Upload folder using huggingface_hub

matchday/agent.py

@@ -78,9 +78,10 @@ TOOL_SPECS: list[dict[str, Any]] = [
         "function": {
             "name": "web_search",
             "description": (
-                "Search the web for supplemental FIFA/Vancouver context (kick-off "
-                "times, venue policies). NOT for prices, flights, hotels, or "
-                "weather — those come from build_trip_packages."
+                "Search the web for supplemental FIFA/Vancouver context and "
+                "current local facts (kick-off times, venue policies, transit, "
+                "current weather right now). NOT for prices, flights, hotels, or "
+                "the match-day weather FORECAST — those come from build_trip_packages."
             ),
             "parameters": WebSearchArgs.model_json_schema(),
         },

matchday/prompts.py

@@ -77,6 +77,18 @@ Tool-use rules:
 - UNDERSTAND FIRST, THEN ACT. Read what the traveler actually wants before
   calling a tool. Capture their intent: the origin city/airport, the match,
   the dates, the budget, the vibe.
+- CLASSIFY THE INTENT, then act once. Pick exactly one path:
+  (a) TRIP — origin + match + date are present or inferable -> build_trip_packages.
+  (b) INCOMPLETE TRIP — a trip is intended but origin OR date is genuinely missing
+      -> clarify with ONE specific question offering concrete options.
+  (c) FACTUAL / LOCAL question — a question ABOUT the World Cup or Vancouver
+      (today's date, the BC Place schedule, kickoff time, how to get to the
+      stadium, the CURRENT weather right now, local info) with NO intent to book
+      -> answer it directly (see the "FACTUAL / LOCAL questions" rule below); do
+      not call build_trip_packages or ask for trip details. "What's the weather
+      today in Vancouver?" and "what's today's date?" are (c), not (a).
+  (d) OUT OF SCOPE — another destination or a non-World-Cup trip -> decline and
+      offer a Vancouver World Cup trip.
 - If the request is clear (you can infer an origin, a match, and at least one
   date), call `build_trip_packages` right away — it runs all searches in
   parallel and returns 3 scored packages. A match written as "playoff winner",
@@ -90,12 +102,21 @@ Tool-use rules:
   the user asks about another city or a non-World-Cup trip, do NOT plan it —
   say you specialize in Vancouver 2026 World Cup trips and offer to plan one
   of those instead.
-- For a factual question (match schedule at BC Place, kickoff time, transit to
-  the stadium) that `build_trip_packages` does not answer, use `web_search`
-  (restricted to fifa.com, vancouverfwc26.ca, bcplace.com, translink.ca) and
-  ANSWER directly from the result. Do not ask for trip details a factual
-  question does not need. Do NOT use web_search for prices, flights, hotels,
-  or weather — those come from `build_trip_packages`.
+- FACTUAL / LOCAL questions (intent (c) above): answer directly and briefly
+  (1-3 lines). If you already have the answer — today's date is in your
+  instructions; an evening kickoff defaults to 7:00 PM Pacific — answer with NO
+  tool call. If you need data (the BC Place schedule, a kickoff time, transit,
+  the CURRENT weather right now), call `web_search` (preferring authoritative
+  sources — fifa.com / vancouverfwc26.ca / bcplace.com / translink.ca for
+  schedule, venue and transit; weather.gc.ca / Environment Canada for current
+  weather) and answer from the result. Do
+  not ask for trip details a factual question does not need, and do not force
+  build_trip_packages for a question that isn't a trip request.
+- TRIP weather, prices, flights and hotels (the match-day FORECAST, airfare, a
+  hotel rate) come ONLY from `build_trip_packages` — never from web_search. So
+  "what's the weather TODAY in Vancouver?" (current, now) -> web_search;
+  "what's the weather for the Canada vs Qatar game?" (a trip forecast) ->
+  build_trip_packages. web_search is never used for trip pricing or trip weather.
 - Do not call a tool you already called with identical arguments this turn.
 - When you have the packages, write a concise comparison: name each package's
   strength (Cheapest / Safest Arrival / Closest to Stadium), cite the real

matchday/wc2026.py

@@ -60,10 +60,21 @@ _FIXTURES: list[Fixture] = [
     Fixture("Bosnia and Herzegovina vs Qatar", date(2026, 6, 24), "Lumen Field", "Seattle", "", "B"),
     # Group C — Brazil's group (Brazil, Morocco, Scotland, Haiti)
     Fixture("Brazil vs Morocco", date(2026, 6, 13), "MetLife Stadium", "New York / New Jersey", "18:00 ET", "C"),
-    # Other BC Place (Vancouver) group-stage games
+    # Group D opener at BC Place (Vancouver) — verified Jun 13.
     Fixture("Australia vs Türkiye", date(2026, 6, 13), "BC Place", "Vancouver", "21:00 PT", "D"),
-    Fixture("New Zealand vs Egypt", date(2026, 6, 22), "BC Place", "Vancouver", "", "G"),
-    Fixture("Argentina vs Austria", date(2026, 6, 22), "BC Place", "Vancouver", "", "G"),
+    # Group G (Belgium / Egypt / IR Iran / New Zealand) — NZ's two BC Place games.
+    # NZ vs Egypt is SUNDAY Jun 21 (FIFA match M40), NOT Jun 22 — verified against
+    # the official FIFA Hospitality Vancouver schedule + Wikipedia Group G
+    # (re-checked 2026-06-15). The prior "Jun 22" entry was off by one day, which
+    # is exactly the "wrong match data" a user saw. NZ vs Belgium is M64, Fri Jun 26.
+    Fixture("New Zealand vs Egypt", date(2026, 6, 21), "BC Place", "Vancouver", "", "G"),
+    Fixture("New Zealand vs Belgium", date(2026, 6, 26), "BC Place", "Vancouver", "", "G"),
+    # Group J (Argentina's group). Argentina vs Austria is a REAL 2026 fixture but
+    # at AT&T Stadium (Dallas/Arlington), NOT Vancouver — corrected from a stale
+    # "BC Place / Group G" entry (neither team is in Group G). Kept (truthful) so a
+    # user who names it gets an honest "that's in Dallas, not Vancouver" redirect
+    # instead of a false "isn't a 2026 fixture".
+    Fixture("Argentina vs Austria", date(2026, 6, 22), "AT&T Stadium", "Dallas", "13:00 ET", "J"),
 ]
 
 # Spoken team-name aliases → canonical token used for matching. Lets "Bosnia"
@@ -87,13 +98,31 @@ def _norm_team(raw: str) -> str:
     return _TEAM_ALIASES.get(t, t)
 
 
+def _trim_team_suffix(side: str) -> str:
+    """Cut a team string at the first standalone dash / 'at' that introduces a
+    trailing venue — e.g. 'New Zealand – BC Place' -> 'New Zealand'. Lets a user
+    name a match *with* its stadium ('Egypt vs New Zealand – BC Place') still
+    resolve instead of being falsely reported as 'not a fixture' (the exact
+    'I added the stadium and it refused' bug)."""
+    import re
+    side = re.split(r"\s+[-–—]\s+", side, maxsplit=1)[0]                       # ' – BC Place'
+    side = re.split(r"\s+(?:at|@)\s+", side, maxsplit=1, flags=re.IGNORECASE)[0]  # ' at BC Place'
+    return side.strip()
+
+
 def _split_match(match_name: str) -> list[str]:
-    """Split 'A vs B' into normalized team tokens (any order)."""
+    """Split 'A vs B' into normalized team tokens (any order).
+
+    Each side is trimmed of a trailing venue suffix before normalization, so
+    'Egypt vs New Zealand – BC Place' yields ['egypt','newzealand'] rather than a
+    contaminated token ('new zealand – bc place') that fails to match the fixture
+    table. Separators accepted: 'vs' / 'v' / 'versus', or a space-padded dash.
+    """
     import re
-    parts = re.split(r"\b(?:vs?\.?|versus|[-–])\b", match_name, maxsplit=1, flags=re.IGNORECASE)
+    parts = re.split(r"\b(?:vs?\.?|versus)\b|\s+[-–—]\s+", match_name, maxsplit=1, flags=re.IGNORECASE)
     if len(parts) != 2:
         return []
-    return [_norm_team(parts[0]), _norm_team(parts[1])]
+    return [_norm_team(_trim_team_suffix(parts[0])), _norm_team(_trim_team_suffix(parts[1]))]
 
 
 def _teams_match(query_teams: list[str], fixture: Fixture) -> bool:
@@ -107,6 +136,12 @@ def _teams_match(query_teams: list[str], fixture: Fixture) -> bool:
         for f in ft:
             if q == f or (len(q) >= 3 and (q in f or f in q)):
                 return f
+            # space-insensitive containment: a venue suffix absorbed into the
+            # team (e.g. the matcher seeing 'New Zealand BC' from '…BC Place')
+            # still matches the canonical 'newzealand'. len(f)>=4 avoids
+            # short-token false positives.
+            if len(f) >= 4 and f in q.replace(" ", ""):
+                return f
         return None
     a, b = hit(query_teams[0]), hit(query_teams[1])
     return bool(a and b and a != b)