Parity: like-for-like tariff matching instead of headline vs lead-in

The check compared Newbook's FIRST tariff (typically flexi B&B) against
Booking.com's lead-in card (often prepaid) — apples vs oranges, e.g.
21 Jul flagged B&B FLEX £289 vs a prepaid BC rate whose true comparable
was B&B PPAY £279.

- classify Newbook tariffs from names/descriptions (PPAY/prepay/advance/
  ADV/NRF/saver = prepaid; DBB/dinner/half board = dinner-inclusive)
- use the scraped BC rate's flags (free_cancellation/no_prepayment/
  breakfast_included) to pick the cheapest COMPARABLE tariff per date,
  excluding tariffs not bookable for that date (success=false, min-stay
  >1, unmet advance-purchase windows)
- tiered fallback (matched -> any non-dinner -> any -> legacy headline),
  recorded per alert in room_category as 'TARIFF vs BC basis'
- shared gather_comparisons() now drives both the daily job and
  GET /competitors/parity (issues gain newbook_tariff, booking_basis,
  match_quality, expected_rate)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

backend/api/competitors.py

@@ -717,93 +717,45 @@ async def get_rate_parity(
     current_user: dict = Depends(get_current_user)
 ):
     """
-    Get rate parity comparison between booking.com and Newbook rates.
+    Like-for-like parity comparison between our Booking.com rate and the
-
+    comparable Newbook tariff (matched on flex/prepaid + board via the
-    Compares scraped booking.com rates for own hotel against Newbook current rates.
+    scraped rate's condition flags). Same logic as the daily alert job.
     """
+    import asyncio
+    from jobs.check_rate_parity import get_parity_config, gather_comparisons, expected_booking_rate
+    from database import SyncSessionLocal
+
     today = date.today()
     start = date.fromisoformat(from_date) if from_date else today
     end = date.fromisoformat(to_date) if to_date else today + timedelta(days=30)
 
-    # Get own hotel's booking.com rates
+    def _run():
-    booking_rates_result = await db.execute(
+        sdb = SyncSessionLocal()
-        text("""
-            SELECT DISTINCT ON (r.rate_date)
-                r.rate_date,
-                r.rate_gross as booking_rate,
-                r.availability_status,
-                r.room_type as booking_room_type,
-                r.scraped_at
-            FROM booking_com_rates r
-            JOIN booking_com_hotels h ON r.hotel_id = h.id
-            WHERE h.tier = 'own'
-              AND r.rate_date >= :from_date AND r.rate_date <= :to_date
-            ORDER BY r.rate_date, r.scraped_at DESC
-        """),
-        {'from_date': start, 'to_date': end}
-    )
-    booking_rates = {row.rate_date: dict(row._mapping) for row in booking_rates_result.fetchall()}
-
-    # Get Newbook rates (cheapest per date across latest category snapshots)
-    newbook_rates_result = await db.execute(
-        text("""
-            SELECT rate_date, MIN(rate_gross) AS newbook_rate FROM (
-                SELECT DISTINCT ON (rate_date, category_id)
-                    rate_date, rate_gross
-                FROM newbook_current_rates
-                WHERE rate_date >= :from_date AND rate_date <= :to_date
-                  AND rate_gross IS NOT NULL AND rate_gross > 0
-                ORDER BY rate_date, category_id, valid_from DESC
-            ) t GROUP BY rate_date
-        """),
-        {'from_date': start, 'to_date': end}
-    )
-    newbook_rates = {row.rate_date: dict(row._mapping) for row in newbook_rates_result.fetchall()}
-
-    # Expected-markup config: we deliberately price Booking.com higher to
-    # cover commission, so parity is measured against Newbook + markup (% or £)
-    from jobs.check_rate_parity import get_parity_config, expected_booking_rate, evaluate_parity
-    from database import SyncSessionLocal
-    cfg_db = SyncSessionLocal()
         try:
-        cfg = get_parity_config(cfg_db)
+            cfg = get_parity_config(sdb)
+            return cfg, gather_comparisons(sdb, start, end, cfg)
         finally:
-        cfg_db.close()
+            sdb.close()
 
-    # Compare rates
+    loop = asyncio.get_event_loop()
-    parity_issues = []
+    cfg, comparisons = await loop.run_in_executor(None, _run)
-    all_dates = set(booking_rates.keys()) | set(newbook_rates.keys())
 
-    for rate_date in sorted(all_dates):
+    parity_issues = [
-        booking = booking_rates.get(rate_date)
+        {
-        newbook = newbook_rates.get(rate_date)
+            'rate_date': c["date"].isoformat(),
-
+            'booking_rate': c["booking_rate"],
-        if not booking or not newbook:
+            'booking_basis': c["booking_basis"],
-            continue
+            'booking_room_type': c["booking_room"],
-
+            'newbook_rate': c["newbook_rate"],
-        booking_rate = booking.get('booking_rate')
+            'newbook_tariff': c["newbook_tariff"],
-        newbook_rate = newbook.get('newbook_rate')
+            'match_quality': c["match_quality"],
-
+            'expected_rate': round(expected_booking_rate(c["newbook_rate"], cfg), 2),
-        if not booking_rate or not newbook_rate:
+            'difference_pct': round(c["dev_pct"], 2),
-            continue
+            'difference_gbp': round(c["dev_gbp"], 2),
-
+            'alert_type': 'higher' if c["dev_pct"] > 0 else 'lower',
-        result = evaluate_parity(float(booking_rate), float(newbook_rate), cfg)
+        }
-        if result is None:
+        for c in comparisons if c["breach"]
-            continue
+    ]
-        diff_pct, diff_gbp, breach = result
-
-        if breach:
-            parity_issues.append({
-                'rate_date': rate_date.isoformat(),
-                'booking_rate': float(booking_rate),
-                'newbook_rate': float(newbook_rate),
-                'expected_rate': round(expected_booking_rate(float(newbook_rate), cfg), 2),
-                'difference_pct': round(diff_pct, 2),
-                'difference_gbp': round(diff_gbp, 2),
-                'alert_type': 'higher' if diff_pct > 0 else 'lower',
-                'booking_room_type': booking.get('booking_room_type'),
-                'availability_status': booking.get('availability_status'),
-            })
 
     return {
         'from_date': start.isoformat(),
@@ -812,6 +764,7 @@ async def get_rate_parity(
         'markup_unit': cfg["markup_unit"],
         'tolerance_value': cfg["tolerance_value"],
         'tolerance_unit': cfg["tolerance_unit"],
+        'dates_compared': len(comparisons),
         'issues_count': len(parity_issues),
         'issues': parity_issues
     }

backend/jobs/check_rate_parity.py

@@ -1,15 +1,26 @@
 """
 Rate Parity Check Job
 
-Compares our own hotel's scraped Booking.com rate against our Newbook rate
+Compares our own hotel's scraped Booking.com rate against the COMPARABLE
-for each date, allowing for a configured expected markup — we deliberately
+Newbook tariff for each date — like-for-like, not headline vs headline.
-price Booking.com higher to cover commission, so parity is measured against
+
-expected = newbook × (1 + markup%), not raw equality. Deviations beyond the
+The scraped Booking.com lead-in rate carries condition flags (breakfast
-tolerance are persisted as rate_parity_alerts rows (feeding the Market View
+included, free cancellation, no prepayment); Newbook's tariffs_data holds
-badge); dates back within tolerance auto-resolve their active alert.
+every tariff by name with per-day rates (e.g. DIRECT B&B FLEX / PPAY / ADV,
-Acknowledged alerts are left alone — acknowledging a date suppresses
+DIRECT DBB ...). We classify tariffs from their names and pick the cheapest
-re-alerting for it until the alert is resolved by the rates coming back
+one matching the Booking.com rate's basis: a prepaid B&B rate on Booking.com
-in line.
+is compared against the cheapest prepaid B&B tariff, not the flexi rate.
+
+Fallback tiers when no exact match exists (each alert records which was used):
+  matched    — terms (flex/prepaid) and board match
+  terms      — terms match, any board except dinner-inclusive
+  any        — any non-dinner tariff
+  headline   — legacy rate_gross (first tariff) when tariffs_data is absent
+
+Parity is then measured against an expected markup — we deliberately price
+Booking.com higher to cover commission: expected = comparable + markup
+(% or flat £), alert when the deviation exceeds the tolerance (% or £).
+Dates back within tolerance auto-resolve; acknowledged dates stay quiet.
 
 Config (system_config):
   parity_check_enabled     true/false (default true)
@@ -21,7 +32,9 @@ Config (system_config):
 
 Schedule: daily at 06:45, after the 05:20 Newbook fetch and 05:30 scrape.
 """
+import json
 import logging
+import re
 from datetime import date, timedelta
 
 from sqlalchemy import text
@@ -32,6 +45,9 @@ logger = logging.getLogger(__name__)
 
 HORIZON_DAYS = 90
 
+PREPAID_PATTERNS = ('ppay', 'prepay', 'pre-pay', 'advance', 'adv', 'non-ref', 'nonref', 'no refund', 'nrf', 'saver')
+DINNER_PATTERNS = ('dbb', 'dinner', 'half board', 'half-board')
+
 
 def get_parity_config(db) -> dict:
     rows = db.execute(
@@ -88,6 +104,136 @@ def evaluate_parity(booking_rate: float, newbook_rate: float, cfg: dict):
     return dev_pct, dev_gbp, breach
 
 
+def _classify_tariff(name: str, description: str = '') -> dict:
+    s = f"{name} {description}".lower()
+    is_dinner = any(p in s for p in DINNER_PATTERNS)
+    is_prepaid = any(re.search(rf'\b{re.escape(p)}\b', s) if p == 'adv' else (p in s)
+                     for p in PREPAID_PATTERNS)
+    return {"prepaid": is_prepaid, "dinner": is_dinner}
+
+
+def _candidate_tariffs(tariffs_data, days_ahead: int) -> list:
+    """Bookable-tonight tariff options from one category's tariffs_data."""
+    if isinstance(tariffs_data, str):
+        try:
+            tariffs_data = json.loads(tariffs_data)
+        except (ValueError, TypeError):
+            return []
+    out = []
+    for t in (tariffs_data or {}).get('tariffs', []):
+        rate = t.get('rate')
+        if not t.get('success') or not rate or rate <= 0:
+            continue
+        if (t.get('min_stay') or 1) > 1:
+            continue
+        if t.get('min_advance_days') and days_ahead < t['min_advance_days']:
+            continue
+        cls = _classify_tariff(t.get('name', ''), t.get('description', ''))
+        out.append({"rate": float(rate), "name": t.get('name', 'Unknown'), **cls})
+    return out
+
+
+def _pick_comparable(candidates: list, booking_flex: bool, booking_breakfast: bool):
+    """Cheapest tariff matching the Booking.com rate's basis, tiered fallback.
+    All our tariffs include breakfast, so board matching reduces to excluding
+    dinner-inclusive tariffs unless nothing else exists.
+    Returns (tariff, match_quality) or (None, None)."""
+    if not candidates:
+        return None, None
+    want_prepaid = not booking_flex
+    tiers = [
+        ([c for c in candidates if c["prepaid"] == want_prepaid and not c["dinner"]], 'matched'),
+        ([c for c in candidates if not c["dinner"]], 'any'),
+        (candidates, 'any'),
+    ]
+    for pool, quality in tiers:
+        if pool:
+            return min(pool, key=lambda c: c["rate"]), quality
+    return None, None
+
+
+def gather_comparisons(db, start: date, end: date, cfg: dict) -> list:
+    """Per-date like-for-like comparison rows for the horizon.
+    Each row: {date, booking_rate, booking_room, booking_basis, newbook_rate,
+               newbook_tariff, match_quality, dev_pct, dev_gbp, breach}"""
+    booking_rows = db.execute(text("""
+        SELECT DISTINCT ON (r.rate_date)
+            r.rate_date, r.rate_gross, r.room_type,
+            r.breakfast_included, r.free_cancellation, r.no_prepayment
+        FROM booking_com_rates r
+        JOIN booking_com_hotels h ON r.hotel_id = h.id
+        WHERE h.tier = 'own'
+          AND r.rate_date BETWEEN :fd AND :td
+          AND r.rate_gross IS NOT NULL
+        ORDER BY r.rate_date, r.scraped_at DESC
+    """), {"fd": start, "td": end}).mappings().all()
+
+    newbook_rows = db.execute(text("""
+        SELECT DISTINCT ON (rate_date, category_id)
+            rate_date, category_id, rate_gross, tariffs_data
+        FROM newbook_current_rates
+        WHERE rate_date BETWEEN :fd AND :td
+        ORDER BY rate_date, category_id, valid_from DESC
+    """), {"fd": start, "td": end}).mappings().all()
+
+    newbook_by_date: dict = {}
+    for r in newbook_rows:
+        newbook_by_date.setdefault(r["rate_date"], []).append(r)
+
+    today = date.today()
+    out = []
+    for b in booking_rows:
+        d = b["rate_date"]
+        cat_rows = newbook_by_date.get(d)
+        if not cat_rows:
+            continue
+
+        booking_rate = float(b["rate_gross"])
+        booking_flex = bool(b["free_cancellation"] or b["no_prepayment"])
+        booking_breakfast = bool(b["breakfast_included"])
+        days_ahead = (d - today).days
+
+        candidates = []
+        for cat in cat_rows:
+            candidates.extend(_candidate_tariffs(cat["tariffs_data"], days_ahead))
+
+        chosen, quality = _pick_comparable(candidates, booking_flex, booking_breakfast)
+        if chosen:
+            newbook_rate = chosen["rate"]
+            tariff_name = chosen["name"]
+        else:
+            # No tariff detail — legacy fallback to cheapest headline rate
+            headline = [float(c["rate_gross"]) for c in cat_rows if c["rate_gross"] and float(c["rate_gross"]) > 0]
+            if not headline:
+                continue
+            newbook_rate = min(headline)
+            tariff_name = 'headline rate'
+            quality = 'headline'
+
+        result = evaluate_parity(booking_rate, newbook_rate, cfg)
+        if result is None:
+            continue
+        dev_pct, dev_gbp, breach = result
+
+        basis_bits = [
+            'flex' if booking_flex else 'prepaid',
+            'B&B' if booking_breakfast else 'room only',
+        ]
+        out.append({
+            "date": d,
+            "booking_rate": booking_rate,
+            "booking_room": (b["room_type"] or '').split('\n')[0].strip() or None,
+            "booking_basis": '/'.join(basis_bits),
+            "newbook_rate": newbook_rate,
+            "newbook_tariff": tariff_name,
+            "match_quality": quality,
+            "dev_pct": dev_pct,
+            "dev_gbp": dev_gbp,
+            "breach": breach,
+        })
+    return out
+
+
 def run_parity_check() -> dict:
     db = SyncSessionLocal()
     try:
@@ -99,31 +245,7 @@ def run_parity_check() -> dict:
         start = date.today()
         end = start + timedelta(days=HORIZON_DAYS)
 
-        # Own hotel's latest Booking.com lead-in rate per date
+        comparisons = gather_comparisons(db, start, end, cfg)
-        booking_rows = db.execute(text("""
-            SELECT DISTINCT ON (r.rate_date)
-                r.rate_date, r.rate_gross, r.room_type
-            FROM booking_com_rates r
-            JOIN booking_com_hotels h ON r.hotel_id = h.id
-            WHERE h.tier = 'own'
-              AND r.rate_date BETWEEN :fd AND :td
-              AND r.rate_gross IS NOT NULL
-            ORDER BY r.rate_date, r.scraped_at DESC
-        """), {"fd": start, "td": end}).fetchall()
-        booking = {r[0]: {"rate": float(r[1]), "room": r[2]} for r in booking_rows}
-
-        # Cheapest current Newbook rate per date (latest snapshot per category)
-        newbook_rows = db.execute(text("""
-            SELECT rate_date, MIN(rate_gross) AS newbook_rate FROM (
-                SELECT DISTINCT ON (rate_date, category_id)
-                    rate_date, rate_gross
-                FROM newbook_current_rates
-                WHERE rate_date BETWEEN :fd AND :td
-                  AND rate_gross IS NOT NULL AND rate_gross > 0
-                ORDER BY rate_date, category_id, valid_from DESC
-            ) t GROUP BY rate_date
-        """), {"fd": start, "td": end}).fetchall()
-        newbook = {r[0]: float(r[1]) for r in newbook_rows}
 
         # Latest alert per date in the horizon
         alert_rows = db.execute(text("""
@@ -135,19 +257,15 @@ def run_parity_check() -> dict:
         """), {"fd": start, "td": end}).fetchall()
         latest_alert = {r[1]: {"id": r[0], "status": r[2], "diff": float(r[3] or 0)} for r in alert_rows}
 
-        common_dates = sorted(set(booking) & set(newbook))
         created = updated = resolved = 0
+        for c in comparisons:
+            alert = latest_alert.get(c["date"])
+            # e.g. "DIRECT B&B PPAY vs BC prepaid/B&B"
+            category_label = f"{c['newbook_tariff']} vs BC {c['booking_basis']}"[:255]
 
-        for d in common_dates:
+            if c["breach"]:
-            result = evaluate_parity(booking[d]["rate"], newbook[d], cfg)
-            if result is None:
-                continue
-            dev, _dev_gbp, breach = result
-            alert = latest_alert.get(d)
-
-            if breach:
                 if alert and alert["status"] == "active":
-                    if round(dev, 2) != round(alert["diff"], 2):
+                    if round(c["dev_pct"], 2) != round(alert["diff"], 2):
                         db.execute(text("""
                             UPDATE rate_parity_alerts
                             SET newbook_rate = :nb, booking_com_rate = :bc,
@@ -155,10 +273,10 @@ def run_parity_check() -> dict:
                                 room_category = :room
                             WHERE id = :id
                         """), {
-                            "id": alert["id"], "nb": newbook[d], "bc": booking[d]["rate"],
+                            "id": alert["id"], "nb": c["newbook_rate"], "bc": c["booking_rate"],
-                            "diff": round(dev, 2),
+                            "diff": round(c["dev_pct"], 2),
-                            "atype": "higher" if dev > 0 else "lower",
+                            "atype": "higher" if c["dev_pct"] > 0 else "lower",
-                            "room": booking[d]["room"],
+                            "room": category_label,
                         })
                         updated += 1
                 elif alert and alert["status"] == "acknowledged":
@@ -170,10 +288,10 @@ def run_parity_check() -> dict:
                              difference_pct, alert_type, alert_status)
                         VALUES (:date, :room, :nb, :bc, :diff, :atype, 'active')
                     """), {
-                        "date": d, "room": booking[d]["room"],
+                        "date": c["date"], "room": category_label,
-                        "nb": newbook[d], "bc": booking[d]["rate"],
+                        "nb": c["newbook_rate"], "bc": c["booking_rate"],
-                        "diff": round(dev, 2),
+                        "diff": round(c["dev_pct"], 2),
-                        "atype": "higher" if dev > 0 else "lower",
+                        "atype": "higher" if c["dev_pct"] > 0 else "lower",
                     })
                     created += 1
             else:
@@ -187,7 +305,7 @@ def run_parity_check() -> dict:
         db.commit()
         summary = {
             "status": "ok",
-            "dates_compared": len(common_dates),
+            "dates_compared": len(comparisons),
             "created": created,
             "updated": updated,
             "resolved": resolved,