Trip Scheduling Choice -- Same Results Single Process and Multi-Process #1005

dhensle · 2025-10-21T17:58:03Z

Fix for #891

This pull request introduces several improvements and refactors to the trip scheduling and departure choice components, settings, and configuration files in the ActivitySim ABM prototype and test suite. The main changes include updating configuration schemas for consistency, refactoring settings classes for better inheritance and maintainability, improving sorting and spec reading logic, and enhancing test coverage and utility functions.

Configuration and Settings Refactoring

Updated the trip scheduling and departure choice configuration files to use SPEC instead of SPECIFICATION, standardized preprocessor settings, and added compute_settings and protect_columns for consistency and clarity. (activitysim/examples/prototype_arc/configs/trip_scheduling_choice.yaml, activitysim/abm/test/test_misc/configs_test_misc/trip_scheduling_choice.yaml, activitysim/abm/test/test_misc/configs_test_misc/trip_departure_choice.yaml, [1] [2] [3]
Refactored the TripSchedulingChoiceSettings class to inherit from LogitComponentSettings and removed redundant fields, streamlining the settings schema for trip scheduling choice. (activitysim/abm/models/trip_scheduling_choice.py, activitysim/abm/models/trip_scheduling_choice.pyL347-R352)

Logic and Code Improvements

Added sorting by tour_id and SCHEDULE_ID in the schedule generation function to ensure deterministic and organized output. (activitysim/abm/models/trip_scheduling_choice.py, activitysim/abm/models/trip_scheduling_choice.pyR85) (This was the actual fix to Reproducible random does not work in trip scheduling choice #891)
Changed the specification file reference from SPECIFICATION to SPEC in the segment spec reading function for consistency with configuration changes. (activitysim/abm/models/trip_scheduling_choice.py, activitysim/abm/models/trip_scheduling_choice.pyL210-R212)

Testing Enhancements

Un-commented and improved the trip departure choice test suite, added utility for canonical directory setup, and updated tests to use the new settings schema and workflow state initialization. (activitysim/abm/test/test_misc/test_trip_departure_choice.py, activitysim/abm/test/test_misc/test_trip_departure_choice.pyL1-R177)

Prototype and Test Configuration Updates

Added new and updated test configuration files for network level-of-service and time period settings, improving test coverage for different scenarios. (activitysim/abm/test/test_misc/configs_test_misc/network_los.yaml, [1]; activitysim/abm/test/test_misc/configs_test_misc/settings_60_min.yaml, [2]
Updated prototype settings to sample households, disable parking location, and enable multiprocessing for more flexible and efficient runs. (activitysim/examples/prototype_arc/configs/settings.yaml, [1] [2]

Miscellaneous

Removed unnecessary .gitignore entries in test output directories for cleaner repository management. (activitysim/examples/prototype_arc/test/output/.gitignore, [1]; activitysim/examples/prototype_arc/test/output/cache/.gitignore, [2]; activitysim/examples/prototype_arc/test/output/trace/.gitignore, [3]

single process run with

…ctivitysim into trip_scheduling_rng_fix

Copilot

Pull Request Overview

This PR makes several updates to the trip scheduling and departure choice models and their test configurations:

Migrates TripSchedulingChoiceSettings to inherit from LogitComponentSettings instead of defining fields manually
Updates configuration files to use standardized field names (SPEC instead of SPECIFICATION, preprocessor instead of PREPROCESSOR)
Un-comments and updates trip scheduling and departure choice tests to be functional again
Adds new test configuration files and fixtures
Removes placeholder .gitignore files from output directories
Adds a new simulation script and test configuration for the prototype_arc example

Reviewed Changes

Copilot reviewed 12 out of 14 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
activitysim/examples/prototype_arc/test/simulation.py	Adds a command-line script to run simulations via ActivitySim's CLI
activitysim/examples/prototype_arc/test/output/trace/.gitignore	Removes .gitignore for trace output directory
activitysim/examples/prototype_arc/test/output/cache/.gitignore	Removes .gitignore for cache output directory
activitysim/examples/prototype_arc/test/output/.gitignore	Removes .gitignore for main output directory
activitysim/examples/prototype_arc/configs/trip_scheduling_choice.yaml	Updates to use standard field names and adds compute_settings with protect_columns
activitysim/examples/prototype_arc/configs/settings.yaml	Adds test configuration settings including sample size, model settings check flag, and multiprocessing settings
activitysim/abm/test/test_misc/test_trip_scheduling_choice.py	Un-comments and updates tests with proper imports and fixtures, including state initialization
activitysim/abm/test/test_misc/test_trip_departure_choice.py	Un-comments and updates tests with state initialization and model settings loading
activitysim/abm/test/test_misc/configs_test_misc/trip_scheduling_choice.yaml	Adds test configuration file with standardized field names
activitysim/abm/test/test_misc/configs_test_misc/trip_departure_choice.yaml	Adds test configuration file using legacy field names
activitysim/abm/test/test_misc/configs_test_misc/settings_60_min.yaml	Adds skim settings configuration for 60-minute time periods
activitysim/abm/test/test_misc/configs_test_misc/network_los.yaml	Adds network level-of-service configuration
activitysim/abm/models/trip_scheduling_choice.py	Refactors TripSchedulingChoiceSettings to inherit from LogitComponentSettings, adds line to sort schedule alternatives

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2025-10-30T16:14:11Z

activitysim/abm/test/test_misc/configs_test_misc/trip_departure_choice.yaml

+#      - start_period
+#      - end_period
+
+SPECIFICATION: trip_departure_choice.csv


The configuration uses the deprecated field name 'SPECIFICATION' instead of 'SPEC'. The TripSchedulingChoiceSettings in trip_scheduling_choice.yaml uses 'SPEC' (the standardized field name), and this should be consistent. Consider updating to 'SPEC' to match the pattern used in LogitComponentSettings.

Suggested change

SPECIFICATION: trip_departure_choice.csv

SPEC: trip_departure_choice.csv

Copilot · 2025-10-30T16:14:12Z

activitysim/abm/test/test_misc/configs_test_misc/trip_departure_choice.yaml

+#COEFFICIENTS: trip_departure_choice_coeff.csv
+
+
+PREPROCESSOR:


The configuration uses the deprecated field name 'PREPROCESSOR' (uppercase) instead of 'preprocessor' (lowercase). The trip_scheduling_choice.yaml file in the same directory uses 'preprocessor', and this inconsistency should be resolved. Consider updating to 'preprocessor' to match the standardized naming convention.

Suggested change

PREPROCESSOR:

preprocessor:

Copilot · 2025-10-30T16:14:12Z

activitysim/abm/test/test_misc/test_trip_scheduling_choice.py

+    A pytest fixture that returns the data folder location.
+    :return: folder location for any necessary data to initialize the tests


The docstring incorrectly states that this fixture 'returns the data folder location', but it actually returns the module name string 'summarize'. The docstring should be updated to accurately describe what the fixture returns and its purpose.

Suggested change

A pytest fixture that returns the data folder location.

:return: folder location for any necessary data to initialize the tests

A pytest fixture that returns the module name string used in the test setup.

:return: module name string ("summarize")

Copilot · 2025-10-30T16:14:13Z

activitysim/abm/test/test_misc/test_trip_departure_choice.py

+    assert len(departures) == len(trips)
+    pd.testing.assert_index_equal(departures.index, trips.index)
+
+    departures = pd.concat([trips, departures], axis=1)


Variable departures is not used.

Suggested change

departures = pd.concat([trips, departures], axis=1)

# Removed unused assignment to departures

Copilot · 2025-10-30T16:14:13Z

activitysim/abm/test/test_misc/test_trip_scheduling_choice.py

+from .setup_utils import setup_dirs
+
+


Import of 'setup_dirs' is not used.

Suggested change

from .setup_utils import setup_dirs

jpn--

Some changes...

jpn-- · 2025-11-01T22:59:13Z

activitysim/abm/models/trip_scheduling_choice.py


    schedules = pd.concat([no_stops, one_way, two_way], sort=True)
    schedules[SCHEDULE_ID] = np.arange(1, schedules.shape[0] + 1)
+    schedules.sort_values(by=["tour_id", SCHEDULE_ID], inplace=True)


This line is the actual "fix" for the non-reproducibility problem... let's put a comment noting that for posterity.

jpn-- · 2025-11-01T23:00:44Z

activitysim/abm/models/trip_scheduling_choice.py


    schedules = pd.concat([no_stops, one_way, two_way], sort=True)
    schedules[SCHEDULE_ID] = np.arange(1, schedules.shape[0] + 1)
+    schedules.sort_values(by=["tour_id", SCHEDULE_ID], inplace=True)


Suggested change

schedules.sort_values(by=["tour_id", SCHEDULE_ID], inplace=True)

# sort by tour id and schedule id to ensure reproducible random results are actually

# stable and reproducible between single and multi-process model runs.

schedules.sort_values(by=["tour_id", SCHEDULE_ID], inplace=True)

jpn-- · 2025-11-01T23:06:05Z

activitysim/abm/test/test_misc/configs_test_misc/trip_departure_choice.yaml

@@ -0,0 +1,20 @@
+#METADATA:


What's "METADATA" for here? Seems like nothing, since it's commented out? We should not add new code/tests with commented out old code.

jpn-- · 2025-11-01T23:06:46Z

activitysim/abm/test/test_misc/configs_test_misc/trip_departure_choice.yaml

+#      - end_period
+
+SPECIFICATION: trip_departure_choice.csv
+#COEFFICIENTS: trip_departure_choice_coeff.csv


jpn-- · 2025-11-01T23:08:26Z

activitysim/abm/test/test_misc/configs_test_misc/trip_scheduling_choice.yaml

@@ -0,0 +1,28 @@
+#METADATA:


should have no extra commented out lines unless it's obvious why

jpn-- · 2025-11-01T23:09:47Z

activitysim/abm/test/test_misc/test_trip_departure_choice.py

+
+@pytest.fixture(scope="module")
+def trips():
+    outbound_array = [True, True, False, False, False, True, True, False, False, True]


Is there a reason why this one single array is written as a variable instead of as a literal in the DF definition below? If so, note why. If not, make it consistent to avoid confusion.

juangacosta and others added 11 commits October 13, 2025 12:41

single process run with

41e9a04

Merge pull request #9 from juangacosta/trip_scheduling_rng_fix

39de183

single process run with

ensured schedule order when generating alternatives

5024ccb

Merge branch 'trip_scheduling_rng_fix' of https://github.com/RSGInc/a…

437aa76

…ctivitysim into trip_scheduling_rng_fix

trip scheduling choice test fix

41edca3

scheduling choice fix with linting

c6b65f8

trip departture choice test fix

75c4024

deleted extra outputs folder

141f876

restoring deleted arc test files

22d662f

deleting extra added files

3246168

more file cleanup

84ac97b

jpn-- mentioned this pull request Oct 23, 2025

2025-10-23 Engineering Team ActivitySim/meeting-notes#37

Open

removing arc from CI test list

f07ec1d

dhensle marked this pull request as ready for review October 30, 2025 03:32

jpn-- requested review from Copilot and jpn-- October 30, 2025 16:11

jpn-- mentioned this pull request Oct 30, 2025

2025-10-30 Engineering Team ActivitySim/meeting-notes#40

Open

Copilot AI reviewed Oct 30, 2025

View reviewed changes

jpn-- requested changes Nov 1, 2025

View reviewed changes

	SPECIFICATION: trip_departure_choice.csv
	SPEC: trip_departure_choice.csv

		A pytest fixture that returns the data folder location.
		:return: folder location for any necessary data to initialize the tests

	departures = pd.concat([trips, departures], axis=1)
	# Removed unused assignment to departures

Trip Scheduling Choice -- Same Results Single Process and Multi-Process #1005

Are you sure you want to change the base?

Trip Scheduling Choice -- Same Results Single Process and Multi-Process #1005

Uh oh!

Conversation

dhensle commented Oct 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Copilot AI Oct 30, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 30, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 30, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 30, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 30, 2025

Choose a reason for hiding this comment

Uh oh!

jpn-- left a comment

Choose a reason for hiding this comment

Uh oh!

jpn-- Nov 1, 2025

Choose a reason for hiding this comment

Uh oh!

jpn-- Nov 1, 2025

Choose a reason for hiding this comment

Uh oh!

jpn-- Nov 1, 2025

Choose a reason for hiding this comment

Uh oh!

jpn-- Nov 1, 2025

Choose a reason for hiding this comment

Uh oh!

jpn-- Nov 1, 2025

Choose a reason for hiding this comment

Uh oh!

jpn-- Nov 1, 2025

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

dhensle commented Oct 21, 2025 •

edited

Loading