Experimented with fuzzy matching but ultimatly not a good idea. 2 false positives and nothing more

sql_analysis.md

@@ -3,7 +3,7 @@
 select sum(room.total_revenue::float)
 	
 	from alpinebits.conversions as con 
-	join alpinebits.room_reservations as room on room.conversion_id = con.id
+	join alpinebits.conversion_rooms as room on room.conversion_id = con.id
 	join alpinebits.reservations as res on res.id = con.reservation_id
 	
 	
@@ -21,7 +21,7 @@ select res.created_at, con.reservation_date, res.start_date, room.arrival_date,r
 	room.room_status
 	
 	from alpinebits.conversions as con 
-	join alpinebits.room_reservations as room on room.conversion_id = con.id
+	join alpinebits.conversion_rooms as room on room.conversion_id = con.id
 	join alpinebits.reservations as res on res.id = con.reservation_id
 	
 	
@@ -37,7 +37,7 @@ select res.created_at, con.reservation_date, res.start_date, room.arrival_date,r
 select round(sum(room.total_revenue::numeric)::numeric, 3), con.advertising_medium
 	
 	from alpinebits.conversions as con 
-	join alpinebits.room_reservations as room on room.conversion_id = con.id
+	join alpinebits.conversion_rooms as room on room.conversion_id = con.id
 
 	
 	

src/alpine_bits_python/api.py

@@ -39,7 +39,7 @@ from .alpinebits_server import (
     Version,
 )
 from .auth import generate_unique_id, validate_api_key
-from .config_loader import load_config
+from .config_loader import load_config, get_username_for_hotel
 from .const import CONF_GOOGLE_ACCOUNT, CONF_HOTEL_ID, CONF_META_ACCOUNT, HttpStatusCode
 from .conversion_service import ConversionService
 from .csv_import import CSVImporter
@@ -1142,15 +1142,25 @@ async def handle_wix_form_test(
 async def _process_csv_import_background(
     csv_content: str,
     filename: str,
-    hotel_code: str | None,
+    hotel_code: str,
     session_maker: SessionMaker,
     config: dict[str, Any],
     log_filename: Path,
 ):
-    """Background task to process CSV import.
+    """Background task to process CSV import with automatic acknowledgement.
 
     This runs in a separate asyncio task after the HTTP response is sent.
     Handles both file saving and database processing.
+    All imported reservations are automatically acknowledged using the username
+    associated with the hotel_code from the config.
+
+    Args:
+        csv_content: CSV content as string
+        filename: Original filename
+        hotel_code: Hotel code (mandatory) - used to get username for acknowledgements
+        session_maker: SessionMaker for creating database sessions
+        config: Application configuration
+        log_filename: Path to save the CSV file
     """
     try:
         # First, save the CSV file (in background)
@@ -1160,27 +1170,41 @@ async def _process_csv_import_background(
         # Now process the CSV import
         _LOGGER.info("Starting database processing of %s", filename)
 
+        # Get username for acknowledgements from config
+        username = get_username_for_hotel(config, hotel_code)
+
         # Create a new session for this background task
-        async with session_maker() as db_session:
+        db_session = await session_maker.create_session()
+        try:
             importer = CSVImporter(db_session, config)
-            stats = await importer.import_csv_file(str(log_filename), hotel_code, dryrun=False)
+            # Import with pre-acknowledgement enabled
+            stats = await importer.import_csv_file(
+                str(log_filename),
+                hotel_code,
+                dryrun=False,
+                pre_acknowledge=True,
+                client_id=hotel_code,
+                username=username
+            )
 
             _LOGGER.info(
                 "CSV import complete for %s: %s", filename, stats
             )
+        finally:
+            await db_session.close()
     except Exception:
         _LOGGER.exception(
             "Error processing CSV import in background for %s", filename
         )
 
 
-@api_router.put("/admin/import-csv/{filename:path}")
+@api_router.put("/admin/import-csv/{hotel_code}/{filename:path}")
 @limiter.limit(BURST_RATE_LIMIT)
 async def import_csv_endpoint(
     request: Request,
     background_tasks: BackgroundTasks,
+    hotel_code: str,
     filename: str,
-    hotel_code: str | None = None,
     credentials: tuple = Depends(validate_basic_auth),
     db_session=Depends(get_async_session),
     session_maker: SessionMaker = Depends(get_session_maker),
@@ -1193,16 +1217,18 @@ async def import_csv_endpoint(
     - fbclid/gclid tracking IDs
 
     Returns immediately with 202 Accepted while processing continues in background.
+    All imported reservations are automatically acknowledged using the username
+    associated with the hotel_code in the config.
 
     Requires basic authentication and saves CSV files to log directory.
     Supports gzip compression via Content-Encoding header.
 
     Args:
+        hotel_code: Hotel code (mandatory) - used to get username for acknowledgements
         filename: Name for the CSV file (used for logging)
-        hotel_code: Optional hotel code to override CSV values
         credentials: Basic auth credentials
 
-    Example: PUT /api/admin/import-csv/reservations.csv
+    Example: PUT /api/admin/import-csv/39054_001/reservations.csv
 
     """
     try:

src/alpine_bits_python/config_loader.py

@@ -329,3 +329,8 @@ class Config:
 # For backward compatibility
 def load_config():
     return Config().config
+
+
+def get_username_for_hotel(config: dict, hotel_code: str) -> str:
+    """Get the username associated with a hotel_code from config."""
+    return next(h.get("username") for h in config.get("alpine_bits_auth", []) if h.get("hotel_id") == hotel_code)

src/alpine_bits_python/conversion_service.py

@@ -2,8 +2,9 @@
 
 import asyncio
 import xml.etree.ElementTree as ET
-from datetime import datetime
+from datetime import datetime, date
 from decimal import Decimal
+from difflib import SequenceMatcher
 from typing import Any
 
 from sqlalchemy import or_, select
@@ -430,6 +431,7 @@ class ConversionService:
                 guest_first_name,
                 guest_last_name,
                 guest_email,
+                guest_birth_date,
                 advertising_partner,
                 session,
             )
@@ -703,6 +705,7 @@ class ConversionService:
         guest_first_name: str | None = None,
         guest_last_name: str | None = None,
         guest_email: str | None = None,
+        guest_birth_date: date | None = None,
         advertising_partner: str | None = None,
         session: AsyncSession | None = None,
     ) -> dict[str, Any]:
@@ -719,6 +722,7 @@ class ConversionService:
             guest_first_name: Guest first name for matching
             guest_last_name: Guest last name for matching
             guest_email: Guest email for matching
+            guest_birth_date: Guest birth date (optional, improves matching confidence)
             advertising_partner: Partner info (matches utm_medium for additional filtering)
             session: AsyncSession to use. If None, uses self.session.
 
@@ -765,7 +769,7 @@ class ConversionService:
             guest_email or guest_first_name or guest_last_name
         ):
             matched_reservation = await self._match_by_guest_details(
-                hotel_id, guest_first_name, guest_last_name, guest_email, session
+                hotel_id, guest_first_name, guest_last_name, guest_email, guest_birth_date, session
             )
 
             if matched_reservation:
@@ -896,9 +900,10 @@ class ConversionService:
         guest_first_name: str | None,
         guest_last_name: str | None,
         guest_email: str | None,
+        guest_birth_date: date | None = None,
         session: AsyncSession | None = None,
     ) -> Reservation | None:
-        """Match reservation by guest name and email using cached data.
+        """Match reservation by guest name, email, and birth date using cached data.
 
         This method uses the reservation cache populated at the start of XML processing.
         If cache is not available, falls back to database queries.
@@ -908,6 +913,7 @@ class ConversionService:
             guest_first_name: Guest first name
             guest_last_name: Guest last name
             guest_email: Guest email
+            guest_birth_date: Guest birth date (optional, improves matching confidence)
             session: AsyncSession to use. If None, uses self.session.
 
         Returns:
@@ -956,7 +962,7 @@ class ConversionService:
         all_reservations = db_result.scalars().all()
 
         return self._match_reservations_by_guest_details(
-            all_reservations, guest_first_name, guest_last_name, guest_email
+            all_reservations, guest_first_name, guest_last_name, guest_email, guest_birth_date
         )
 
     def _match_reservations_by_guest_details(
@@ -965,65 +971,148 @@ class ConversionService:
         guest_first_name: str | None,
         guest_last_name: str | None,
         guest_email: str | None,
+        guest_birth_date: date | None = None,
     ) -> Reservation | None:
-        """Match a reservation from a list by guest name and email (non-async).
+        """Match a reservation from a list by guest name, email, and birth date (non-async).
+
+        Uses a scoring system to find the best match:
+        - Exact email match: 100 points (highest confidence)
+        - Exact name + exact birth date: 90 points (very high confidence)
+        - Fuzzy name + exact birth date: 75-85 points (medium-high confidence)
+        - Exact name only: 50 points (lower confidence, could be common names)
+        - Fuzzy name only: 50-60 points (requires high similarity, works without birth date)
 
         Args:
             reservations: List of reservations to search through
             guest_first_name: Guest first name
             guest_last_name: Guest last name
             guest_email: Guest email
+            guest_birth_date: Guest birth date (optional, improves confidence)
 
         Returns:
-            Matched Reservation or None
+            Matched Reservation with highest score (only if score >= 50), or None
 
         """
-        # Filter by guest details
+        FUZZY_THRESHOLD = 0.75  # Minimum similarity ratio for fuzzy matching
+        MIN_SCORE = 50  # Minimum score to consider a match valid
+
         candidates = []
         for reservation in reservations:
             customer = reservation.customer
             if not customer:
                 continue
 
-            # Match by email (highest priority)
+            score = 0
-            if guest_email:
+            match_details = []
-                if (
+
-                    customer.email_address
+            # Strategy 1: Match by email (highest priority)
-                    and customer.email_address.lower() == guest_email.lower()
+            if guest_email and customer.email_address:
-                ):
+                if customer.email_address.lower() == guest_email.lower():
+                    score = 100
+                    match_details.append("exact_email")
                     _LOGGER.info(
-                        "Found exact email match for %s (reservation_id=%s)",
+                        "Found exact email match for %s (reservation_id=%s, score=%d)",
                         guest_email,
                         reservation.id,
+                        score,
                     )
-                    candidates.append((reservation, 3))  # Highest score
+                    candidates.append((reservation, score))
                     continue
 
-            # Match by name (first + last)
+            # Strategy 2: Match by name and birth date
             if guest_first_name and guest_last_name:
-                first_match = (
+                # Exact name matching
+                first_exact = (
                     customer.given_name
                     and customer.given_name.lower() == guest_first_name.lower()
                 )
-                last_match = (
+                last_exact = (
                     customer.surname
                     and customer.surname.lower() == guest_last_name.lower()
                 )
 
-                if first_match and last_match:
+                # Fuzzy name matching
+                first_fuzzy_ratio = 0.0
+                last_fuzzy_ratio = 0.0
+                if customer.given_name:
+                    first_fuzzy_ratio = SequenceMatcher(
+                        None,
+                        customer.given_name.lower(),
+                        guest_first_name.lower(),
+                    ).ratio()
+                if customer.surname:
+                    last_fuzzy_ratio = SequenceMatcher(
+                        None,
+                        customer.surname.lower(),
+                        guest_last_name.lower(),
+                    ).ratio()
+
+                # Birth date matching
+                birth_date_match = (
+                    guest_birth_date
+                    and customer.birth_date
+                    and customer.birth_date == str(guest_birth_date)
+                )
+
+                # Calculate score based on name matching
+                if first_exact and last_exact:
+                    if birth_date_match:
+                        score = 90  # Exact name + exact birth date
+                        match_details.append("exact_name_exact_birthdate")
+                    else:
+                        score = 50  # Exact name only
+                        match_details.append("exact_name_only")
                     _LOGGER.info(
-                        "Found exact name match for %s %s (reservation_id=%s)",
+                        "Found exact name match for %s %s (reservation_id=%s, score=%d, match_details=%s)",
                         guest_first_name,
                         guest_last_name,
                         reservation.id,
+                        score,
+                        match_details,
                     )
-                    candidates.append((reservation, 2))  # Medium-high score
+                elif (
-                    continue
+                    first_fuzzy_ratio >= FUZZY_THRESHOLD
+                    and last_fuzzy_ratio >= FUZZY_THRESHOLD
+                ):
+                    # Both first and last names are fuzzy matched
+                    fuzzy_score = min(first_fuzzy_ratio, last_fuzzy_ratio)
+                    if birth_date_match:
+                        # Fuzzy name + exact birth date (high confidence)
+                        score = int(75 + (fuzzy_score - FUZZY_THRESHOLD) * 40)
+                        match_details.append(
+                            f"fuzzy_name_exact_birthdate(ratio={fuzzy_score:.2f})"
+                        )
+                    else:
+                        # Fuzzy name only (lower confidence, but still usable)
+                        # Scale from 50-60 based on similarity ratio above threshold
+                        score = int(50 + (fuzzy_score - FUZZY_THRESHOLD) * 40)
+                        match_details.append(f"fuzzy_name_only(ratio={fuzzy_score:.2f})")
+
+                    _LOGGER.info(
+                        "Found fuzzy name match for %s %s (reservation_id=%s, score=%d, match_details=%s)",
+                        guest_first_name,
+                        guest_last_name,
+                        reservation.id,
+                        score,
+                        match_details,
+                    )
+
+            # Only add candidates that meet minimum score threshold
+            if score >= MIN_SCORE:
+                candidates.append((reservation, score))
 
         # Return highest-scoring match
         if candidates:
             candidates.sort(key=lambda x: x[1], reverse=True)
-            return candidates[0][0]
+            best_match = candidates[0][0]
+            best_score = candidates[0][1]
+            _LOGGER.debug(
+                "Selected best match (reservation_id=%s) with score=%d out of %d candidates",
+                best_match.id,
+                best_score,
+                len(candidates),
+            )
+            return best_match
 
         return None
 

src/alpine_bits_python/csv_import.py

@@ -116,6 +116,21 @@ class CSVImporter:
         self.customer_service = CustomerService(db_session)
         self.reservation_service = ReservationService(db_session)
 
+    def _get_hotel_info(self, hotel_code: str) -> tuple[str, str]:
+        """Get hotel name from config by hotel_code.
+
+        Args:
+            hotel_code: Hotel code to look up
+
+        Returns:
+            Tuple of (hotel_code, hotel_name) from config
+        """
+        for hotel in self.config.get("alpine_bits_auth", []):
+            if hotel.get("hotel_id") == hotel_code:
+                return hotel_code, hotel.get("hotel_name", "")
+        # Fallback to default if not found
+        return hotel_code, self.config.get("default_hotel_name", "Frangart Inn")
+
     async def find_duplicate_reservation(
         self,
         first_name: str,
@@ -184,13 +199,13 @@ class CSVImporter:
         return None
 
     async def import_csv_file(
-        self, csv_file_path: str, hotel_code: Optional[str] = None, dryrun: bool = False, pre_acknowledge: bool = False, client_id: Optional[str] = None, username: Optional[str] = None
+        self, csv_file_path: str, hotel_code: str, dryrun: bool = False, pre_acknowledge: bool = False, client_id: Optional[str] = None, username: Optional[str] = None
     ) -> dict[str, Any]:
         """Import reservations from a CSV file.
 
         Args:
             csv_file_path: Path to CSV file
-            hotel_code: Optional hotel code to override CSV values
+            hotel_code: Hotel code (mandatory) - used to look up hotel name from config
             dryrun: If True, parse and print first 10 rows as JSON without importing
             pre_acknowledge: If True, pre-acknowledges all imported reservations
             client_id: Client ID for pre-acknowledgement (required if pre_acknowledge=True)
@@ -432,16 +447,8 @@ class CSVImporter:
                 else:
                     submission_id = f"csv_import_{row_num}_{datetime.now().isoformat()}"
 
-                # Determine hotel code and name
+                # Determine hotel code and name (from config)
-                final_hotel_code = (
+                final_hotel_code, final_hotel_name = self._get_hotel_info(hotel_code)
-                    hotel_code
-                    or str(row.get("hotel_id", "")).strip()
-                    or self.config.get("default_hotel_code", "123")
-                )
-                final_hotel_name = (
-                    str(row.get("hotel_name", "")).strip()
-                    or self.config.get("default_hotel_name", "Frangart Inn")
-                )
 
                 # Parse room type fields if available
                 room_type_code = str(row.get("room_type_code", "")).strip() or None