project context updated

PROJECT_CONTEXT.md

@@ -57,11 +57,14 @@ knowledgebase/
 │   ├── essays/                 → 23 essays across 6 groups
 │   ├── journals/               → Jaladuta Diary, The Beginning
 │   ├── other-documents/        → 12 miscellaneous documents
-│   ├── glossary/               → 2,749 searchable Sanskrit terms
+│   ├── glossary/
+│   │   ├── page.tsx            → Redirects to /glossary/a
+│   │   ├── glossary-client.tsx → Client component (sidebar terms, single entry view, prev/next nav)
+│   │   └── [letter]/page.tsx   → Dynamic route per letter (A-Z)
 │   ├── disciples/              → 4,747 initiated disciples directory
 │   │   ├── page.tsx            → Disciples browser (letter nav, suffix filter, search, sort, pagination)
 │   │   └── disciples-browser.tsx → Client component
-│   ├── transcriptions/page.tsx → Coming Soon page (lectures & conversations)
+│   ├── transcriptions/page.tsx → Coming Soon page
 │   └── api/
 │       ├── og/route.tsx        → Dynamic OG image generator (Noto Sans, Prabhupāda image)
 │       ├── books/              → Book/verse API routes
@@ -662,6 +665,7 @@ SSR and DS have parent chapters with sub-chapters stored as individual `prose_ch
 - **Copy** appends source reference: title, chapter/verse, highlight URL
 - All toolbar surfaces use dark contrast treatment (bg-black, border-white/10)
 - Kautilya lotus icon for the AI companion button
+- On iOS Safari mobile, selection is saved/restored when drawers open/close to prevent focus conflicts
 
 ### Text Highlight Links
 - Selected text encoded as compact URL fragment: `?hl=prefix§suffix§length`
@@ -670,14 +674,15 @@ SSR and DS have parent chapters with sub-chapters stored as individual `prose_ch
 - Incoming `?hl=` URLs: text is found in DOM, wrapped in `<mark>`, scrolled to, highlighted permanently
 - `TextHighlighter` component runs on page load via `useSearchParams`
 - Server-side `extractMatchedText()` searches paragraphs for OG image quote extraction
+- Case-insensitive matching for both DOM highlighting and server-side text extraction
+- Cross-paragraph matching via position mapping (spaces inserted at element boundaries)
 
 ### Dynamic OG Images
 - `/api/og` route generates 1200×630 PNG images using `next/og` (Satori)
 - Left panel: Prabhupāda image with quote mark and branding (from PNG)
 - Right panel: title (bold), detail, quote text, "Read more at prabhupadasays.org"
-- Font: Noto Sans (Regular + Bold TTF from `public/fonts/Noto_Sans/static/`)
-- Full IAST diacritical support (ā, ī, ū, ṛ, ṣ, ṇ, ṭ, ḍ, ś, ñ, ṅ, ḥ, ṁ)
-- Every content page has `generateMetadata` with contextual fallback descriptions
+- Font: Noto Sans Regular + Bold loaded from `public/fonts/Noto_Sans/static/` — full IAST diacritical support
+- All 104 content pages have `generateMetadata` with contextual fallback descriptions
 - When `?hl=` is present, the actual selected text is extracted from DB content for the OG quote
 
 ### Site Footer
@@ -692,7 +697,7 @@ SSR and DS have parent chapters with sub-chapters stored as individual `prose_ch
 
 ```
 DATABASE_URL=mongodb://user:pass@host:port/spsays?authSource=admin
-NEXT_PUBLIC_SITE_URL=https://prabhupadasays.org   # used for OG image URLs
+NEXT_PUBLIC_SITE_URL=https://prabhupadasays.org
 ```
 
 ---
@@ -710,6 +715,9 @@ npx prisma generate                  # Regenerate Prisma client after schema cha
 npx next build
 
 # Extraction & seeding scripts are in extractor/ — see extractor/scripts/
+
+# BFF API server (from ../spsays-bff/)
+make run                             # Build + start Go API on :8080
 ```
 
 ---
@@ -793,6 +801,8 @@ npx next build
 }
 ```
 
+Glossary is browsed via `/glossary/{letter}` routes (A-Z). Each letter page shows terms grouped by base letter (diacritical variants mapped: Ā→A, Ś→S, Ṛ→R, etc.). Single entry view with prev/next navigation.
+
 ---
 
 ## Known Issues