add category to top 20 files

analytics.py

@@ -4,6 +4,7 @@ from datetime import datetime, timedelta, timezone
 import geoip2.database
 from auth import require_secret
 from collections import defaultdict
+import pandas as pd
 import json
 import os
 import auth
@@ -389,6 +390,16 @@ def dashboard():
         cursor = log_db.execute(query, params_for_filter)
         rows = cursor.fetchall()
 
+    # Convert rows to a list of dictionaries and add category
+    rows = [
+        {
+            'rel_path': rel_path,
+            'access_count': access_count,
+            'category': hf.extract_structure_from_string(rel_path)[0]
+        }
+        for rel_path, access_count in rows
+    ]
+
     # 2. Distinct device trend
     #    We'll group by hour if "today", by day if "7days"/"30days", by month if "365days"
     if session['timeframe'] == 'last24hours':
@@ -588,9 +599,6 @@ def dashboard():
     location_data.sort(key=lambda x: x['count'], reverse=True)
     location_data = location_data[:20]
 
-    # Convert the top-files rows to a list of dictionaries
-    rows = [dict(rel_path=r[0], access_count=r[1]) for r in rows]
-
     title_short = app_config.get('TITLE_SHORT', 'Default Title')
     title_long  = app_config.get('TITLE_LONG' , 'Default Title')
 
@@ -614,7 +622,6 @@ def dashboard():
 
 def export_to_excel():
     """Export search_db to an Excel file and store it locally."""
-    import pandas as pd
 
     # Query all data from the search_db
     query = "SELECT * FROM files"

helperfunctions.py

@@ -1,5 +1,6 @@
 from datetime import datetime
 import re
+import os
 
 def extract_date_from_string(string_with_date):
     # grab X.Y.Z where X,Y,Z are 1–4 digits
@@ -43,4 +44,49 @@ def extract_date_from_string(string_with_date):
         dt = datetime.strptime(date_str, fmt)
         return dt.strftime('%Y-%m-%d')
     except ValueError:
-        return None
+        return None
+
+
+def extract_structure_from_string(input_string):
+    # extract category and titel from filename
+    filename_ext = os.path.splitext(input_string)[0]
+    left_side, right_side = filename_ext.split('-', 1) if '-' in filename_ext else (filename_ext, None)
+    try:
+        int(left_side.strip())
+        # first part is only a number
+        previous_right_side = right_side
+        left_side, right_side = previous_right_side.split('-', 1) if '-' in previous_right_side else (previous_right_side, None)
+    except:
+        # first part not a number
+        pass
+    
+    if 'predig' in left_side.lower():
+        category = 'Predigt'
+    elif 'wort' in left_side.lower() or 'einladung' in left_side.lower():
+        category = 'Vorwort'
+    elif 'chor' in left_side.lower():
+        category = 'Chor'
+    elif 'orchester' in left_side.lower():
+        category = 'Orchester'
+    elif 'gruppenlied' in left_side.lower() or 'jugendlied' in left_side.lower():
+        category = 'Gruppenlied'
+    elif 'gemeinsam' in left_side.lower() or 'gesang' in left_side.lower() or 'lied' in left_side.lower():
+        category = 'Gemeinsamer Gesang'
+    elif 'gedicht' in left_side.lower():
+        category = 'Gedicht'
+    elif 'instrumental' in left_side.lower() or 'musikstück' in left_side.lower():
+        category = 'Instrumental'
+    else:
+        category = None
+
+    if right_side:
+        titel, name = right_side.split('-', 1) if '-' in right_side else (right_side, None)
+        if category == 'Predigt' or category == 'Vorwort' or category == 'Gedicht':
+            if not name: # kein Titel, nur Name
+                name = titel
+                titel = None
+    else:
+        titel = None
+        name = None
+
+    return category, titel, name

index_for_search.py

@@ -110,13 +110,13 @@ def updatefileindex():
                 # Retrieve the hit count for this file.
                 hit_count = get_hit_count(relative_path)
                 
-                category, titel, name, performance_date, site = None, None, None, None, None
-                
                 # Determine the site
                 if foldername == 'Gottesdienste Speyer':
                     site = 'Speyer'
                 elif foldername == 'Gottesdienste Schwegenheim':
                     site = 'Schwegenheim'
+                else:
+                    site = None
 
                 # Check for a corresponding transcript file in a sibling "Transkription" folder.
                 parent_dir = os.path.dirname(entry_path)
@@ -130,46 +130,7 @@ def updatefileindex():
                     except Exception:
                         transcript = None
 
-                # extract category and titel from filename
+                category, titel, name = hf.extract_structure_from_string(entry.name)
-                filename_ext = os.path.splitext(entry.name)[0]
-                left_side, right_side = filename_ext.split('-', 1) if '-' in filename_ext else (filename_ext, None)
-                try:
-                    int(left_side.strip())
-                    # first part is only a number
-                    previous_right_side = right_side
-                    left_side, right_side = previous_right_side.split('-', 1) if '-' in previous_right_side else (previous_right_side, None)
-                except:
-                    # first part not a number
-                    pass
-                
-                if 'predig' in left_side.lower():
-                    category = 'Predigt'
-                elif 'wort' in left_side.lower() or 'einladung' in left_side.lower():
-                    category = 'Vorwort'
-                elif 'chor' in left_side.lower():
-                    category = 'Chor'
-                elif 'orchester' in left_side.lower():
-                    category = 'Orchester'
-                elif 'gruppenlied' in left_side.lower() or 'jugendlied' in left_side.lower():
-                    category = 'Gruppenlied'
-                elif 'gemeinsam' in left_side.lower() or 'gesang' in left_side.lower() or 'lied' in left_side.lower():
-                    category = 'Gemeinsamer Gesang'
-                elif 'gedicht' in left_side.lower():
-                    category = 'Gedicht'
-                elif 'instrumental' in left_side.lower() or 'musikstück' in left_side.lower():
-                    category = 'Instrumental'
-                else:
-                    category = None
-
-                if right_side:
-                    titel, name = right_side.split('-', 1) if '-' in right_side else (right_side, None)
-                    if category == 'Predigt' or category == 'Vorwort' or category == 'Gedicht':
-                        if not name: # kein Titel, nur Name
-                            name = titel
-                            titel = None
-                else:
-                    titel = None
-                    name = None
                     
                 performance_date = hf.extract_date_from_string(relative_path)
 

templates/dashboard.html

@@ -222,6 +222,7 @@
                 <tr>
                   <th>File Path</th>
                   <th>Access Count</th>
+                  <th>Kategorie</th>
                 </tr>
               </thead>
               <tbody>
@@ -229,6 +230,7 @@
                 <tr>
                   <td>{{ row.rel_path }}</td>
                   <td>{{ row.access_count }}</td>
+                  <td>{{ row.category }}</td>
                 </tr>
                 {% else %}
                 <tr>