Add NEW badge for recent listings, text input for price filter, cleanup

- New listings (≤1 day) show yellow NEW badge instead of oversized marker
- Price filter changed from dropdown to text input (max 14M)
- Cap price filter at 14M in JS
- Remove unused regen_map.py
- Remove unused HTMLParser import in scrape_idnes.py

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

merge_and_map.py

@@ -79,6 +79,10 @@ def main():
         if key in seen_keys:
             dupes += 1
             existing = seen_keys[key]
+            # Preserve earliest first_seen across sources
+            dup_fs = e.get("first_seen", "")
+            if dup_fs and (not existing.get("first_seen") or dup_fs < existing["first_seen"]):
+                existing["first_seen"] = dup_fs
             # Log it
             print(f"  Duplikát: {e['locality']} | {format_price(e['price'])} | {e.get('area', '?')} m² "
                   f"({e.get('source', '?')} vs {existing.get('source', '?')})")

regen_map.py

@@ -1,114 +0,0 @@
-#!/usr/bin/env python3
-"""
-Přegeneruje mapu z již stažených dat (byty_sreality.json).
-Doplní chybějící plochy ze Sreality API, opraví URL, aplikuje filtry.
-"""
-from __future__ import annotations
-
-import json
-import time
-import urllib.request
-from pathlib import Path
-
-from scrape_and_map import (
-    generate_map, format_price, MIN_AREA, HEADERS, DETAIL_API
-)
-
-
-def api_get(url: str) -> dict:
-    req = urllib.request.Request(url, headers=HEADERS)
-    with urllib.request.urlopen(req, timeout=30) as resp:
-        return json.loads(resp.read().decode("utf-8"))
-
-
-def fix_sreality_url(estate: dict) -> str:
-    """Fix the Sreality URL to include disposition segment (only if missing)."""
-    disp = estate.get("disposition", "")
-    slug_map = {
-        "1+kk": "1+kk", "1+1": "1+1", "2+kk": "2+kk", "2+1": "2+1",
-        "3+kk": "3+kk", "3+1": "3+1", "4+kk": "4+kk", "4+1": "4+1",
-        "5+kk": "5+kk", "5+1": "5+1", "6+": "6-a-vice", "Atypický": "atypicky",
-    }
-    slug = slug_map.get(disp, "byt")
-    old_url = estate.get("url", "")
-    parts = old_url.split("/")
-    try:
-        byt_idx = parts.index("byt")
-        # Only insert if disposition slug is not already there
-        if byt_idx + 1 < len(parts) and parts[byt_idx + 1] == slug:
-            return old_url  # already correct
-        parts.insert(byt_idx + 1, slug)
-        return "/".join(parts)
-    except ValueError:
-        return old_url
-
-
-def fetch_area(hash_id: int) -> int | None:
-    """Fetch area from detail API."""
-    try:
-        url = DETAIL_API.format(hash_id)
-        detail = api_get(url)
-        for item in detail.get("items", []):
-            name = item.get("name", "")
-            if "žitná ploch" in name or "zitna ploch" in name.lower():
-                return int(item["value"])
-    except Exception:
-        pass
-    return None
-
-
-def main():
-    json_path = Path("byty_sreality.json")
-    if not json_path.exists():
-        print("Soubor byty_sreality.json nenalezen. Nejprve spusť scrape_and_map.py")
-        return
-
-    estates = json.loads(json_path.read_text(encoding="utf-8"))
-    print(f"Načteno {len(estates)} bytů z byty_sreality.json")
-
-    # Step 1: Fetch missing areas
-    missing_area = [e for e in estates if e.get("area") is None]
-    print(f"Doplňuji plochu u {len(missing_area)} bytů...")
-
-    for i, e in enumerate(missing_area):
-        time.sleep(0.3)
-        area = fetch_area(e["hash_id"])
-        if area is not None:
-            e["area"] = area
-        if (i + 1) % 50 == 0:
-            print(f"  {i + 1}/{len(missing_area)} ...")
-
-    # Count results
-    with_area = sum(1 for e in estates if e.get("area") is not None)
-    print(f"Plocha doplněna: {with_area}/{len(estates)}")
-
-    # Step 2: Fix URLs
-    for e in estates:
-        e["url"] = fix_sreality_url(e)
-
-    # Step 3: Filter by min area
-    filtered = []
-    excluded = 0
-    for e in estates:
-        area = e.get("area")
-        if area is not None and area < MIN_AREA:
-            excluded += 1
-            continue
-        filtered.append(e)
-
-    print(f"Vyloučeno (< {MIN_AREA} m²): {excluded}")
-    print(f"Zbývá: {len(filtered)} bytů")
-
-    # Save updated data
-    filtered_path = Path("byty_sreality.json")
-    filtered_path.write_text(
-        json.dumps(filtered, ensure_ascii=False, indent=2),
-        encoding="utf-8",
-    )
-
-    # Generate map
-    generate_map(filtered)
-
-
-if __name__ == "__main__":
-    main()

scrape_and_map.py

@@ -13,7 +13,7 @@ import math
 import time
 import urllib.request
 import urllib.parse
-from datetime import datetime
+from datetime import datetime, timedelta
 from pathlib import Path
 from scraper_stats import write_stats
 
@@ -353,6 +353,8 @@ def scrape(max_pages: int | None = None, max_properties: int | None = None):
             "url": sreality_url(hash_id, seo),
             "image": (estate.get("_links", {}).get("images", [{}])[0].get("href", "") if estate.get("_links", {}).get("images") else ""),
             "scraped_at": datetime.now().strftime("%Y-%m-%d"),
+            "first_seen": cached.get("first_seen", datetime.now().strftime("%Y-%m-%d")) if cached else datetime.now().strftime("%Y-%m-%d"),
+            "last_changed": datetime.now().strftime("%Y-%m-%d"),
         }
         results.append(result)
         details_fetched += 1
@@ -429,18 +431,30 @@ def generate_map(estates: list[dict], output_path: str = "mapa_bytu.html"):
     ]
     for bcolor, blabel in bands:
         price_legend_items += (
-            f'<div style="display:flex;align-items:center;gap:6px;margin:2px 0;">'
+            f'<div class="price-band" data-color="{bcolor}" onclick="toggleColorFilter(\'{bcolor}\')" '
+            f'style="display:flex;align-items:center;gap:6px;margin:2px 0;padding:2px 4px;'
+            f'border-radius:4px;border:2px solid transparent;">'
             f'<span style="width:14px;height:14px;border-radius:50%;background:{bcolor};'
             f'display:inline-block;border:2px solid white;box-shadow:0 1px 3px rgba(0,0,0,0.3);flex-shrink:0;"></span>'
             f'<span>{blabel}</span></div>'
         )
+    price_legend_items += (
+        '<div id="price-filter-reset" style="display:none;margin:3px 0 0 4px;">'
+        '<a href="#" onclick="resetColorFilter();return false;" '
+        'style="font-size:11px;color:#1976D2;text-decoration:none;">✕ Zobrazit všechny ceny</a>'
+        '</div>'
+    )
     # New marker indicator — bigger dot, no extra border
     price_legend_items += (
         '<div style="display:flex;align-items:center;gap:6px;margin:6px 0 0 0;'
         'padding-top:6px;border-top:1px solid #eee;">'
-        '<span style="width:18px;height:18px;border-radius:50%;background:#66BB6A;'
+        '<span style="display:inline-flex;align-items:center;gap:3px;flex-shrink:0;">'
-        'display:inline-block;box-shadow:0 1px 4px rgba(0,0,0,0.35);flex-shrink:0;"></span>'
+        '<span style="width:14px;height:14px;border-radius:50%;background:#66BB6A;'
-        '<span>Nové (z dnešního scrapu) — větší</span></div>'
+        'display:inline-block;box-shadow:0 1px 3px rgba(0,0,0,0.3);"></span>'
+        '<span style="font-size:8px;font-weight:700;background:#FFD600;color:#333;'
+        'padding:1px 3px;border-radius:2px;">NEW</span>'
+        '</span>'
+        '<span>Nové (≤ 1 den)</span></div>'
     )
 
     markers_js = ""
@@ -464,16 +478,30 @@ def generate_map(estates: list[dict], output_path: str = "mapa_bytu.html"):
 
         hash_id = e.get("hash_id", "")
 
-        scraped_at = e.get("scraped_at", "")
+        first_seen = e.get("first_seen", "")
-        is_new = scraped_at == datetime.now().strftime("%Y-%m-%d")
+        last_changed = e.get("last_changed", "")
+        today = datetime.now().strftime("%Y-%m-%d")
+        yesterday = (datetime.now() - timedelta(days=1)).strftime("%Y-%m-%d")
+        is_new = first_seen in (today, yesterday)
 
         new_badge = (
             '<span style="margin-left:6px;font-size:11px;background:#FFD600;color:#333;'
             'padding:1px 6px;border-radius:3px;font-weight:bold;">NOVÉ</span>'
             if is_new else ""
         )
+
+        date_parts = []
+        if first_seen:
+            date_parts.append(f'Přidáno: {first_seen}')
+        if last_changed and last_changed != first_seen:
+            date_parts.append(f'Změněno: {last_changed}')
+        date_row = (
+            f'<span style="font-size:11px;color:#888;">{"&nbsp;·&nbsp;".join(date_parts)}</span><br>'
+            if date_parts else ""
+        )
+
         popup = (
-            f'<div style="min-width:280px;font-family:system-ui,sans-serif;" data-hashid="{hash_id}">'
+            f'<div style="min-width:280px;font-family:system-ui,sans-serif;" data-hashid="{hash_id}" data-first-seen="{first_seen}" data-last-changed="{last_changed}">'
             f'<b style="font-size:14px;">{format_price(e["price"])}</b>'
             f'<span style="margin-left:8px;font-size:11px;background:{source_color};color:white;'
             f'padding:1px 6px;border-radius:3px;">{source_label}</span>{new_badge}<br>'
@@ -481,7 +509,9 @@ def generate_map(estates: list[dict], output_path: str = "mapa_bytu.html"):
             f'{floor_note}<br><br>'
             f'<b>{e["locality"]}</b><br>'
             f'Stavba: {building_text}<br>'
-            f'Vlastnictví: {ownership_text}<br><br>'
+            f'Vlastnictví: {ownership_text}<br>'
+            f'{date_row}'
+            f'<br>'
             f'<a href="{e["url"]}" target="_blank" '
             f'style="color:{source_color};text-decoration:none;font-weight:bold;">'
             f'→ Otevřít na {source_label}</a>'
@@ -513,7 +543,7 @@ def generate_map(estates: list[dict], output_path: str = "mapa_bytu.html"):
         else:
             marker_fn = "addMarker"
         markers_js += (
-            f"  {marker_fn}({e['lat']}, {e['lon']}, '{color}', '{popup}', '{hash_id}');\n"
+            f"  {marker_fn}({e['lat']}, {e['lon']}, '{color}', '{popup}', '{hash_id}', '{first_seen}', '{last_changed}');\n"
         )
 
     # Build legend — price per m² bands + disposition counts
@@ -579,12 +609,12 @@ def generate_map(estates: list[dict], output_path: str = "mapa_bytu.html"):
   .heart-icon-fav svg path {{ stroke: gold !important; stroke-width: 2.5 !important; filter: drop-shadow(0 0 4px rgba(255,193,7,0.7)); }}
   .heart-icon-rej {{ opacity: 0.4 !important; filter: grayscale(1); }}
   .reject-overlay {{ background: none !important; border: none !important; pointer-events: none !important; }}
-  @keyframes pulse-new {{
+  .new-badge-icon {{ background: none !important; border: none !important; pointer-events: none !important; }}
-    0%   {{ stroke-opacity: 1;   stroke-width: 3px; r: 11; }}
+  .new-badge {{
-    50%  {{ stroke-opacity: 0.4; stroke-width: 6px; r: 12; }}
+    font-size: 9px; font-weight: 700; color: #333; background: #FFD600;
-    100% {{ stroke-opacity: 1;   stroke-width: 3px; r: 11; }}
+    padding: 1px 4px; border-radius: 3px; white-space: nowrap;
+    box-shadow: 0 1px 3px rgba(0,0,0,0.3); letter-spacing: 0.5px;
   }}
-  .marker-new {{ animation: pulse-new 2s ease-in-out infinite; }}
   .info-panel {{
     position: absolute; top: 10px; right: 10px; z-index: 1000;
     background: white; padding: 16px; border-radius: 10px;
@@ -617,6 +647,10 @@ def generate_map(estates: list[dict], output_path: str = "mapa_bytu.html"):
   .info-panel .stats {{ color: #666; margin-bottom: 10px; padding-bottom: 10px; border-bottom: 1px solid #eee; }}
   .filter-section {{ margin-top: 10px; padding-top: 10px; border-top: 1px solid #eee; }}
   .filter-section label {{ display: flex; align-items: center; gap: 6px; margin: 3px 0; cursor: pointer; }}
+  .price-band {{ cursor: pointer; transition: background 0.12s; }}
+  .price-band:hover {{ background: #f0f0f0; }}
+  .price-band.active {{ border-color: #333 !important; background: #e8f0fe; }}
+  .price-band.dimmed {{ opacity: 0.35; }}
   .filter-section input[type="checkbox"] {{ accent-color: #1976D2; }}
   #floor-filter {{ margin-top: 8px; }}
   #floor-filter select {{ width: 100%; padding: 4px; border-radius: 4px; border: 1px solid #ccc; }}
@@ -655,11 +689,23 @@ def generate_map(estates: list[dict], output_path: str = "mapa_bytu.html"):
     </div>
     <div style="margin-top:6px;">
       <label>Max cena:
-        <select id="max-price" onchange="applyFilters()">
+        <input type="number" id="max-price" value="13500000" max="14000000" step="500000"
-          <option value="13500000">13 500 000 Kč</option>
+          style="width:130px;padding:2px 4px;border:1px solid #ccc;border-radius:3px;"
-          <option value="12000000">12 000 000 Kč</option>
+          onchange="applyFilters()" onkeyup="applyFilters()"> Kč
-          <option value="10000000">10 000 000 Kč</option>
+      </label>
-          <option value="8000000">8 000 000 Kč</option>
+    </div>
+    <div style="margin-top:6px;">
+      <label>Přidáno / změněno:
+        <select id="days-filter" onchange="applyFilters()" style="width:100%;padding:4px;border-radius:4px;border:1px solid #ccc;">
+          <option value="0">Vše</option>
+          <option value="1">za 1 den</option>
+          <option value="2">za 2 dny</option>
+          <option value="3">za 3 dny</option>
+          <option value="4">za 4 dny</option>
+          <option value="5">za 5 dní</option>
+          <option value="7">za 7 dní</option>
+          <option value="14">za 14 dní</option>
+          <option value="30">za 30 dní</option>
         </select>
       </label>
     </div>
@@ -693,9 +739,39 @@ L.tileLayer('https://{{s}}.basemaps.cartocdn.com/light_only_labels/{{z}}/{{x}}/{
   pane: 'shadowPane',
 }}).addTo(map);
 
+var selectedColors = [];
+
+function toggleColorFilter(color) {{
+  var idx = selectedColors.indexOf(color);
+  if (idx >= 0) selectedColors.splice(idx, 1);
+  else selectedColors.push(color);
+  document.querySelectorAll('.price-band').forEach(function(el) {{
+    var c = el.getAttribute('data-color');
+    if (selectedColors.length === 0) {{
+      el.classList.remove('active', 'dimmed');
+    }} else if (selectedColors.indexOf(c) >= 0) {{
+      el.classList.add('active'); el.classList.remove('dimmed');
+    }} else {{
+      el.classList.add('dimmed'); el.classList.remove('active');
+    }}
+  }});
+  document.getElementById('price-filter-reset').style.display =
+    selectedColors.length > 0 ? 'block' : 'none';
+  applyFilters();
+}}
+
+function resetColorFilter() {{
+  selectedColors = [];
+  document.querySelectorAll('.price-band').forEach(function(el) {{
+    el.classList.remove('active', 'dimmed');
+  }});
+  document.getElementById('price-filter-reset').style.display = 'none';
+  applyFilters();
+}}
+
 var allMarkers = [];
 
-function addMarker(lat, lon, color, popup, hashId) {{
+function addMarker(lat, lon, color, popup, hashId, firstSeen, lastChanged) {{
   var marker = L.circleMarker([lat, lon], {{
     radius: 8,
     fillColor: color,
@@ -704,26 +780,35 @@ function addMarker(lat, lon, color, popup, hashId) {{
     opacity: 1,
     fillOpacity: 0.85,
   }}).bindPopup(popup);
-  marker._data = {{ lat: lat, lon: lon, color: color, hashId: hashId }};
+  marker._data = {{ lat: lat, lon: lon, color: color, hashId: hashId, firstSeen: firstSeen || '', lastChanged: lastChanged || '' }};
   allMarkers.push(marker);
   marker.addTo(map);
 }}
 
-function addNewMarker(lat, lon, color, popup, hashId) {{
+function addNewMarker(lat, lon, color, popup, hashId, firstSeen, lastChanged) {{
   var marker = L.circleMarker([lat, lon], {{
-    radius: 12,
+    radius: 8,
     fillColor: color,
-    color: color,
+    color: '#fff',
-    weight: 4,
+    weight: 2,
-    opacity: 0.35,
+    opacity: 1,
-    fillOpacity: 0.95,
+    fillOpacity: 0.85,
   }}).bindPopup(popup);
-  marker._data = {{ lat: lat, lon: lon, color: color, hashId: hashId, isNew: true }};
+  marker._data = {{ lat: lat, lon: lon, color: color, hashId: hashId, isNew: true, firstSeen: firstSeen || '', lastChanged: lastChanged || '' }};
   allMarkers.push(marker);
   marker.addTo(map);
-  marker.on('add', function() {{
+  var badge = L.marker([lat, lon], {{
-    if (marker._path) marker._path.classList.add('marker-new');
+    icon: L.divIcon({{
+      className: 'new-badge-icon',
+      html: '<span class="new-badge">NEW</span>',
+      iconSize: [32, 14],
+      iconAnchor: [-6, 7],
+    }}),
+    interactive: false,
+    pane: 'markerPane',
   }});
+  badge.addTo(map);
+  marker._newBadge = badge;
 }}
 
 function heartIcon(color) {{
@@ -756,11 +841,11 @@ function starIcon() {{
   }});
 }}
 
-function addHeartMarker(lat, lon, color, popup, hashId) {{
+function addHeartMarker(lat, lon, color, popup, hashId, firstSeen, lastChanged) {{
   var marker = L.marker([lat, lon], {{
     icon: heartIcon(color),
   }}).bindPopup(popup);
-  marker._data = {{ lat: lat, lon: lon, color: color, hashId: hashId, isHeart: true }};
+  marker._data = {{ lat: lat, lon: lon, color: color, hashId: hashId, isHeart: true, firstSeen: firstSeen || '', lastChanged: lastChanged || '' }};
   allMarkers.push(marker);
   marker.addTo(map);
 }}
@@ -826,6 +911,7 @@ function applyMarkerStyle(marker, status) {{
   }} else {{
     if (status === 'fav') {{
       removeRejectStrike(marker);
+      if (marker._newBadge && map.hasLayer(marker._newBadge)) map.removeLayer(marker._newBadge);
       if (!marker._data._origCircle) marker._data._origCircle = true;
       var popup = marker.getPopup();
       var popupContent = popup ? popup.getContent() : '';
@@ -849,6 +935,7 @@ function applyMarkerStyle(marker, status) {{
       }}
       // Add strikethrough line over the marker
       addRejectStrike(marker);
+      if (marker._newBadge && map.hasLayer(marker._newBadge)) map.removeLayer(marker._newBadge);
     }} else {{
       if (marker._data._origCircle && !(marker instanceof L.CircleMarker)) {{
         revertToCircle(marker, {{ radius: 8, fillColor: marker._data.color, color: '#fff', weight: 2, fillOpacity: 0.85 }});
@@ -861,6 +948,7 @@ function applyMarkerStyle(marker, status) {{
       }}
       if (marker._path) marker._path.classList.remove('marker-rejected');
       removeRejectStrike(marker);
+      if (marker._newBadge && !map.hasLayer(marker._newBadge)) marker._newBadge.addTo(map);
     }}
   }}
 }}
@@ -1016,11 +1104,21 @@ map.on('popupopen', function(e) {{
 // ── Filters ────────────────────────────────────────────────────
 function applyFilters() {{
   var minFloor = parseInt(document.getElementById('min-floor').value);
-  var maxPrice = parseInt(document.getElementById('max-price').value);
+  var maxPriceEl = document.getElementById('max-price');
+  var maxPrice = parseInt(maxPriceEl.value) || 14000000;
+  if (maxPrice > 14000000) {{ maxPrice = 14000000; maxPriceEl.value = 14000000; }}
   var hideRejected = document.getElementById('hide-rejected').checked;
+  var daysFilter = parseInt(document.getElementById('days-filter').value) || 0;
   var ratings = loadRatings();
   var visible = 0;
 
+  var cutoff = null;
+  if (daysFilter > 0) {{
+    cutoff = new Date();
+    cutoff.setDate(cutoff.getDate() - daysFilter);
+    cutoff.setHours(0, 0, 0, 0);
+  }}
+
   allMarkers.forEach(function(m) {{
     var popup = m.getPopup().getContent();
     var floorMatch = popup.match(/(\\d+)\\. NP/);
@@ -1033,6 +1131,14 @@ function applyFilters() {{
     if (floor !== null && floor < minFloor) show = false;
     if (price > maxPrice) show = false;
 
+    if (cutoff) {{
+      var fs = m._data.firstSeen ? new Date(m._data.firstSeen) : null;
+      var lc = m._data.lastChanged ? new Date(m._data.lastChanged) : null;
+      if (!((fs && fs >= cutoff) || (lc && lc >= cutoff))) show = false;
+    }}
+
+    if (selectedColors.length > 0 && selectedColors.indexOf(m._data.color) < 0) show = false;
+
     var r = ratings[m._data.hashId];
     if (hideRejected && r && r.status === 'reject') show = false;
 
@@ -1041,10 +1147,12 @@ function applyFilters() {{
       visible++;
       // Show strike line if rejected and visible
       if (m._rejectStrike && !map.hasLayer(m._rejectStrike)) m._rejectStrike.addTo(map);
+      if (m._newBadge && !map.hasLayer(m._newBadge)) m._newBadge.addTo(map);
     }} else {{
       if (map.hasLayer(m)) map.removeLayer(m);
       // Hide strike line when marker hidden
       if (m._rejectStrike && map.hasLayer(m._rejectStrike)) map.removeLayer(m._rejectStrike);
+      if (m._newBadge && map.hasLayer(m._newBadge)) map.removeLayer(m._newBadge);
     }}
   }});
 

scrape_bezrealitky.py

@@ -362,6 +362,8 @@ def scrape(max_pages: int | None = None, max_properties: int | None = None):
             "source": "bezrealitky",
             "image": "",
             "scraped_at": datetime.now().strftime("%Y-%m-%d"),
+            "first_seen": cached.get("first_seen", datetime.now().strftime("%Y-%m-%d")) if cached else datetime.now().strftime("%Y-%m-%d"),
+            "last_changed": datetime.now().strftime("%Y-%m-%d"),
         }
         results.append(result)
         properties_fetched += 1

scrape_cityhome.py

@@ -255,6 +255,16 @@ def scrape(max_pages: int | None = None, max_properties: int | None = None):
         else:
             logger.info(f"✗ {slug}: GPS nenalezeno")
 
+    # Load previous output for first_seen/last_changed tracking
+    _prev_cache: dict[str, dict] = {}
+    _prev_path = Path("byty_cityhome.json")
+    if _prev_path.exists():
+        try:
+            for _item in json.loads(_prev_path.read_text(encoding="utf-8")):
+                _prev_cache[str(_item["hash_id"])] = _item
+        except Exception:
+            pass
+
     # Step 3: Filter listings
     logger.info(f"\nFáze 3: Filtrování...")
     results = []
@@ -362,6 +372,8 @@ def scrape(max_pages: int | None = None, max_properties: int | None = None):
             "source": "cityhome",
             "image": "",
             "scraped_at": datetime.now().strftime("%Y-%m-%d"),
+            "first_seen": _prev_cache.get(f"cityhome_{slug}_{listing['unit_name']}", {}).get("first_seen", datetime.now().strftime("%Y-%m-%d")),
+            "last_changed": datetime.now().strftime("%Y-%m-%d") if _prev_cache.get(f"cityhome_{slug}_{listing['unit_name']}", {}).get("price") != price else _prev_cache[f"cityhome_{slug}_{listing['unit_name']}"].get("last_changed", datetime.now().strftime("%Y-%m-%d")),
         }
         results.append(result)
         properties_fetched += 1

scrape_idnes.py

@@ -15,7 +15,6 @@ import re
 import time
 import urllib.request
 import urllib.parse
-from html.parser import HTMLParser
 from pathlib import Path
 from scraper_stats import write_stats
 
@@ -465,6 +464,8 @@ def scrape(max_pages: int | None = None, max_properties: int | None = None):
             "source": "idnes",
             "image": "",
             "scraped_at": datetime.now().strftime("%Y-%m-%d"),
+            "first_seen": cached.get("first_seen", datetime.now().strftime("%Y-%m-%d")) if cached else datetime.now().strftime("%Y-%m-%d"),
+            "last_changed": datetime.now().strftime("%Y-%m-%d"),
         }
         results.append(result)
         properties_fetched += 1

scrape_psn.py

@@ -112,6 +112,16 @@ def scrape(max_properties: int | None = None):
     all_units = data.get("units", {}).get("data", [])
     logger.info(f"Staženo jednotek celkem: {len(all_units)}")
 
+    # Load previous output for first_seen/last_changed tracking
+    _prev_cache: dict[str, dict] = {}
+    _prev_path = Path("byty_psn.json")
+    if _prev_path.exists():
+        try:
+            for _item in json.loads(_prev_path.read_text(encoding="utf-8")):
+                _prev_cache[str(_item["hash_id"])] = _item
+        except Exception:
+            pass
+
     # Filtrování
     results = []
     excluded = {
@@ -242,6 +252,8 @@ def scrape(max_properties: int | None = None):
             "source": "psn",
             "image": "",
             "scraped_at": datetime.now().strftime("%Y-%m-%d"),
+            "first_seen": _prev_cache.get(str(unit_id), {}).get("first_seen", datetime.now().strftime("%Y-%m-%d")),
+            "last_changed": datetime.now().strftime("%Y-%m-%d") if _prev_cache.get(str(unit_id), {}).get("price") != int(price) else _prev_cache[str(unit_id)].get("last_changed", datetime.now().strftime("%Y-%m-%d")),
         }
         results.append(result)
         properties_fetched += 1

scrape_realingo.py

@@ -321,6 +321,8 @@ def scrape(max_pages: int | None = None, max_properties: int | None = None):
             "source": "realingo",
             "image": "",
             "scraped_at": datetime.now().strftime("%Y-%m-%d"),
+            "first_seen": cached.get("first_seen", datetime.now().strftime("%Y-%m-%d")) if cached else datetime.now().strftime("%Y-%m-%d"),
+            "last_changed": datetime.now().strftime("%Y-%m-%d"),
         }
         results.append(result)
         properties_fetched += 1