rmf_fleet_adapter_python

Python bindings for rmf_fleet_adapter

Introduction

The rmf_fleet_adapter package defines C++ classes that can interact with rmf_core. Specifically an Adapter and a MockAdapter class. This package provides bindings for both and most of their dependencies.

The main difference between rmf_fleet_adapter and rmf_mock_adapter is the fact that rmf_fleet_adapter objects can interact with the ROS2 graph.

The MockAdapter and Adapter both do this. But the difference is that MockAdapter can be instantiated with no issues, whereas Adapter will block to wait for a scheduler node from rmf_core to make its presence known.

Installation

• Place the pybind_ament and rmf_fleet_adapter_python packages in your ROS2 workspace, and run colcon build as per normal.

Running Examples

Integration test script for rmf loop and delivery task:

ros2 run rmf_fleet_adapter_python test_adapter.py

You may then probe the effects on the ROS2 graph by subscribing to the following topics with ros2 topic echo <TOPIC_NAME>:

• : `/dispenser_requests`, `/dispenser_results`, `ingestor_requests`, `ingestor_results`, `/task_summaries`

Traffic Light Example

This will showcase an example of having 2 “traffic light” robots using RMF.

# First Terminal
ros2 run rmf_fleet_adapter_python schedule_blockade_nodes

# Second Terminal
ros2 run rmf_fleet_adapter_python traffic_light

For more details, please read this README.

Notes

• The py api bindings are mainly experimental. Use with caution.
• Current CI and docs gen are using rolling release of RMF
• More bindings, tests, and examples are welcome!

Using the Bindings

Ensure that you have built this package and sourced its environment. Then it is as simple as importing the module containing the Python bindings!

import rmf_adapter as adpt

# You may then check the available bindings
print(dir(adpt))

Description

Fleet adapters allow for interactions between rmf_core and robot fleets.

High level pathing or task assignments can be issued to robots from rmf_core, and a fleet adapter will take those assignments and handle it in an implementation specific manner to direct individual robots to execute certain actions while monitoring and updating rmf_core on their high-level states via rmf_core robot update handles.

For more information it is helpful to look at the docs for rmf_core

Usage

The bound APIs are almost identical to the rmf_core ones as well, so if the written documentation for these packages, and the Python usage examples in the scripts or tests are insufficient, you can also look into the relevant C++ implementations as a last resort.

Alternatively, you may also explore any interfaces with:

# To list any available members or methods
dir(adpt.<WHATEVER_CLASS_YOU_WANT_TO_INSPECT>)

# To see function signatures
help(adpt.<WHATEVER_CLASS_YOU_WANT_TO_INSPECT>)

Binding Details

This package contains Python bindings to C++ code that is found in rmf_core and its associated packages. It was implemented using Pybind11.

Do note that as a result of that, if one needs to inspect the source, there are no Python implementations of any of the bound code, and it would be advisable to look directly into the appropriate rmf_fleet_adapter package, that contains the C++ code.. Additionally, C++ objects will get instantiated (generally) as references that can be used or called using the appropriate bound methods on the Python side.

Some of these bindings allow you to override virtual methods that would normally be called in the C++ side with Python code that you write in your Python scripts!

Features

There are enough functionalities bound to implement your own fleet adapters to communicate with the rest of the rmf_core systems in the ROS2 graph. Allowing you to:

• Adapter/MockAdapter: Communicate with rmf_core by adding fleets or requesting deliveries
• RobotCommandHandle: Specify custom robot command handles that can take commands from the rmf_core packages and implement or execute the following robot behaviors:
  • Docking

File truncated at 100 lines see the full file

Changelog for package rmf_fleet_adapter_python

2.1.8 (2023-12-19)

2.1.7 (2023-08-10)

• Adding initiator and request time to booking (#284)
• Contributors: Aaron Chong

2.1.6 (2023-06-02)

• Switch to rst changelogs (#276)
• Contributors: Yadunund

2.1.5 (2023-05-20)

2.1.4 (2023-04-27)

2.1.3 (2023-04-26)

2.1.2 (2022-10-10)

2.1.0 (2022-10-03)

• Make async behaviors more robust: (#228)
• Allow fleet adapters to change schedule participant profiles: (#229)
• Allow robots to be decommissioned from the task dispatch system: (#233)
• Allow manual toggling of stubborn negotiation: (#196)
• Allow users to specify a custom update listener: (#198)
• Fix various segfaults related to pybind: (#205)
• Allow [ResponsiveWait]{.title-ref} to be enabled and disabled: (#209)
• Allow robot status to be overridden by the user: (#191)
• Add API to report status for `perform_action`: (#190)
• Changes for humble compatibility: (#215)

2.0.0 (2022-03-18)

No changes yet

1.5.0 (2022-02-14)

• Support flexible task definitions (#168)
• Add lane speed limit to graph parsing function (#124)

1.3.0 (2021-06-07)

• Modifications to support refactor of rmf_task (#51)
• Fix symlink-install compilation (#32)
• Updated package.xml (#26)
• Fix/rmf task ros2 cleanup (#21)
• Feature/python binding planner (#11)
• Adding reference_internal tag to function bindings that return raw pointers (#6)
• Feature/add unstable participant api (#11)
• Feature/add simple docs (#9)
• Support apis for task dispatcher (#10)
• differentiate functions to prevent overloading (#8)
• support ez traffic light (#7)
• Update/release 1.1 (#6)
• Implement binding for Duration optional
• Make integration test even stricter
• Add reference capture for posterity
• Add clarifying printouts and fix multi-timer bug
• Integrate compute_plan_starts into integration test
• Implement type tests
• Bind optional constructors and delivery msg interfaces
• Bind compute_plan_starts
• Add update_position overload
• Implement Python Bindings for rmf_fleet_adapter (#1)
• Contributors: Aaron Chong, Charayaphan Nakorn Boon Han, Geoffrey Biggs, Grey, Marco A. Gutiérrez, Yadu, methylDragon, youliang