pulse-api 1.8.0.dev20200625134840

Python API for Pulse Robotic Arm with useful utilities

Pulse Robot Python API

This folder contains Python wrapper for the Pulse Robot REST API. Tested with Python 3. Compatibility with Python 2 is not guaranteed but the underlying API (called pdhttp) supports Python 2.

• Pulse Robot Python API
  • Requirements
  • Installation
  • Software compatibility table
    • Getting started
      • Quickstart
      • API initialization
      • Motion control
      • Controlling accessories and signals
      • Controlling accessories and signals during trajectory execution
      • Tool API
      • Base API
      • Environment API
      • Exception handling
      • Versions API
      • Documentation and further information

Documentation and further information

Requirements

Python 3.5+

Installation

To install from the Python Package Index:

pip3 install pulse-api

or, for a specific version

pip3 install pulse-api==v1.v2.v3

Alternatively, to get the version from our repository, use the following command:

pip3 install pulse-api -i https://pip.rozum.com/simple

To install a specific version:

pip3 install pulse-api==v1.v2.v3 -i https://pip.rozum.com/simple where v1, v2, and v3 (e.g., pulse-api==1.4.3) are version numbers as listed below in the compatibility table.

Note: To install the underlying API (pdhttp), use: pip3 install pdhttp -i https://pip.rozum.com/simple

Software compatibility table

Changelog

Pulse Desk UI version	Python API version
1.4.3	1.4.3
1.4.4	1.4.4
1.5.0, 1.5.1, 1.5.2	1.5.0
1.6.0	1.6.0
1.7.0	1.7.0
1.8.0	1.8.0

Getting started

Examples use the latest version of the library.

Quickstart

WARNING! Before launching this example, make sure that there are no facilities within 0.6 meters around the manipulator.

import math
from pulseapi import (
    RobotPulse,
    pose,
    position,
    PulseApiException,
    MT_LINEAR,
    Session,
)

host = "http://127.0.0.1:8081"  # replace with a valid robot address


# create session in read-write mode
with Session(host, Session.READ_WRITE) as session:
    # create an instance of the API wrapper class using initialized session
    robot = RobotPulse(session)

    # create motion targets
    home_pose = pose([0, -90, 0, -90, -90, 0])
    start_pose = pose([0, -90, 90, -90, -90, 0])
    pose_targets = [
        pose([10, -90, 90, -90, -90, 0]),
        pose([10, -90, 0, -90, -90, 0]),
    ]
    position_target = position([-0.42, -0.12, 0.35], [math.pi, 0, 0])
    position_targets = [
        position([-0.37, -0.12, 0.35], [math.pi, 0, 0]),
        position([-0.42, -0.12, 0.35], [math.pi, 0, 0]),
        position([-0.42, -0.17, 0.35], [math.pi, 0, 0]),
        position([-0.37, -0.17, 0.35], [math.pi, 0, 0]),
    ]

    # set the desired speed (controls both motor velocity and acceleration)
    SPEED = 30
    # set the desired motor velocity
    VELOCITY = 40
    # set the desired motor acceleration
    ACCELERATION = 50
    # set the desired tcp velocity
    TCP_VELOCITY_1CM = 0.01
    TCP_VELOCITY_10CM = 0.1

    while True:
        try:
            robot.set_pose(home_pose, speed=SPEED)
            # checks every 0.1 s whether the motion is finished
            robot.await_stop()

            robot.set_pose(
                start_pose, velocity=VELOCITY, acceleration=ACCELERATION
            )
            robot.await_stop()

            robot.set_position(
                position_target, velocity=VELOCITY, acceleration=ACCELERATION
            )
            robot.await_stop()

            # command the robot to go through multiple position waypoints
            # (execute a trajectory)
            robot.run_positions(position_targets, SPEED)
            robot.await_stop()

            # set the linear motion type
            robot.run_positions(
                position_targets,
                velocity=VELOCITY,
                acceleration=ACCELERATION,
                motion_type=MT_LINEAR,
            )
            robot.await_stop()

            # limit the TCP velocity not to exceed 0.01 m/s (1 cm/s)
            robot.run_positions(
                position_targets,
                tcp_max_velocity=TCP_VELOCITY_1CM,
                motion_type=MT_LINEAR,
            )
            # checks every 0.5 s whether the motion is finished
            robot.await_stop(0.5)

            # limit the TCP velocity not to exceed 0.1 m/s (10 cm/s)
            robot.run_poses(pose_targets, tcp_max_velocity=TCP_VELOCITY_1CM)

        except PulseApiException as e:
            # handle possible errors
            print(
                "Exception {} while calling robot at {} ".format(e, robot.host)
            )
            break

Back to the table of contents

API initialization

In order to control remote access to the robot arm you must use Sessions API (since v1.8.0). The session itself could be initialized in two modes:

• Session.READ_WRITE (default) provides access to all the methods including ones that change state of the robot arm (e.g. motion, changing tool and base, etc.). Note: only one read-write session could exist at a given time on a robot. Attempts to open more than one read-write session will result in error.
• Session.READ_ONLY provides access only to the methods that read state from the robot (e.g current position, tool, base, etc.). There is almost no limit on the read-only sessions count.

Examples, using context manager (recommended way to manage sessions):

from pulseapi import RobotPulse, Session, pose
host = "http://127.0.0.1:8081"  # replace with a valid robot address

target_pose = pose([0, -90, 0, -90, -90, 0]) # create motion target

# open read-write session
with Session(host, Session.READ_WRITE) as session:
    # create an instance of the API wrapper class
    robot = RobotPulse(session)
    # do the necessary actions, for example
    robot.set_pose(target_pose, 100)
    robot.await_stop()
# session is automatically closed here

# open read-only session
with Session(host, Session.READ_ONLY) as session:
    # create an instance of the API wrapper class
    robot = RobotPulse(session)
    # do the necessary actions, for example
    print(robot.get_position())

Examples, using manual session control:

from pulseapi import RobotPulse, Session, pose
host = "http://127.0.0.1:8081"  # replace with a valid robot address

target_pose = pose([0, -90, 0, -90, -90, 0]) # create motion target

# open read-write session
session = Session(host, Session.READ_WRITE)
session.open_session()
# create an instance of the API wrapper class
robot = RobotPulse(session)
# do the necessary actions, for example
robot.set_pose(target_pose, 100)
robot.await_stop()
# close session
session.close_session()

# open read-only session
session = Session(host, Session.READ_ONLY)
session.open_session()
# create an instance of the API wrapper class
robot = RobotPulse(session)
# do the necessary actions, for example
print(robot.get_position())
session.close_session()

Note: if you do not close session, it expires only after 60 seconds.

Back to the table of contents

Motion control

Possible motion targets:

• Positions (set_position, run_positions and get_position methods) - to control the location of the robot's TCP (tool center point). Use the position helper function to create a motion target.
• Poses (set_pose, run_poses and get_pose methods) - to control motor angles. Use the pose helper function to create a motion target.
• Jogging (jogging method) - enter the 'jogging' mode. If the robotic arm already in the 'jogging' mode, use this method to control the direction of the movement. Use jog helper function to create motion target. The motion target has six components ('x', 'y', 'z', 'rx', 'ry', 'rz'). Components are optional with default value equal to 0. Components control accelerations along the corresponding axis relative to the base coordinate system of the robotic arm. To disable the mode, pass jog motion target, where all components are zeros.

Possible motion types:

• Joint (MT_JOINT, default)
• Linear (MT_LINEAR)

Auxiliary methods:

• await_motion - periodically requests robot status (default: every 0.1 s) and waits until the robot finishes movements. Deprecated, use await_stop
• await_stop - periodically requests robot status (default: every 0.1 s) and waits until the robot finishes movements.
• status_motion - returns the actual state of the robotic arm: running (arm in motion), idle (arm not in motion), in the zero gravity mode, or in error state. Deprecated, use status
• status - returns the actual state of the robotic arm - whether it is initializing, or twisted, or running (in motion), or active (not in motion), or in the zero gravity mode, or failed (broken, failed initializing or in emergency).
• freeze - sets the arm in the "freeze" state. The arm stops moving, retaining its last position.
  Note: In the state, it is not advisable to move the arm by hand as this can cause damage to its components.
• relax - sets the arm in the "relaxed" state. The arm stops moving without retaining its last position. In this state, the user can move the robotic arm by hand (e.g., to verify/test a motion trajectory).
• pack - asking the arm to reach a compact pose for transportation.
• status_motors - returns the actual states of the six servo motors integrated into the joints of the robotic arm.

WARNING! The following example is sample code. Before running, you must replace reference motion targets in the sample with the ones applicable to your specific case. Before launching this example, make sure that manipulator would not cause any damage to your facilities.

import math
import time
from pulseapi import (
    position,
    pose,
    RobotPulse,
    MT_LINEAR,
    SystemState,
    jog,
    Session,
)

host = "http://127.0.0.1:8081"  # replace with a valid robot address

with Session(host, Session.READ_WRITE) as session:
    robot = RobotPulse(sesion)

    # create motion targets
    pose_target = pose([0, -90, 90, -90, -90, 0])
    position_target = position([-0.42, -0.12, 0.35], [math.pi, 0, 0])
    position_targets = [
        position([-0.37, -0.12, 0.35], [math.pi, 0, 0]),
        position([-0.42, -0.12, 0.35], [math.pi, 0, 0]),
        position([-0.42, -0.17, 0.35], [math.pi, 0, 0]),
        position([-0.37, -0.17, 0.35], [math.pi, 0, 0]),
    ]
    SPEED = 30  # set the desired speed
    TCP_VELOCITY_1CM = 0.01

    # use the status command as shown below
    def my_await_stop(robot_instance, asking_interval=0.1):
        status = robot_instance.status()
        while status.state == SystemState.MOTION:
            time.sleep(asking_interval)
            status = robot_instance.status()

    robot.set_pose(pose_target, SPEED)
    robot.await_stop()  # checks every 0.1 s whether the motion is finished
    print("Current pose:\n{}".format(robot.get_pose()))

    robot.set_position(position_target, SPEED)
    robot.await_stop(0.5)  # checks every 0.5 s whether the motion is finished
    print("Current position:\n{}".format(robot.get_position()))

    # command the robot to go through multiple position waypoints
    # (execute a trajectory)
    robot.run_positions(position_targets, SPEED)
    my_await_stop(robot)

    # set the linear motion type
    robot.run_positions(position_targets, SPEED, motion_type=MT_LINEAR)
    robot.await_stop()

    # limit the TCP velocity not to exceed 0.01 m/s (1 cm/s)
    robot.run_positions(
        position_targets,
        tcp_max_velocity=TCP_VELOCITY_1CM,
        motion_type=MT_LINEAR,
    )
    robot.await_stop()

    # stop the arm in the last position
    robot.freeze()

    # get status from motors
    print(robot.status_motors())

    # jogging example
    # command the robot to execute preparatory motion targets
    robot.set_pose(pose([0, -90, 0, -90, -90, 0]), SPEED)
    robot.set_position(position([-0.45, -0, 0.33], [math.pi, 0, 0]), SPEED)
    robot.await_stop()
    # start the jogging mode and execute motion targets
    robot.jogging(jog(x=-1, y=-1))
    time.sleep(2)
    robot.jogging(jog(x=1, y=1))
    time.sleep(7)
    robot.jogging(jog(rx=1, rz=-1))
    time.sleep(5)
    robot.jogging(jog(-0.1, -0.8, 0.1, 0, -1, 0.7))
    time.sleep(5)
    # disable the jogging mode
    robot.jogging(jog())

Back to the table of contents

Controlling accessories and signals

Available methods:

• close_gripper, open_gripper with a preset timeout before executing further commands (default: 500 ms). Supported grippers: Schunk and OnRobot.
• set_digital_output_high set_digital_output_low, get_digital_output - to work with output ports on the controlbox.
• get_digital_input to work with input ports on the controlbox.

Signals:

• SIG_LOW - port is inactive
• SIG_HIGH - port is active

from pulseapi import RobotPulse, SIG_LOW, SIG_HIGH, Session

host = "http://127.0.0.1:8081"  # replace with a valid robot address

with Session(host, Session.READ_WRITE) as session:
    robot = RobotPulse(session)

    # ask the robot to close the gripper and continue execution of
    # commands after 500 ms
    robot.close_gripper()

    # ask the robot to open the gripper and begin to execute further
    # commands after 100 ms
    robot.open_gripper(100)

    # set the first output port to the active state
    robot.set_digital_output_high(1)

    # execute required operations when input port 3 is active
    if robot.get_digital_input(3) == SIG_HIGH:
        print("Input port 3 is active")
    # execute required operations when input port 1 is inactive
    if robot.get_digital_input(1) == SIG_LOW:
        print("Input port 1 is inactive")

Back to the table of contents

Controlling accessories and signals during trajectory execution

Use output_action(), open_gripper_action(), close_gripper_action() functions combined with pose() and position() helper functions to control gripper/output signals during trajectory execution.

Note: actions are performed asynchronously.

import math
from pulseapi import (
    RobotPulse,
    SIG_LOW,
    SIG_HIGH,
    output_action,
    position,
    pose,
    open_gripper_action,
    close_gripper_action,
    Session,
)

host = "http://127.0.0.1:8081"  # replace with a valid robot address

with Session(host, Session.READ_WRITE) as session:
    robot = RobotPulse(session)

    # create motion targets with actions

    # ask the robot to set output signal to SIG_LOW value on port 1
    # when it reaches the specified pose
    pose_target = pose([0, -90, 90, -90, -90, 0], [output_action(1, SIG_LOW)])

    # ask the robot to set output signal to SIG_HIGH value on port 1
    # when it reaches the specified position
    position_target = position(
        [-0.42, -0.12, 0.35], [math.pi, 0, 0], [output_action(1, SIG_HIGH)]
    )

    position_targets = [
        # ask the robot to open gripper at the specified position
        position(
            [-0.37, -0.12, 0.35], [math.pi, 0, 0], [close_gripper_action()]
        ),
        position([-0.42, -0.12, 0.35], [math.pi, 0, 0]),
        position([-0.42, -0.17, 0.35], [math.pi, 0, 0]),
        # ask the robot to close gripper at the specified position and
        # to set output signal to SIG_LOW value on port 1 at the specified position
        position(
            [-0.37, -0.17, 0.35],
            [math.pi, 0, 0],
            [open_gripper_action(), output_action(1, SIG_LOW),],
        ),
    ]
    SPEED = 30  # set the desired speed

    robot.set_pose(pose_target, SPEED)
    robot.await_stop()

    robot.set_position(position_target, SPEED)
    robot.await_stop()

    robot.run_positions(position_targets, SPEED)
    robot.await_stop()

Back to the table of contents

Tool API

Use the Tool API methods when you need to calculate a robot motion trajectory with regard to the used tool and to take the tool into account when the robot calculates collisions.

Available methods:

• change_tool_info - set tool info for trajectory calculations.
• change_tool_shape - set tool shape for collision validation.
• get_tool_info, get_tool_shape - receive information about current tool settings.

Helper functions:

• tool_info - creates a tool info instance to be passed into change_tool_info method.
• tool_shape - creates a tool shape instance to be passed into change_tool_shape method.

from pulseapi import RobotPulse, position, Point, Session
from pulseapi import create_simple_capsule_obstacle, tool_shape, tool_info

host = "http://127.0.0.1:8081"  # replace with a valid robot address

with Session(host, Session.READ_WRITE) as session:
    robot = RobotPulse(session)

    # get info about the current tool
    current_tool_info = robot.get_tool_info()
    current_tool_shape = robot.get_tool_shape()
    print("Current tool info\n{}".format(current_tool_info))
    print("Current tool shape\n{}".format(current_tool_shape))

    # create new tool properties
    new_tool_info = tool_info(
        position([0, 0, 0.07], [0, 0, 0]), name="CupHolder"
    )
    new_tool_shape = tool_shape(
        [
            create_simple_capsule_obstacle(
                0.03, Point(0, 0, 0), Point(0, 0, 0.07)
            )
        ]
    )

    # change tool properties
    robot.change_tool_info(new_tool_info)
    robot.change_tool_shape(new_tool_shape)
    print("New tool info\n{}".format(robot.get_tool_info()))
    print("New tool shape\n{}".format(robot.get_tool_shape()))

Back to the table of contents

Base API

Use the Base API methods when you need to calculate a robot motion trajectory relative to a specific point in space.

Available methods:

• change_base
• get_base

from pulseapi import RobotPulse, position, Session

host = "http://127.0.0.1:8081"  # replace with a valid robot address

with Session(host, Session.READ_WRITE) as session:
    robot = RobotPulse(session)

    current_base = robot.get_base()
    print("Current base\n{}".format(current_base))

    # move the new base point along x and y axes
    new_base = position([0.05, 0.05, 0], [0, 0, 0])
    robot.change_base(new_base)

    print("New base\n{}".format(robot.get_base()))

Back to the table of contents

Environment API

Use the Environment API to add virtual obstacles to be taken into account when calculating collisions.

Available methods:

• add_to_environment - adds an obstacle to an environment. Use the helper functions below to describe obstacles.
• get_all_from_environment - returns all obstacles from an environment.
• get_from_environment_by_name - returns an obstacle with a specific name from an environment.
• remove_all_from_environment - removes all obstacles from an environment.
• remove_from_environment_by_name - removes an obstacle with a specific name from an environment.

Helper functions:

• create_box_obstacle
• create_capsule_obstacle
• create_plane_obstacle

from pulseapi import RobotPulse, Point, position, Session
from pulseapi import (
    create_plane_obstacle,
    create_box_obstacle,
    create_capsule_obstacle,
)

host = "http://127.0.0.1:8081"  # replace with a valid robot address

with Session(host, Session.READ_WRITE) as session:
    robot = RobotPulse(session)

    print("Current environment\n{}".format(robot.get_all_from_environment()))
    # add obstacles to the environment for calculating collisions
    box = create_box_obstacle(
        Point(0.1, 0.1, 0.1), position((1, 1, 1), (0, 0, 0)), "example_box"
    )
    capsule = create_capsule_obstacle(
        0.1, Point(0.5, 0.5, 0.2), Point(0.5, 0.5, 0.5), "example_capsule"
    )
    plane = create_plane_obstacle(
        [Point(-0.5, 0.4, 0), Point(-0.5, 0, 0), Point(-0.5, 0, 0.1)],
        "example_plane",
    )
    robot.add_to_environment(box)
    robot.add_to_environment(capsule)
    robot.add_to_environment(plane)
    print("New environment\n{}".format(robot.get_all_from_environment()))
    print(
        "Get example box\n{}".format(
            robot.get_from_environment_by_name(box.name)
        )
    )
    # remove specific obstacles
    robot.remove_from_environment_by_name(box.name)
    print(
        "Environment without box\n{}".format(robot.get_all_from_environment())
    )
    # remove all obstacles from an environment
    robot.remove_all_from_environment()
    print("Empty environment\n{}".format(robot.get_all_from_environment()))

Back to the table of contents

Exception handling

For information about errors, see the API reference. The client wraps errors from the robot into PulseApiException.

Available methods:

• recover - the function recovers the arm after an emergency, setting its motion status to IDLE. Recovery is possible only after an emergency that is not fatal (corresponds to the ERROR status).

For example, we can trigger an API exception by sending pose into set_position method.

from pulseapi import RobotPulse, PulseApiException, pose, SystemState, Session

host = "http://127.0.0.1:8081"  # replace with a valid robot address

with Session(host, Session.READ_WRITE) as session:
    robot = RobotPulse(session)

    try:
        robot.set_position(pose([0, -90, 90, -90, -90, 0]), 10)
        robot.await_stop()
    except PulseApiException as e:
        print("Exception {}while calling robot at {} ".format(e, robot.host))
        status = robot.status()
        if status.state == SystemState.BROKEN:
            robot.recover()
            print(
                "Robot recovered from error. Error message: {}".format(
                    status.message
                )
            )

Back to the table of contents

Versions API

Use the Version API methods to get information about the software and hardware versions. You may need to use the methods for contacting support specialists when you notice strange robot behaviour.

Available methods:

• hardware - returns the hardware versions for all motors, the USB-CAN dongle, safety board and wrist.
• software - returns the software version for all motors, the USB-CAN dongle, safety board and wrist.
• robot_software - returns the version of the robot control software.

from pulseapi import Versions, Session

host = "http://127.0.0.1:8081"  # replace with a valid robot address

with Session(host, Session.READ_ONLY) as session:
    versions = Versions(session)

    print(versions.hardware())
    print(versions.software())
    print(versions.robot_software())

Back to the table of contents

Documentation and further information

For further details, see the API reference guide.