Curated indoor kids venues source (פעלטון, ג'ימבורי, trampoline parks) with Shabbat hours

README.md

@@ -80,6 +80,9 @@ python3 sources/parks.py             # -> data/raw/parks.json
 # tiuli (coordinates already in the pages; resumable):
 python3 sources/tiuli.py             # -> data/raw/tiuli.json
 
+# kidsfun (curated indoor kids venues — geocodes sources/kidsfun_venues.json; resumable):
+python3 sources/kidsfun.py           # -> data/raw/kidsfun.json
+
 # enrichment (optional but recommended; both are resumable):
 python3 enrich/osm.py                # -> data/enrichment/osm_pois.json (one bulk Overpass query)
 python3 enrich/wikipedia.py          # -> data/enrichment/wikipedia.json (~1 req/sec, ~20 min)
@@ -124,6 +127,15 @@ up the new source filter automatically.
   present across every category, so no geocoding. Events (time-bound) and flora/fauna
   (species pages) are intentionally excluded. The on-page description is templated SEO
   filler, so it is left blank for Wikipedia enrichment to fill.
+- **kidsfun** is not a scraper but a hand-curated list (`sources/kidsfun_venues.json`)
+  of indoor play venues for small kids — פעלטון-style משחקיות, ג'ימבורי, soft-play —
+  a category the trip-blog sources don't cover at all. Every entry carries verified
+  opening hours (Saturday included: these venues are typically OPEN on Shabbat, unlike
+  most of the corpus) and an age range in the description. `sources/kidsfun.py`
+  geocodes each venue with Nominatim — mall name first, then street address — rejects
+  hits >12km from the stated city, and caches results in
+  `sources/cache/kidsfun_geocode.json`. To add a venue, append it to the JSON and
+  re-run the script + `build.py`.
 - **OSM enrichment** (`enrich/osm.py`) issues a single bulk Overpass query for named
   POIs across Israel, and `build.py` copies hours/phone/website/wheelchair/fee onto a
   place only on a confident name+distance match (a wrong match would show another

build.py

@@ -22,6 +22,7 @@
     "coffeetrail": "Coffee Trail",
     "parks": "רשות הטבע והגנים",
     "tiuli": "טיולי",
+    "kidsfun": "בילוי מקורה לילדים",
 }
 COFFEE_TYPE = "עגלות קפה ופוד טראק"
 

enrich/tag_places.py

@@ -28,6 +28,7 @@
 
 COFFEE_TYPE = "עגלות קפה ופוד טראק"
 MUSEUM_TYPE = "מוזיאון, אתר מורשת, מרכז מבקרים ועתיקות"
+KIDSFUN_TYPE = "משחקיות ובילוי מקורה"  # sources/kidsfun.py curated venues
 NATURE_TYPES = ["טיולי פריחה", "נקודות עניין בטבע", "שמורות טבע וגנים לאומיים",
                 "נקודות תצפית", "פארקים לפיקניק", "מסלולי טיול"]
 KIDS_TYPES = ["פארק שעשועים", "טיולים עם עגלות", "טיולים עם חיות"]
@@ -36,8 +37,9 @@
     r"מוזיאון|מרכז מדע|מדע וטכנולוגיה|מרכז מבקרים|מרכז המבקרים|משחקיי|משחקיה|"
     r"אסקייפ|escape|חדר בריחה|חדרי בריחה|טרמפולינ|ג'ימבו|ג׳ימבו|באולינג|"
     r"פלנטריום|פלנת|גלריה|לייזר|נינג'ה|נינג׳ה|אקווריום|מצפה כוכבים|"
-    r"חלל המופלא|אורבניה|בית הראשונים|יקב|קיר טיפוס|תיאטרון|סינמה|קולנוע")
-KIDS_RE = re.compile(r"טרמפולינ|ג'ימבו|ג׳ימבו|משחקיי|משחקיה|ילדים|משפח|שעשוע")
+    r"חלל המופלא|אורבניה|בית הראשונים|יקב|קיר טיפוס|תיאטרון|סינמה|קולנוע|"
+    r"פעלטון|לונדע|סופט פליי|ג'ימבורי|ג׳ימבורי")
+KIDS_RE = re.compile(r"טרמפולינ|ג'ימבו|ג׳ימבו|משחקיי|משחקיה|ילדים|משפח|שעשוע|פעלטון|לונדע")
 BEACH_RE = re.compile(r"חוף|טיילת")
 
 
@@ -48,10 +50,10 @@ def rule_tags(a):
     hay = (a.get("title", "") + " " + " ".join(a.get("keywords", [])) + " " + " ".join(types))
     tags = set()
     is_coffee = COFFEE_TYPE in types
-    is_indoor = MUSEUM_TYPE in types or bool(INDOOR_RE.search(hay))
+    is_indoor = MUSEUM_TYPE in types or KIDSFUN_TYPE in types or bool(INDOOR_RE.search(hay))
     is_beach = "טיולי מים" in types or bool(BEACH_RE.search(hay))
     is_nature = any(t in types for t in NATURE_TYPES)
-    is_kids = any(t in types for t in KIDS_TYPES) or bool(KIDS_RE.search(hay))
+    is_kids = any(t in types for t in KIDS_TYPES) or KIDSFUN_TYPE in types or bool(KIDS_RE.search(hay))
 
     if is_coffee:
         tags |= {"outdoor", "food", "quick-stop"}

sources/kidsfun.py

@@ -0,0 +1,163 @@
+#!/usr/bin/env python3
+"""Curated source: indoor play venues for small kids (פעלטון-style משחקיות,
+ג'ימבורי, soft-play), hand-maintained in sources/kidsfun_venues.json.
+
+Unlike the scraped sources, this one is a verified editorial list: every venue
+carries opening hours (Saturday matters most — these are the places that ARE
+open on Shabbat when the heritage/nature corpus is closed) and an age range.
+Coordinates are geocoded via Nominatim (mall name first, then street address),
+cached in sources/cache/kidsfun_geocode.json, and sanity-checked against the
+venue's city centroid so a bad geocode can't land a branch in the wrong town.
+
+Emits the shared raw schema to data/raw/kidsfun.json with hours/phone/website
+passed through (build.py's merge copies those fields from raw members).
+"""
+import json
+import os
+import re
+import time
+import urllib.parse
+import urllib.request
+
+HERE = os.path.dirname(os.path.abspath(__file__))
+ROOT = os.path.dirname(HERE)
+VENUES = os.path.join(HERE, "kidsfun_venues.json")
+CACHE = os.path.join(HERE, "cache", "kidsfun_geocode.json")
+OUT = os.path.join(ROOT, "data", "raw", "kidsfun.json")
+
+UA = "attractions-map-builder/1.0 (amara.nir@gmail.com)"
+LAT_MIN, LAT_MAX = 29.45, 33.40
+LON_MIN, LON_MAX = 34.20, 35.95
+KIDSFUN_TYPE = "משחקיות ובילוי מקורה"
+
+_last_req = [0.0]
+
+
+def _nominatim(params):
+    wait = 1.1 - (time.time() - _last_req[0])
+    if wait > 0:
+        time.sleep(wait)
+    url = "https://nominatim.openstreetmap.org/search?" + urllib.parse.urlencode(params)
+    req = urllib.request.Request(url, headers={"User-Agent": UA})
+    with urllib.request.urlopen(req, timeout=30) as r:
+        data = json.load(r)
+    _last_req[0] = time.time()
+    return data
+
+
+def geocode(query):
+    # Two attempts: countrycodes=il first, then a bounded viewbox without it —
+    # West Bank places (e.g. Maale Adumim) aren't tagged 'il' in Nominatim.
+    for params in (
+        {"q": query, "countrycodes": "il", "format": "json", "limit": "1",
+         "viewbox": f"{LON_MIN},{LAT_MAX},{LON_MAX},{LAT_MIN}"},
+        {"q": query, "format": "json", "limit": "1", "bounded": "1",
+         "viewbox": f"{LON_MIN},{LAT_MAX},{LON_MAX},{LAT_MIN}"},
+    ):
+        data = _nominatim(params)
+        if not data:
+            continue
+        lat, lon = float(data[0]["lat"]), float(data[0]["lon"])
+        if LAT_MIN <= lat <= LAT_MAX and LON_MIN <= lon <= LON_MAX:
+            return {"lat": lat, "lng": lon, "display": data[0].get("display_name", "")}
+    return None
+
+
+def geocode_city(city):
+    data = _nominatim({
+        "city": city, "format": "json", "limit": "1",
+        "viewbox": f"{LON_MIN},{LAT_MAX},{LON_MAX},{LAT_MIN}", "bounded": "1",
+    })
+    if not data:
+        return None
+    return {"lat": float(data[0]["lat"]), "lng": float(data[0]["lon"])}
+
+
+def dist_km(a, b):
+    import math
+    dlat = (a["lat"] - b["lat"]) * 111.0
+    dlng = (a["lng"] - b["lng"]) * 111.0 * math.cos(math.radians((a["lat"] + b["lat"]) / 2))
+    return math.hypot(dlat, dlng)
+
+
+def load_cache():
+    if os.path.exists(CACHE):
+        return json.load(open(CACHE, encoding="utf-8"))
+    return {}
+
+
+def main():
+    venues = json.load(open(VENUES, encoding="utf-8"))
+    cache = load_cache()
+
+    def cached(key, fn, *args):
+        if key not in cache:
+            try:
+                cache[key] = fn(*args)
+            except Exception:
+                cache[key] = None
+        return cache[key]
+
+    out, misses = [], []
+    for v in venues:
+        city = v["city"]
+        # Mall name resolves to the exact building; the street address and the
+        # venue name are progressively weaker fallbacks.
+        queries = []
+        if v.get("geocode_hint"):
+            queries.append(v["geocode_hint"])
+        if v.get("mall"):
+            queries.append(f"{v['mall']}, {city}")
+        if v.get("address"):
+            queries.append(f"{v['address']}, {city}")
+        queries.append(f"{v['title']}, {city}")
+
+        geo = None
+        for q in queries:
+            geo = cached(q, geocode, q)
+            if not geo:
+                continue
+            centroid = cached("city:" + city, geocode_city, city)
+            # Reject hits far from the stated city (same-named mall/street in
+            # another town); 12km covers metro sprawl like Beer Sheva.
+            if centroid and dist_km(geo, centroid) > 12:
+                geo = None
+                continue
+            break
+        if not geo:
+            misses.append(v["title"])
+            continue
+
+        hours = v.get("hours")
+        keywords = ["משחקייה", "ילדים", "בילוי מקורה"] + v.get("keywords", [])
+        # "Sa <time>" (not "Sa off") ⇒ open on Shabbat — the killer search term.
+        if hours and re.search(r"Sa[^;]*\d", hours) and not re.search(r"Sa (off|closed)", hours):
+            keywords.append("פתוח בשבת")
+
+        out.append({
+            "source": "kidsfun",
+            "title": v["title"],
+            "lat": round(geo["lat"], 6),
+            "lng": round(geo["lng"], 6),
+            "region": None,  # build.py derives from coords
+            "types": [KIDSFUN_TYPE],
+            "keywords": keywords,
+            "image": v.get("image"),
+            "url": v["url"],
+            "address": f"{v['address']}, {city}" if v.get("address") else city,
+            "description": v.get("description"),
+            "hours": v.get("hours"),
+            "phone": v.get("phone"),
+            "website": v.get("website") or v["url"],
+        })
+
+    json.dump(cache, open(CACHE, "w", encoding="utf-8"), ensure_ascii=False)
+    os.makedirs(os.path.dirname(OUT), exist_ok=True)
+    json.dump(out, open(OUT, "w", encoding="utf-8"), ensure_ascii=False)
+    print(f"kidsfun: {len(out)} venues placed -> {OUT}")
+    if misses:
+        print(f"  {len(misses)} could not be geocoded: " + "; ".join(misses))
+
+
+if __name__ == "__main__":
+    main()