Link previous trip shapes for GPS positioning on circular routes (#111)

Co-authored-by: Ariel Costas Guerrero <ariel@costas.dev>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: arielcostas <94913521+arielcostas@users.noreply.github.com>

5 files changed, 202 insertions, 65 deletions

diff --git a/src/gtfs_vigo_stops/src/proto/stop_schedule_pb2.py b/src/gtfs_vigo_stops/src/proto/stop_schedule_pb2.py
index d9f8e52..cb4f336 100644
--- a/src/gtfs_vigo_stops/src/proto/stop_schedule_pb2.py
+++ b/src/gtfs_vigo_stops/src/proto/stop_schedule_pb2.py
@@ -1,22 +1,11 @@
 # -*- coding: utf-8 -*-
 # Generated by the protocol buffer compiler.  DO NOT EDIT!
-# NO CHECKED-IN PROTOBUF GENCODE
 # source: stop_schedule.proto
-# Protobuf Python Version: 6.33.0
 """Generated protocol buffer code."""
+from google.protobuf.internal import builder as _builder
 from google.protobuf import descriptor as _descriptor
 from google.protobuf import descriptor_pool as _descriptor_pool
-from google.protobuf import runtime_version as _runtime_version
 from google.protobuf import symbol_database as _symbol_database
-from google.protobuf.internal import builder as _builder
-_runtime_version.ValidateProtobufRuntimeVersion(
-    _runtime_version.Domain.PUBLIC,
-    6,
-    33,
-    0,
-    '',
-    'stop_schedule.proto'
-)
 # @@protoc_insertion_point(imports)
 
 _sym_db = _symbol_database.Default()
@@ -24,20 +13,20 @@ _sym_db = _symbol_database.Default()
 
 
 
-DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x13stop_schedule.proto\x12\x05proto\"!\n\tEpsg25829\x12\t\n\x01x\x18\x01 \x01(\x01\x12\t\n\x01y\x18\x02 \x01(\x01\"\xe3\x03\n\x0cStopArrivals\x12\x0f\n\x07stop_id\x18\x01 \x01(\t\x12\"\n\x08location\x18\x03 \x01(\x0b\x32\x10.proto.Epsg25829\x12\x36\n\x08\x61rrivals\x18\x05 \x03(\x0b\x32$.proto.StopArrivals.ScheduledArrival\x1a\xe5\x02\n\x10ScheduledArrival\x12\x12\n\nservice_id\x18\x01 \x01(\t\x12\x0f\n\x07trip_id\x18\x02 \x01(\t\x12\x0c\n\x04line\x18\x03 \x01(\t\x12\r\n\x05route\x18\x04 \x01(\t\x12\x10\n\x08shape_id\x18\x05 \x01(\t\x12\x1b\n\x13shape_dist_traveled\x18\x06 \x01(\x01\x12\x15\n\rstop_sequence\x18\x0b \x01(\r\x12\x14\n\x0cnext_streets\x18\x0c \x03(\t\x12\x15\n\rstarting_code\x18\x15 \x01(\t\x12\x15\n\rstarting_name\x18\x16 \x01(\t\x12\x15\n\rstarting_time\x18\x17 \x01(\t\x12\x14\n\x0c\x63\x61lling_time\x18! \x01(\t\x12\x13\n\x0b\x63\x61lling_ssm\x18\" \x01(\r\x12\x15\n\rterminus_code\x18) \x01(\t\x12\x15\n\rterminus_name\x18* \x01(\t\x12\x15\n\rterminus_time\x18+ \x01(\t\";\n\x05Shape\x12\x10\n\x08shape_id\x18\x01 \x01(\t\x12 \n\x06points\x18\x03 \x03(\x0b\x32\x10.proto.Epsg25829B$\xaa\x02!Costasdev.Busurbano.Backend.Typesb\x06proto3')
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x13stop_schedule.proto\x12\x05proto\"!\n\tEpsg25829\x12\t\n\x01x\x18\x01 \x01(\x01\x12\t\n\x01y\x18\x02 \x01(\x01\"\x83\x04\n\x0cStopArrivals\x12\x0f\n\x07stop_id\x18\x01 \x01(\t\x12\"\n\x08location\x18\x03 \x01(\x0b\x32\x10.proto.Epsg25829\x12\x36\n\x08\x61rrivals\x18\x05 \x03(\x0b\x32$.proto.StopArrivals.ScheduledArrival\x1a\x85\x03\n\x10ScheduledArrival\x12\x12\n\nservice_id\x18\x01 \x01(\t\x12\x0f\n\x07trip_id\x18\x02 \x01(\t\x12\x0c\n\x04line\x18\x03 \x01(\t\x12\r\n\x05route\x18\x04 \x01(\t\x12\x10\n\x08shape_id\x18\x05 \x01(\t\x12\x1b\n\x13shape_dist_traveled\x18\x06 \x01(\x01\x12\x15\n\rstop_sequence\x18\x0b \x01(\r\x12\x14\n\x0cnext_streets\x18\x0c \x03(\t\x12\x15\n\rstarting_code\x18\x15 \x01(\t\x12\x15\n\rstarting_name\x18\x16 \x01(\t\x12\x15\n\rstarting_time\x18\x17 \x01(\t\x12\x14\n\x0c\x63\x61lling_time\x18! \x01(\t\x12\x13\n\x0b\x63\x61lling_ssm\x18\" \x01(\r\x12\x15\n\rterminus_code\x18) \x01(\t\x12\x15\n\rterminus_name\x18* \x01(\t\x12\x15\n\rterminus_time\x18+ \x01(\t\x12\x1e\n\x16previous_trip_shape_id\x18\x33 \x01(\t\";\n\x05Shape\x12\x10\n\x08shape_id\x18\x01 \x01(\t\x12 \n\x06points\x18\x03 \x03(\x0b\x32\x10.proto.Epsg25829B$\xaa\x02!Costasdev.Busurbano.Backend.Typesb\x06proto3')
+
+_builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, globals())
+_builder.BuildTopDescriptorsAndMessages(DESCRIPTOR, 'stop_schedule_pb2', globals())
+if _descriptor._USE_C_DESCRIPTORS == False:
 
-_globals = globals()
-_builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
-_builder.BuildTopDescriptorsAndMessages(DESCRIPTOR, 'stop_schedule_pb2', _globals)
-if not _descriptor._USE_C_DESCRIPTORS:
-  _globals['DESCRIPTOR']._loaded_options = None
-  _globals['DESCRIPTOR']._serialized_options = b'\252\002!Costasdev.Busurbano.Backend.Types'
-  _globals['_EPSG25829']._serialized_start=30
-  _globals['_EPSG25829']._serialized_end=63
-  _globals['_STOPARRIVALS']._serialized_start=66
-  _globals['_STOPARRIVALS']._serialized_end=549
-  _globals['_STOPARRIVALS_SCHEDULEDARRIVAL']._serialized_start=192
-  _globals['_STOPARRIVALS_SCHEDULEDARRIVAL']._serialized_end=549
-  _globals['_SHAPE']._serialized_start=551
-  _globals['_SHAPE']._serialized_end=610
+  DESCRIPTOR._options = None
+  DESCRIPTOR._serialized_options = b'\252\002!Costasdev.Busurbano.Backend.Types'
+  _EPSG25829._serialized_start=30
+  _EPSG25829._serialized_end=63
+  _STOPARRIVALS._serialized_start=66
+  _STOPARRIVALS._serialized_end=581
+  _STOPARRIVALS_SCHEDULEDARRIVAL._serialized_start=192
+  _STOPARRIVALS_SCHEDULEDARRIVAL._serialized_end=581
+  _SHAPE._serialized_start=583
+  _SHAPE._serialized_end=642
 # @@protoc_insertion_point(module_scope)
diff --git a/src/gtfs_vigo_stops/src/proto/stop_schedule_pb2.pyi b/src/gtfs_vigo_stops/src/proto/stop_schedule_pb2.pyi
index 615999f..355798f 100644
--- a/src/gtfs_vigo_stops/src/proto/stop_schedule_pb2.pyi
+++ b/src/gtfs_vigo_stops/src/proto/stop_schedule_pb2.pyi
@@ -1,68 +1,69 @@
 from google.protobuf.internal import containers as _containers
 from google.protobuf import descriptor as _descriptor
 from google.protobuf import message as _message
-from collections.abc import Iterable as _Iterable, Mapping as _Mapping
-from typing import ClassVar as _ClassVar, Optional as _Optional, Union as _Union
+from typing import ClassVar as _ClassVar, Iterable as _Iterable, Mapping as _Mapping, Optional as _Optional, Union as _Union
 
 DESCRIPTOR: _descriptor.FileDescriptor
 
 class Epsg25829(_message.Message):
-    __slots__ = ()
+    __slots__ = ["x", "y"]
     X_FIELD_NUMBER: _ClassVar[int]
     Y_FIELD_NUMBER: _ClassVar[int]
     x: float
     y: float
     def __init__(self, x: _Optional[float] = ..., y: _Optional[float] = ...) -> None: ...
 
+class Shape(_message.Message):
+    __slots__ = ["points", "shape_id"]
+    POINTS_FIELD_NUMBER: _ClassVar[int]
+    SHAPE_ID_FIELD_NUMBER: _ClassVar[int]
+    points: _containers.RepeatedCompositeFieldContainer[Epsg25829]
+    shape_id: str
+    def __init__(self, shape_id: _Optional[str] = ..., points: _Optional[_Iterable[_Union[Epsg25829, _Mapping]]] = ...) -> None: ...
+
 class StopArrivals(_message.Message):
-    __slots__ = ()
+    __slots__ = ["arrivals", "location", "stop_id"]
     class ScheduledArrival(_message.Message):
-        __slots__ = ()
-        SERVICE_ID_FIELD_NUMBER: _ClassVar[int]
-        TRIP_ID_FIELD_NUMBER: _ClassVar[int]
+        __slots__ = ["calling_ssm", "calling_time", "line", "next_streets", "previous_trip_shape_id", "route", "service_id", "shape_dist_traveled", "shape_id", "starting_code", "starting_name", "starting_time", "stop_sequence", "terminus_code", "terminus_name", "terminus_time", "trip_id"]
+        CALLING_SSM_FIELD_NUMBER: _ClassVar[int]
+        CALLING_TIME_FIELD_NUMBER: _ClassVar[int]
         LINE_FIELD_NUMBER: _ClassVar[int]
+        NEXT_STREETS_FIELD_NUMBER: _ClassVar[int]
+        PREVIOUS_TRIP_SHAPE_ID_FIELD_NUMBER: _ClassVar[int]
         ROUTE_FIELD_NUMBER: _ClassVar[int]
-        SHAPE_ID_FIELD_NUMBER: _ClassVar[int]
+        SERVICE_ID_FIELD_NUMBER: _ClassVar[int]
         SHAPE_DIST_TRAVELED_FIELD_NUMBER: _ClassVar[int]
-        STOP_SEQUENCE_FIELD_NUMBER: _ClassVar[int]
-        NEXT_STREETS_FIELD_NUMBER: _ClassVar[int]
+        SHAPE_ID_FIELD_NUMBER: _ClassVar[int]
         STARTING_CODE_FIELD_NUMBER: _ClassVar[int]
         STARTING_NAME_FIELD_NUMBER: _ClassVar[int]
         STARTING_TIME_FIELD_NUMBER: _ClassVar[int]
-        CALLING_TIME_FIELD_NUMBER: _ClassVar[int]
-        CALLING_SSM_FIELD_NUMBER: _ClassVar[int]
+        STOP_SEQUENCE_FIELD_NUMBER: _ClassVar[int]
         TERMINUS_CODE_FIELD_NUMBER: _ClassVar[int]
         TERMINUS_NAME_FIELD_NUMBER: _ClassVar[int]
         TERMINUS_TIME_FIELD_NUMBER: _ClassVar[int]
-        service_id: str
-        trip_id: str
+        TRIP_ID_FIELD_NUMBER: _ClassVar[int]
+        calling_ssm: int
+        calling_time: str
         line: str
+        next_streets: _containers.RepeatedScalarFieldContainer[str]
+        previous_trip_shape_id: str
         route: str
-        shape_id: str
+        service_id: str
         shape_dist_traveled: float
-        stop_sequence: int
-        next_streets: _containers.RepeatedScalarFieldContainer[str]
+        shape_id: str
         starting_code: str
         starting_name: str
         starting_time: str
-        calling_time: str
-        calling_ssm: int
+        stop_sequence: int
         terminus_code: str
         terminus_name: str
         terminus_time: str
-        def __init__(self, service_id: _Optional[str] = ..., trip_id: _Optional[str] = ..., line: _Optional[str] = ..., route: _Optional[str] = ..., shape_id: _Optional[str] = ..., shape_dist_traveled: _Optional[float] = ..., stop_sequence: _Optional[int] = ..., next_streets: _Optional[_Iterable[str]] = ..., starting_code: _Optional[str] = ..., starting_name: _Optional[str] = ..., starting_time: _Optional[str] = ..., calling_time: _Optional[str] = ..., calling_ssm: _Optional[int] = ..., terminus_code: _Optional[str] = ..., terminus_name: _Optional[str] = ..., terminus_time: _Optional[str] = ...) -> None: ...
-    STOP_ID_FIELD_NUMBER: _ClassVar[int]
-    LOCATION_FIELD_NUMBER: _ClassVar[int]
+        trip_id: str
+        def __init__(self, service_id: _Optional[str] = ..., trip_id: _Optional[str] = ..., line: _Optional[str] = ..., route: _Optional[str] = ..., shape_id: _Optional[str] = ..., shape_dist_traveled: _Optional[float] = ..., stop_sequence: _Optional[int] = ..., next_streets: _Optional[_Iterable[str]] = ..., starting_code: _Optional[str] = ..., starting_name: _Optional[str] = ..., starting_time: _Optional[str] = ..., calling_time: _Optional[str] = ..., calling_ssm: _Optional[int] = ..., terminus_code: _Optional[str] = ..., terminus_name: _Optional[str] = ..., terminus_time: _Optional[str] = ..., previous_trip_shape_id: _Optional[str] = ...) -> None: ...
     ARRIVALS_FIELD_NUMBER: _ClassVar[int]
-    stop_id: str
-    location: Epsg25829
+    LOCATION_FIELD_NUMBER: _ClassVar[int]
+    STOP_ID_FIELD_NUMBER: _ClassVar[int]
     arrivals: _containers.RepeatedCompositeFieldContainer[StopArrivals.ScheduledArrival]
+    location: Epsg25829
+    stop_id: str
     def __init__(self, stop_id: _Optional[str] = ..., location: _Optional[_Union[Epsg25829, _Mapping]] = ..., arrivals: _Optional[_Iterable[_Union[StopArrivals.ScheduledArrival, _Mapping]]] = ...) -> None: ...
-
-class Shape(_message.Message):
-    __slots__ = ()
-    SHAPE_ID_FIELD_NUMBER: _ClassVar[int]
-    POINTS_FIELD_NUMBER: _ClassVar[int]
-    shape_id: str
-    points: _containers.RepeatedCompositeFieldContainer[Epsg25829]
-    def __init__(self, shape_id: _Optional[str] = ..., points: _Optional[_Iterable[_Union[Epsg25829, _Mapping]]] = ...) -> None: ...
diff --git a/src/gtfs_vigo_stops/src/report_writer.py b/src/gtfs_vigo_stops/src/report_writer.py
index 695931f..f6d8763 100644
--- a/src/gtfs_vigo_stops/src/report_writer.py
+++ b/src/gtfs_vigo_stops/src/report_writer.py
@@ -51,6 +51,7 @@ def write_stop_protobuf(
                 terminus_code=arrival["terminus_code"],
                 terminus_name=arrival["terminus_name"],
                 terminus_time=arrival["terminus_time"],
+                previous_trip_shape_id=arrival.get("previous_trip_shape_id", ""),
             )
             for arrival in arrivals
         ],
diff --git a/src/gtfs_vigo_stops/src/trips.py b/src/gtfs_vigo_stops/src/trips.py
index 0c1375c..0cedd26 100644
--- a/src/gtfs_vigo_stops/src/trips.py
+++ b/src/gtfs_vigo_stops/src/trips.py
@@ -10,18 +10,19 @@ class TripLine:
     """
     Class representing a trip line in the GTFS data.
     """
-    def __init__(self, route_id: str, service_id: str, trip_id: str, headsign: str, direction_id: int, shape_id: str|None = None):
+    def __init__(self, route_id: str, service_id: str, trip_id: str, headsign: str, direction_id: int, shape_id: str|None = None, block_id: str|None = None):
         self.route_id = route_id
         self.service_id = service_id
         self.trip_id = trip_id
         self.headsign = headsign
         self.direction_id = direction_id
         self.shape_id = shape_id
+        self.block_id = block_id
         self.route_short_name = ""
         self.route_color = ""
 
     def __str__(self):
-        return f"TripLine({self.route_id=}, {self.service_id=}, {self.trip_id=}, {self.headsign=}, {self.direction_id=}, {self.shape_id=})"
+        return f"TripLine({self.route_id=}, {self.service_id=}, {self.trip_id=}, {self.headsign=}, {self.direction_id=}, {self.shape_id=}, {self.block_id=})"
 
 
 TRIPS_BY_SERVICE_ID: dict[str, dict[str, list[TripLine]]] = {}
@@ -74,6 +75,13 @@ def get_trips_for_services(feed_dir: str, service_ids: list[str]) -> dict[str, l
             else:
                 logger.warning("shape_id column not found in trips.txt")
 
+            # Check if block_id column exists
+            block_id_index = None
+            if 'block_id' in header:
+                block_id_index = header.index('block_id')
+            else:
+                logger.info("block_id column not found in trips.txt")
+
             # Initialize cache for this feed directory
             TRIPS_BY_SERVICE_ID[feed_dir] = {}
             
@@ -96,6 +104,11 @@ def get_trips_for_services(feed_dir: str, service_ids: list[str]) -> dict[str, l
                 if shape_id_index is not None and shape_id_index < len(parts):
                     shape_id = parts[shape_id_index] if parts[shape_id_index] else None
 
+                # Get block_id if available
+                block_id = None
+                if block_id_index is not None and block_id_index < len(parts):
+                    block_id = parts[block_id_index] if parts[block_id_index] else None
+
                 trip_line = TripLine(
                     route_id=parts[route_id_index],
                     service_id=service_id,
@@ -103,7 +116,8 @@ def get_trips_for_services(feed_dir: str, service_ids: list[str]) -> dict[str, l
                     headsign=parts[headsign_index],
                     direction_id=int(
                         parts[direction_id_index] if parts[direction_id_index] else -1),
-                    shape_id=shape_id
+                    shape_id=shape_id,
+                    block_id=block_id
                 )
                 
                 TRIPS_BY_SERVICE_ID[feed_dir][service_id].append(trip_line)
diff --git a/src/gtfs_vigo_stops/stop_report.py b/src/gtfs_vigo_stops/stop_report.py
index a827eaa..cee11ea 100644
--- a/src/gtfs_vigo_stops/stop_report.py
+++ b/src/gtfs_vigo_stops/stop_report.py
@@ -4,7 +4,7 @@ import shutil
 import sys
 import time
 import traceback
-from typing import Any, Dict, List
+from typing import Any, Dict, List, Optional, Tuple
 
 from src.shapes import process_shapes
 from src.common import get_all_feed_dates
@@ -13,10 +13,10 @@ from src.logger import get_logger
 from src.report_writer import write_stop_json, write_stop_protobuf
 from src.routes import load_routes
 from src.services import get_active_services
-from src.stop_times import get_stops_for_trips
+from src.stop_times import get_stops_for_trips, StopTime
 from src.stops import get_all_stops, get_all_stops_by_code, get_numeric_code
 from src.street_name import get_street_name, normalise_stop_name
-from src.trips import get_trips_for_services
+from src.trips import get_trips_for_services, TripLine
 
 logger = get_logger("stop_report")
 
@@ -143,6 +143,130 @@ def is_next_day_service(time_str: str) -> bool:
         return False
 
 
+def parse_trip_id_components(trip_id: str) -> Optional[Tuple[str, str, int]]:
+    """
+    Parse a trip ID in format XXXYYY-Z or XXXYYY_Z where:
+    - XXX = line number (e.g., 003)
+    - YYY = shift/internal ID (e.g., 001)
+    - Z = trip number (e.g., 12)
+    
+    Supported formats:
+    1. ..._XXXYYY_Z (e.g. "C1 01SNA00_001001_18")
+    2. ..._XXXYYY-Z (e.g. "VIGO_20241122_003001-12")
+    
+    Returns tuple of (line, shift_id, trip_number) or None if parsing fails.
+    """
+    try:
+        parts = trip_id.split("_")
+        if len(parts) < 2:
+            return None
+
+        # Try format 1: ..._XXXYYY_Z
+        # Check if second to last part is 6 digits (XXXYYY) and last part is numeric
+        if len(parts) >= 2:
+            shift_part = parts[-2]
+            trip_num_str = parts[-1]
+            
+            if len(shift_part) == 6 and shift_part.isdigit() and trip_num_str.isdigit():
+                line = shift_part[:3]
+                shift_id = shift_part[3:6]
+                trip_number = int(trip_num_str)
+                return (line, shift_id, trip_number)
+
+        # Try format 2: ..._XXXYYY-Z
+        # The trip ID is the last part in format XXXYYY-Z
+        trip_part = parts[-1]
+        
+        if "-" in trip_part:
+            shift_part, trip_num_str = trip_part.split("-", 1)
+            
+            # shift_part should be 6 digits: XXXYYY
+            if len(shift_part) == 6 and shift_part.isdigit():
+                line = shift_part[:3]  # First 3 digits
+                shift_id = shift_part[3:6]  # Next 3 digits
+                trip_number = int(trip_num_str)
+                return (line, shift_id, trip_number)
+                
+        return None
+    except (ValueError, IndexError):
+        return None
+
+
+def build_trip_previous_shape_map(
+    trips: Dict[str, List[TripLine]],
+    stops_for_all_trips: Dict[str, List[StopTime]],
+) -> Dict[str, Optional[str]]:
+    """
+    Build a mapping from trip_id to previous_trip_shape_id.
+    
+    Links trips based on trip ID structure (XXXYYY-Z) where trips with the same
+    XXX (line) and YYY (shift) and sequential Z (trip numbers) are connected
+    if the terminus of trip N matches the start of trip N+1.
+    
+    Args:
+        trips: Dictionary of service_id -> list of trips
+        stops_for_all_trips: Dictionary of trip_id -> list of stop times
+        
+    Returns:
+        Dictionary mapping trip_id to previous_trip_shape_id (or None)
+    """
+    trip_previous_shape: Dict[str, Optional[str]] = {}
+    
+    # Collect all trips across all services
+    all_trips_list: List[TripLine] = []
+    for trip_list in trips.values():
+        all_trips_list.extend(trip_list)
+    
+    # Group trips by shift ID (line + shift combination)
+    trips_by_shift: Dict[str, List[Tuple[TripLine, int, str, str]]] = {}
+    
+    for trip in all_trips_list:
+        parsed = parse_trip_id_components(trip.trip_id)
+        if not parsed:
+            continue
+        
+        line, shift_id, trip_number = parsed
+        shift_key = f"{line}{shift_id}"
+        
+        trip_stops = stops_for_all_trips.get(trip.trip_id)
+        if not trip_stops or len(trip_stops) < 2:
+            continue
+        
+        first_stop = trip_stops[0]
+        last_stop = trip_stops[-1]
+        
+        if shift_key not in trips_by_shift:
+            trips_by_shift[shift_key] = []
+        
+        trips_by_shift[shift_key].append((
+            trip,
+            trip_number,
+            first_stop.stop_id,
+            last_stop.stop_id
+        ))
+    
+    # For each shift, sort trips by trip number and link consecutive trips
+    for shift_key, shift_trips in trips_by_shift.items():
+        # Sort by trip number
+        shift_trips.sort(key=lambda x: x[1])
+        
+        # Link consecutive trips if their stops match
+        for i in range(1, len(shift_trips)):
+            current_trip, current_num, current_start_stop, _ = shift_trips[i]
+            prev_trip, prev_num, _, prev_end_stop = shift_trips[i - 1]
+            
+            # Check if trips are consecutive (trip numbers differ by 1),
+            # if previous trip's terminus matches current trip's start,
+            # and if both trips have valid shape IDs
+            if (current_num == prev_num + 1 and 
+                prev_end_stop == current_start_stop and 
+                prev_trip.shape_id and
+                current_trip.shape_id):
+                trip_previous_shape[current_trip.trip_id] = prev_trip.shape_id
+    
+    return trip_previous_shape
+
+
 def get_stop_arrivals(feed_dir: str, date: str) -> Dict[str, List[Dict[str, Any]]]:
     """
     Process trips for the given date and organize stop arrivals.
@@ -192,6 +316,10 @@ def get_stop_arrivals(feed_dir: str, date: str) -> Dict[str, List[Dict[str, Any]
     stops_for_all_trips = get_stops_for_trips(feed_dir, all_trip_ids)
     logger.info(f"Precomputed stops for {len(stops_for_all_trips)} trips.")
 
+    # Build mapping from trip_id to previous trip's shape_id
+    trip_previous_shape_map = build_trip_previous_shape_map(trips, stops_for_all_trips)
+    logger.info(f"Built previous trip shape mapping for {len(trip_previous_shape_map)} trips.")
+
     # Load routes information
     routes = load_routes(feed_dir)
     logger.info(f"Loaded {len(routes)} routes from feed.")
@@ -335,6 +463,9 @@ def get_stop_arrivals(feed_dir: str, date: str) -> Dict[str, List[Dict[str, Any]
                         next_streets = []
 
                     trip_id_fmt = "_".join(trip_id.split("_")[1:3])
+                    
+                    # Get previous trip shape_id if available
+                    previous_trip_shape_id = trip_previous_shape_map.get(trip_id, "")
 
                     stop_arrivals[stop_code].append(
                         {
@@ -356,6 +487,7 @@ def get_stop_arrivals(feed_dir: str, date: str) -> Dict[str, List[Dict[str, Any]
                             "terminus_code": terminus_code,
                             "terminus_name": terminus_name,
                             "terminus_time": final_terminus_time,
+                            "previous_trip_shape_id": previous_trip_shape_id,
                         }
                     )