feat: implement comprehensive ZIP export per v9c specification

Complete rewrite of ZIP export to match CLAUDE.md specification.

**Structure:**
- README.txt (format explanation)
- profile.json (no password hash, includes stats)
- data/*.csv (5 separate CSV files)
- insights/ai_insights.json
- photos/*.jpg

**CSV Format:**
- Delimiter: semicolon (;)
- Encoding: UTF-8 with BOM (Excel compatible)
- Date format: YYYY-MM-DD
- Decimal separator: dot (.)
- NULL values: empty string
- First row: header

**Files:**
- data/weight.csv (id, date, weight, note, source, created)
- data/circumferences.csv (8 measurement points)
- data/caliper.csv (skinfold measurements + bf%)
- data/nutrition.csv (kcal, protein, fat, carbs)
- data/activity.csv (type, duration, kcal, HR, distance)

**Filename:** mitai-export-{name}-{YYYY-MM-DD}.zip

Ready for import functionality (v9c).

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

backend/main.py

@@ -1519,67 +1519,196 @@ def export_json(x_profile_id: Optional[str]=Header(default=None), session: dict=
 
 @app.get("/api/export/zip")
 def export_zip(x_profile_id: Optional[str]=Header(default=None), session: dict=Depends(require_auth)):
-    """Export all data as ZIP (JSON + photos)."""
+    """Export all data as ZIP (CSV + JSON + photos) per specification."""
     pid = get_pid(x_profile_id)
 
-    # Check export permission
+    # Check export permission & get profile
     with get_db() as conn:
         cur = get_cursor(conn)
-        cur.execute("SELECT export_enabled FROM profiles WHERE id=%s", (pid,))
+        cur.execute("SELECT * FROM profiles WHERE id=%s", (pid,))
-        prof = cur.fetchone()
+        prof = r2d(cur.fetchone())
-        if not prof or not prof['export_enabled']:
+        if not prof or not prof.get('export_enabled'):
             raise HTTPException(403, "Export ist für dieses Profil deaktiviert")
 
-    # Create ZIP in memory
+    # Helper: CSV writer with UTF-8 BOM + semicolon
+    def write_csv(zf, filename, rows, columns):
+        if not rows:
+            return
+        output = io.StringIO()
+        writer = csv.writer(output, delimiter=';')
+        writer.writerow(columns)
+        for r in rows:
+            writer.writerow([
+                '' if r.get(col) is None else
+                (float(r[col]) if isinstance(r.get(col), Decimal) else r[col])
+                for col in columns
+            ])
+        # UTF-8 with BOM for Excel
+        csv_bytes = '\ufeff'.encode('utf-8') + output.getvalue().encode('utf-8')
+        zf.writestr(f"data/{filename}", csv_bytes)
+
+    # Create ZIP
     zip_buffer = io.BytesIO()
+    export_date = datetime.now().strftime('%Y-%m-%d')
+    profile_name = prof.get('name', 'export')
+
     with zipfile.ZipFile(zip_buffer, 'w', zipfile.ZIP_DEFLATED) as zf:
-        # Add JSON data
-        data = {}
         with get_db() as conn:
             cur = get_cursor(conn)
 
-            cur.execute("SELECT * FROM profiles WHERE id=%s", (pid,))
+            # 1. README.txt
-            data['profile'] = r2d(cur.fetchone())
+            readme = f"""Mitai Jinkendo – Datenexport
+Version: 2
+Exportiert am: {export_date}
+Profil: {profile_name}
 
-            cur.execute("SELECT * FROM weight_log WHERE profile_id=%s ORDER BY date", (pid,))
+Inhalt:
-            data['weight'] = [r2d(r) for r in cur.fetchall()]
+- profile.json: Profildaten und Einstellungen
+- data/*.csv:   Messdaten (Semikolon-getrennt, UTF-8)
+- insights/:    KI-Auswertungen (JSON)
+- photos/:      Progress-Fotos (JPEG)
 
-            cur.execute("SELECT * FROM circumference_log WHERE profile_id=%s ORDER BY date", (pid,))
+Import:
-            data['circumferences'] = [r2d(r) for r in cur.fetchall()]
+Dieser Export kann in Mitai Jinkendo unter
+Einstellungen → Import → "Mitai Backup importieren"
+wieder eingespielt werden.
 
-            cur.execute("SELECT * FROM caliper_log WHERE profile_id=%s ORDER BY date", (pid,))
+Format-Version 2 (ab v9b):
-            data['caliper'] = [r2d(r) for r in cur.fetchall()]
+Alle CSV-Dateien sind UTF-8 mit BOM kodiert.
+Trennzeichen: Semikolon (;)
+Datumsformat: YYYY-MM-DD
+"""
+            zf.writestr("README.txt", readme.encode('utf-8'))
 
-            cur.execute("SELECT * FROM nutrition_log WHERE profile_id=%s ORDER BY date", (pid,))
+            # 2. profile.json (ohne Passwort-Hash)
-            data['nutrition'] = [r2d(r) for r in cur.fetchall()]
+            cur.execute("SELECT COUNT(*) as c FROM weight_log WHERE profile_id=%s", (pid,))
+            w_count = cur.fetchone()['c']
+            cur.execute("SELECT COUNT(*) as c FROM nutrition_log WHERE profile_id=%s", (pid,))
+            n_count = cur.fetchone()['c']
+            cur.execute("SELECT COUNT(*) as c FROM activity_log WHERE profile_id=%s", (pid,))
+            a_count = cur.fetchone()['c']
+            cur.execute("SELECT COUNT(*) as c FROM photos WHERE profile_id=%s", (pid,))
+            p_count = cur.fetchone()['c']
 
-            cur.execute("SELECT * FROM activity_log WHERE profile_id=%s ORDER BY date", (pid,))
+            profile_data = {
-            data['activity'] = [r2d(r) for r in cur.fetchall()]
+                "export_version": "2",
+                "export_date": export_date,
+                "app": "Mitai Jinkendo",
+                "profile": {
+                    "name": prof.get('name'),
+                    "email": prof.get('email'),
+                    "sex": prof.get('sex'),
+                    "height": float(prof['height']) if prof.get('height') else None,
+                    "birth_year": int(prof['dob'][:4]) if prof.get('dob') else None,
+                    "goal_weight": float(prof['goal_weight']) if prof.get('goal_weight') else None,
+                    "goal_bf_pct": float(prof['goal_bf_pct']) if prof.get('goal_bf_pct') else None,
+                    "avatar_color": prof.get('avatar_color'),
+                    "auth_type": prof.get('auth_type'),
+                    "session_days": prof.get('session_days'),
+                    "ai_enabled": prof.get('ai_enabled'),
+                    "tier": prof.get('tier')
+                },
+                "stats": {
+                    "weight_entries": w_count,
+                    "nutrition_entries": n_count,
+                    "activity_entries": a_count,
+                    "photos": p_count
+                }
+            }
+            zf.writestr("profile.json", json.dumps(profile_data, indent=2, ensure_ascii=False).encode('utf-8'))
 
-            cur.execute("SELECT * FROM ai_insights WHERE profile_id=%s ORDER BY created DESC", (pid,))
+            # 3. data/weight.csv
-            data['insights'] = [r2d(r) for r in cur.fetchall()]
+            cur.execute("SELECT id, date, weight, note, source, created FROM weight_log WHERE profile_id=%s ORDER BY date", (pid,))
+            write_csv(zf, "weight.csv", [r2d(r) for r in cur.fetchall()],
+                     ['id','date','weight','note','source','created'])
 
-        def decimal_handler(obj):
+            # 4. data/circumferences.csv
-            if isinstance(obj, Decimal):
+            cur.execute("SELECT id, date, c_waist, c_hip, c_chest, c_neck, c_arm, c_thigh, c_calf, notes, created FROM circumference_log WHERE profile_id=%s ORDER BY date", (pid,))
-                return float(obj)
+            rows = [r2d(r) for r in cur.fetchall()]
-            return str(obj)
+            # Rename columns to match spec
+            for r in rows:
+                r['waist'] = r.pop('c_waist', None)
+                r['hip'] = r.pop('c_hip', None)
+                r['chest'] = r.pop('c_chest', None)
+                r['neck'] = r.pop('c_neck', None)
+                r['upper_arm'] = r.pop('c_arm', None)
+                r['thigh'] = r.pop('c_thigh', None)
+                r['calf'] = r.pop('c_calf', None)
+                r['forearm'] = None  # not tracked
+                r['note'] = r.pop('notes', None)
+            write_csv(zf, "circumferences.csv", rows,
+                     ['id','date','waist','hip','chest','neck','upper_arm','thigh','calf','forearm','note','created'])
 
-        zf.writestr("data.json", json.dumps(data, indent=2, default=decimal_handler))
+            # 5. data/caliper.csv
+            cur.execute("SELECT id, date, sf_chest, sf_abdomen, sf_thigh, sf_triceps, sf_subscap, sf_suprailiac, sf_axilla, sf_method, body_fat_pct, notes, created FROM caliper_log WHERE profile_id=%s ORDER BY date", (pid,))
+            rows = [r2d(r) for r in cur.fetchall()]
+            for r in rows:
+                r['chest'] = r.pop('sf_chest', None)
+                r['abdomen'] = r.pop('sf_abdomen', None)
+                r['thigh'] = r.pop('sf_thigh', None)
+                r['tricep'] = r.pop('sf_triceps', None)
+                r['subscapular'] = r.pop('sf_subscap', None)
+                r['suprailiac'] = r.pop('sf_suprailiac', None)
+                r['midaxillary'] = r.pop('sf_axilla', None)
+                r['method'] = r.pop('sf_method', None)
+                r['bf_percent'] = r.pop('body_fat_pct', None)
+                r['note'] = r.pop('notes', None)
+            write_csv(zf, "caliper.csv", rows,
+                     ['id','date','chest','abdomen','thigh','tricep','subscapular','suprailiac','midaxillary','method','bf_percent','note','created'])
 
-        # Add photos if they exist
+            # 6. data/nutrition.csv
-        with get_db() as conn:
+            cur.execute("SELECT id, date, kcal, protein_g, fat_g, carbs_g, source, created FROM nutrition_log WHERE profile_id=%s ORDER BY date", (pid,))
-            cur = get_cursor(conn)
+            rows = [r2d(r) for r in cur.fetchall()]
+            for r in rows:
+                r['meal_name'] = ''  # not tracked per meal
+                r['protein'] = r.pop('protein_g', None)
+                r['fat'] = r.pop('fat_g', None)
+                r['carbs'] = r.pop('carbs_g', None)
+                r['fiber'] = None  # not tracked
+                r['note'] = ''
+            write_csv(zf, "nutrition.csv", rows,
+                     ['id','date','meal_name','kcal','protein','fat','carbs','fiber','note','source','created'])
+
+            # 7. data/activity.csv
+            cur.execute("SELECT id, date, activity_type, duration_min, kcal_active, hr_avg, hr_max, distance_km, notes, source, created FROM activity_log WHERE profile_id=%s ORDER BY date", (pid,))
+            rows = [r2d(r) for r in cur.fetchall()]
+            for r in rows:
+                r['name'] = r['activity_type']
+                r['type'] = r.pop('activity_type', None)
+                r['kcal'] = r.pop('kcal_active', None)
+                r['heart_rate_avg'] = r.pop('hr_avg', None)
+                r['heart_rate_max'] = r.pop('hr_max', None)
+                r['note'] = r.pop('notes', None)
+            write_csv(zf, "activity.csv", rows,
+                     ['id','date','name','type','duration_min','kcal','heart_rate_avg','heart_rate_max','distance_km','note','source','created'])
+
+            # 8. insights/ai_insights.json
+            cur.execute("SELECT id, scope, content, created FROM ai_insights WHERE profile_id=%s ORDER BY created DESC", (pid,))
+            insights = []
+            for r in cur.fetchall():
+                rd = r2d(r)
+                insights.append({
+                    "id": rd['id'],
+                    "scope": rd['scope'],
+                    "created": rd['created'].isoformat() if hasattr(rd['created'], 'isoformat') else str(rd['created']),
+                    "result": rd['content']
+                })
+            if insights:
+                zf.writestr("insights/ai_insights.json", json.dumps(insights, indent=2, ensure_ascii=False).encode('utf-8'))
+
+            # 9. photos/
             cur.execute("SELECT * FROM photos WHERE profile_id=%s ORDER BY date", (pid,))
             photos = [r2d(r) for r in cur.fetchall()]
-
             for i, photo in enumerate(photos):
                 photo_path = Path(PHOTOS_DIR) / photo['path']
                 if photo_path.exists():
-                    zf.write(photo_path, f"photos/{photo['date'] or i}_{photo_path.name}")
+                    filename = f"{photo.get('date') or export_date}_{i+1}{photo_path.suffix}"
+                    zf.write(photo_path, f"photos/{filename}")
 
     zip_buffer.seek(0)
+    filename = f"mitai-export-{profile_name.replace(' ','-')}-{export_date}.zip"
     return StreamingResponse(
         iter([zip_buffer.getvalue()]),
         media_type="application/zip",
-        headers={"Content-Disposition": f"attachment; filename=mitai-export-{pid}.zip"}
+        headers={"Content-Disposition": f"attachment; filename={filename}"}
     )