Add savetv_country_filter: exclude DE/FR production countries from film selection

homelab-ai-bot/tools/savetv_country_filter.py

@@ -0,0 +1,79 @@
+"""Save.TV — Filter: Deutschland/Frankreich (Produktionsland via Wikidata-Cache).
+
+Nutzt dieselbe .filminfo_cache.json wie savetv_extra_routes (Wikidata countries).
+Ohne Cache-Eintrag fuer einen Titel wird nicht gefiltert (Film bleibt sichtbar).
+
+Deaktivieren: Umgebungsvariable SAVETV_FILTER_DE_FR=0
+"""
+from __future__ import annotations
+
+import json
+import os
+import re
+from pathlib import Path
+
+FILMINFO_CACHE = Path("/mnt/savetv/.filminfo_cache.json")
+
+# Deutschland / Frankreich (haeufige Wikidata-Sprachvarianten)
+_COUNTRY_EXCL = re.compile(
+    r"\b("
+    r"deutschland|germany|allemagne|bundesrepublik(\s+deutschland)?|"
+    r"west\s+germany|east\s+germany|german\s+democratic\s+republic|"
+    r"rfa|\bbrd\b|\bddr\b|\bgdr\b|"
+    r"frankreich|france|republique\s+francaise"
+    r")\b",
+    re.IGNORECASE,
+)
+
+
+def _enabled() -> bool:
+    v = (os.environ.get("SAVETV_FILTER_DE_FR") or "1").strip().lower()
+    return v not in ("0", "false", "no", "off")
+
+
+def _norm_title_key(s: str) -> str:
+    return re.sub(r"\s+", " ", re.sub(r"[^\w\s]", " ", s or "")).strip().lower()
+
+
+def load_filminfo_cache() -> dict:
+    if not FILMINFO_CACHE.exists():
+        return {}
+    try:
+        data = json.loads(FILMINFO_CACHE.read_text(encoding="utf-8"))
+        return data if isinstance(data, dict) else {}
+    except Exception:
+        return {}
+
+
+def _countries_for_title(title: str, cache: dict) -> list[str] | None:
+    if not title or not cache:
+        return None
+    entry = cache.get(title)
+    if entry is None:
+        nk = _norm_title_key(title)
+        for k, v in cache.items():
+            if isinstance(k, str) and _norm_title_key(k) == nk:
+                entry = v
+                break
+    if not entry or not isinstance(entry, dict):
+        return None
+    c = entry.get("countries")
+    if not c:
+        return None
+    if isinstance(c, list):
+        return [str(x) for x in c if x]
+    return [str(c)]
+
+
+def should_exclude_production_country(title: str, cache: dict | None = None) -> bool:
+    """True, wenn ein Produktionsland DE oder FR ist (laut Wikidata-Cache)."""
+    if not _enabled():
+        return False
+    cache = load_filminfo_filter() if cache is None else cache
+    countries = _countries_for_title(title, cache)
+    if not countries:
+        return False
+    for name in countries:
+        if name and _COUNTRY_EXCL.search(name):
+            return True
+    return False