.hilali.txt, but was tagged English-Piped which fell through to the generic else branch — copying raw lines verbatim and leaking the 1|5| prefix into the HTML.

# Before — falls into else branch, copies "1|5|Thee do we worship..." verbatim
('data/en.yusufali.txt', 'output/quran_english_yusufali.txt', 'Abdullah Yusuf Ali', 'English-Piped'),

# After — uses split('|', 2)[2] to strip the sura|ayah| prefix
('data/en.yusufali.txt', 'output/quran_english_yusufali.txt', 'Abdullah Yusuf Ali', 'English-SuraAyah'),

Output files regenerated

• output/quran_hindi_mokhtasar.txt: was 4 lines (header only, zip was an LFS pointer) → now 6,582 lines of full Hindi Tafsir
• All 13 output files in output/ rebuilt from real source data

HTML docs regenerated

All 114 surah pages, index.html, search-data.js, and search.html rebuilt. Every ayah now renders complete content across all translation rows: Hindi Mokhtasar (full tafsir), Gujarati (Rabila), English Abridged, Hilali-Khan, and Yusuf Ali without the prefix artifact.

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.