PipedSubscriber.__init__() - Code Metrics - UC3MSocialRobots/rospy_utils - Measure and Improve Code Quality continuously with Scrutinizer

PipedSubscriber.init() A
last analyzed 2016-06-03 10:45 UTC

↳ Parent: PipedSubscriber
Complexity

Conditions
Size

Total Lines
Duplication

Lines	0
Ratio	0 %
Importance

Changes	3
Bugs	0	Features	0
Metric	Value
cc	1
c	3
b	0
f	0
dl	0
loc	3
rs	10
# :version:      0.1.0
# :copyright:    Copyright (C) 2014 Universidad Carlos III de Madrid.
#                Todos los derechos reservados.
# :license       LASR_UC3M v1.0, ver LICENCIA.txt

# Este programa es software libre: puede redistribuirlo y/o modificarlo
# bajo los terminos de la Licencia Academica Social Robotics Lab - UC3M
# publicada por la Universidad Carlos III de Madrid, tanto en su version 1.0
# como en una version posterior.

# Este programa se distribuye con la intencion de que sea util,
# pero SIN NINGUNA GARANTIA. Para mas detalles, consulte la
# Licencia Academica Social Robotics Lab - UC3M version 1.0 o posterior.

# Usted ha recibido una copia de la Licencia Academica Social
# Robotics Lab - UC3M en el fichero LICENCIA.txt, que tambien se encuentra
# disponible en <URL a la LASR_UC3Mv1.0>.

"""
Coroutines that ease the data manipulation and communication in ROS.

:author: Victor Gonzalez ()
:maintainer: Victor Gonzalez
:version: 0.2

"""
from __future__ import print_function

import rospy
# .scrutinizer.yml
before_commands:
    - sudo pip install abc # Python2
    - sudo pip3 install abc # Python3
from collections import deque
from functools import wraps

# from decorator import decorator


# @decorator
def coroutine(func):
    """
    Decorator function that takes care of starting a coroutine automatically.

    Example

    >>> @coroutine
    >>> def grep(pattern):
    >>> print "Looking for %s" % pattern
    >>> while True:
    >>>     line = (yield)
    >>>     if pattern in line:
    >>>         print line,
    >>> g = grep("python")
    >>> # Notice how you don't need a next() call here
    >>> g.send("Yeah, but no, but yeah, but no")
    >>> g.send("A series of tubes")
    >>> g.send("python generators rock!")

    Unshamely taken from: http://dabeaz.com/coroutines/coroutine.py
    BTW: If you still havent't checked out his great tutorials on coroutines
    do it right now and you will learn some Python magic tricks, litte bastard:
    http://dabeaz.com/coroutines
    """
    @wraps(func)
    def start(*args, **kwargs):
        """Coroutine wrapper that sets up the coroutine."""
        cr = func(*args, **kwargs)
        next(cr)
        return cr
    return start


###############################################################################
# Filter Coroutines (intermediate steps in the data pipeline)
###############################################################################


@coroutine
def accumulator(num_items, target=None):
    """
    Accumulate items in a list to send it to target when len == num_items.

    It clears the list after it is sent.
    If you do not specify the target, you will have to send it later.


    :param int num_items: (int) Num of items to accumulate before each
    :param target: (default: None) Next coroutine in the data pipeline
        Note that if you don't specify instantiate the coroutine
        specifying the ``target`` you'll have to send it later using
        ``accumulator.send(target)`` method.
    :type target: coroutine or None

    Example

    >>> accum = accumulator(3, printer())
    >>> accum.send(1)     # Nothing sent to printer coroutine
    >>> accum.send(2)     # Still nothing sent
    >>> accum.send(3)
    [1, 2, 3]
    >>> accumf.send(4)
    >>> accumf.send(5)
    >>> accumf.send(6)
    [4, 5, 6]


    Note that you can call ``accumulator`` withouth specifying the target.
    If you don't specify the target, you'll have to send it manually:

    >>> accum = accumulator(3)
    >>> accum.send(printer())     # First you send the target
    >>> accum.send(1)             # And then you can send regular data
    >>> accum.send(2)
    >>> accum.send(3)
    [1, 2, 3]
    """
    items = list()
    if not target:
        target = (yield)
    while True:
        items.append((yield))
        if len(items) == num_items:
            target.send(items)
            del items[:]


@coroutine
def sliding_window(size, target=None):
    """
    Send the last size recived elements to target.

    :param int size: Size of the sliding window
    :param target: (default: None) Next coroutine in the data pipeline
        Note that if you don't specify instantiate the coroutine
        specifying the ``target`` you'll have to send it later using
        ``sliding_window.send(target)`` method.
    :type target: coroutine or None


    Example

    >>> window = sliding_window(3, printer())
    >>> for i in xrange(5):
    >>>     window.send(i)
    [0]
    [0, 1]
    [0, 1, 2]
    [1, 2, 3]
    [2, 3, 4]
    """
    window = deque([], size)
    if not target:
        target = (yield)
    while True:
        window.append((yield))
        target.send(list(window))


@coroutine
def do(f, target=None):
    """
    Apply f on x, send x to the next coroutine.

    Note that ``do`` does not send f(x) but x itself. That means that only
    side effects of ``f`` are relevant.
    This is a good coroutine for logging or publishing messages through
    the pipeline.

    :param callable f: Function with side effects to apply to incoming messages
    :param target: (default: None) Next coroutine in the data pipeline
        Note that if you don't specify instantiate the coroutine
        specifying the ``target`` you'll have to send it later using
        ``do.send(target)`` method.
    :type target: coroutine or None

    Example

    >>> def print_msg(msg):
    >>>     print msg
    >>> t = do(print_msg, transformer(lambda x: x+1, printer()))
    >>> t.send(1)
    1
    2
    >>> t.send(10)
    10
    11
    """
    if not target:
        target = (yield)
    while True:
        msg = (yield)
        f(msg)
        target.send(msg)


@coroutine
def transformer(f, target=None):
    """
    Apply f to incoming data and send the result to target coroutine.

    :param callable f: Function to apply to every incoming message
    :param target: (default: None) Next coroutine in the data pipeline
        Note that if you don't specify instantiate the coroutine
        specifying the ``target`` you'll have to send it later using
        ``transformer.send(target)`` method.
    :type target: coroutine or None

    Example

    >>> t = transformer(lambda x: x+1, printer())
    >>> t.send(1)
    2
    >>> t.send(10)
    11
    """
    if not target:
        target = (yield)
    while True:
        msg = f((yield))
        target.send(msg)


mapper = transformer   # alias


@coroutine
def startransformer(f, target=None, *fargs, **fkwargs):
    """
    Apply f (where f has multiple args) to incoming data. Send result to target.

    This is an extension of the transformer coroutine that uses functions
    that require more than one argument. Note that the first argument of f
    is received by the coroutine, while the remaining args are pre-set in
    fargs and fkwargs.

    :param callable f: Function to apply to every incoming message
    :param target: (default: None) Next coroutine in the data pipeline
        Note that if you don't specify instantiate the coroutine
        specifying the ``target`` you'll have to send it later using
        ``transformer.send(target)`` method.
    :type target: coroutine or None
    :param fargs: Extra arguments to pass to `f`
    :param fkwargs: keyword arguments to pass to `f`

    Example:

    >>> import operator as op
    >>> increment = transformer(op.add, printer(), 1)
    >>> increment.send(1)
    2
    >>> increment.send(2)
    3
    >>> increment.send(3)
    4
    >>> increment.send(10)
    11

    Note that if you want to use the curried version of startransformer
    (E.g. for their use in pipes), you mustset the target coroutine as `None`:

    >>> import operator as op
    >>> increment = transformer(op.add, None, 1)  # Explicitly say "No target".
    >>> increment.send(printer())          # Send the target first.
    >>> increment.send(1)                  # And then you can send regular data.
    2
    >>> increment.send(2)
    3
    >>> increment.send(3)
    4
    """
    if not target:
        target = (yield)
    while True:
        msg = f((yield), *fargs, **fkwargs)
        target.send(msg)


starmapper = startransformer   # alias


@coroutine
def filterer(pred, target=None):
    """
    Coroutine that Filters its messages with pred function.

    :param callable pred: Predicate that evaluates every incoming message
    :param target: (default: None) Next coroutine in the data pipeline
        Note that if you don't specify instantiate the coroutine
        specifying the ``target`` you'll have to send it later using
        ``filterer.send(target)`` method.
    :type target: coroutine or None


    Example

    >>> is_even = lambda x: x % 2 == 0
    >>> evens = filterer(is_even, printer())
    >>> for i in xrange(5):
    >>>     evens.send(i)
    0
    2
    4

    """
    if not target:
        target = (yield)
    while True:
        msg = (yield)
        if pred(msg):
            target.send(msg)


@coroutine
def starfilterer(pred, target=None, *fargs, **fkwargs):
    """
    Multiple argument version of filterer.

    This is an extension of the filterer coroutine which uses predicates
    that require more than one argument. Note that the first argument of f
    is received by the coroutine, while the remaining args are pre-set in
    fargs and fkwargs.

    Coroutine that Filters its messages with pred function.

    :param callable pred: Predicate that evaluates every incoming message
    :param target: (default: None) Next coroutine in the data pipeline
        Note that if you don't specify instantiate the coroutine
        specifying the ``target`` you'll have to send it later using
        ``filterer.send(target)`` method.
    :type target: coroutine or None
    :param fargs: Extra arguments to pass to `f`
    :param fkwargs: keyword arguments to pass to `f`


    Example

    >>> greater_than = lambda x, y: x > y
    >>> greater_than_two = filterer(greater_than, printer(), 2)
    >>> for i in xrange(5):
    >>>     greater_than_two.send(i)
    3
    4

    Note that if you want to use the curried version of starfilterer
    (E.g. for their use in pipes), you mustset the target coroutine as `None`:

    >>> greater_than = lambda x, y: x > y
    >>> greater_than_two = filterer(greater_than, None, 2)
    >>> greater_than_two.send(printer())   # First you set the target coroutine
    >>> for i in xrange(5):      # And then you can send it regular data
    >>>     greater_than_two.send(i)
    3
    4

    """
    if not target:
        target = (yield)
    while True:
        msg = (yield)
        if pred(msg, *fargs, **fkwargs):
            target.send(msg)


@coroutine
def splitter(*coroutines):
    """
    Send the data to all the passed coroutines.

    :param coroutines: coroutines at which the incoming data will be sent

    Example:
    >>> s = splitter(printer(),
                     printer(suffix='!!!'),
                     printer(suffix='World!'))
    >>> s.send('Hello')
    'Hello'
    'Hello!!!'
    'Hello World!'

    If you do not specify the coroutines at coroutine instantiation, you'll
    have to do it later by sending a ``tuple`` or a ``list`` using the
    ``send`` method.

    >>> s = splitter()   # You'll need to specify the coroutines later
    >>> targets = [printer(),
                   printer(suffix='!!!'),
                   printer(suffix='World!')]
    >>> s.send(targets)     # Now you can send regular data.
    >>> s.send('Hello')
    'Hello'
    'Hello!!!'
    'Hello World!'
    """
    if not coroutines:
        coroutines = (yield)
    while True:
        data = (yield)
        for c in coroutines:
            c.send(data)


@coroutine
def either(pred, targets=(None, None)):
    """
    Split an incoming message in two coroutintes according to a predicate.

    The predicate is evaluated against incoming data to decide to which
    coroutine resend the incoming message.
    If the predicate produces a ``True``, then the incoming message will be
    sent to ``targets[0]``. If produces ``False`` it will be sent to
    ``targets[1]``

    :param callable pred: Predicate to decide to which target send the data
    :param targets: A pair of coroutines to send the data.
              If you don't instantiate the coroutine with this param,
              you will need to send the targets afterwards prior to start
              sending it data.
    :type targets: tuple(coroutine, coroutine)

    A possible use of this coroutine is to send data to loggers if some
    preconditions fail:

    Example

    >>> data_processor = transformer(lambda x: x**2, printer())
    >>> error_logger = printer(prefix="ERROR: value too high!")
    >>> ei = either(lambda x: x > 10)
    >>> ei.send((data_processor, error_logger))
    >>> ei.send(2)
    4
    >>> ei.send(12)
    "ERROR: value too high!"
    >>> ei.send(5)
    25
    """
    if not all(targets):
        targets = (yield)
    trues, falses = targets
    while True:
        msg = (yield)
        if pred(msg):
            trues.send(msg)
        else:
            falses.send(msg)


@coroutine
def folder(binop, init_value, target=None):
    """
    The reduce equivalent for coroutines.

    Applies binop to each received value and the previous result
    and sends the result to target

    :param callable binop: binary operation to apply to each received msg
    :param init_value: value used the first time the coroutine receives data
    :param target: (default: None) Next coroutine in the data pipeline
        Note that if you don't specify instantiate the coroutine
        specifying the ``target`` you'll have to send it later using
        ``folder.send(target)`` method.
    :type target: coroutine or None

    Example:

    >>> import operator as op
    >>> mul = fold(op.mul, 1, printer())
    >>> for _ in xrange(5):
            mul.send(2)
    2
    4
    8
    16
    32
    """
    value = init_value
    target = target or (yield)
    # target = kwargs.get('target', (yield))
    # if not target:
    #     target = (yield)
    while True:
        value = binop(value, (yield))
        target.send(value)


@coroutine
def dropwhile(pred, target=None):
    """
    Drop received elements while pred is True.

    :param callable pred: Predicate that evaluates incoming data
    :param target: (default: None) Next coroutine in the data pipeline
        Note that if you don't specify instantiate the coroutine
        specifying the ``target`` you'll have to send it later using
        ``buffer.send(target)`` method.
    :type target: coroutine or None

    Example:

    >>> tw = dropwhile(lambda x: 0 <= x <= 1, printer())
    >>> tw.send(-1)     # Nothing is printed
    >>> tw.send(2)      # Nothing is printed
    >>> tw.send(0.2)
    0.2
    >>> tw.send(42)
    42
    """
    if not target:
        target = (yield)
    while True:
        value = (yield)
        if not pred(value):
            target.send(value)
            break
    while True:
        target.send((yield))


@coroutine
def takewhile(pred, target=None):
    """
    Send elements to target until pred returns False.

    :param callable pred: Predicate that evaluates incoming data
    :param target: (default: None) Next coroutine in the data pipeline
        Note that if you don't specify instantiate the coroutine
        specifying the ``target`` you'll have to send it later using
        ``buffer.send(target)`` method.
    :type target: coroutine or None

    Example:
    --------
    >>> tw = takewhile(lambda x: 0 <= x <= 1, printer())
    >>> tw.send(0.1)
    0.1
    >>> tw.send(0.5)
    0.5
    >>> tw.send(2)      # Nothing is printed anymore
    >>> tw.send(0.2)    # Nothing is printed
    """
    if not target:
        target = ((yield))
    while True:
        val = (yield)
        if pred(val):
            target.send(val)
        else:
            break
    while True:
        (yield)

###############################################################################
# Consumer Coroutines (sinks in the data pipeline)
###############################################################################


@coroutine
def publisher(topic, msg_type, *args, **kwargs):
    """
    A coroutine-based rospy.publisher.

    :topic: (str) Name of the topic to publish
    :msg_type: type of the message to publish

    Example of use:

    >>> from std_msgs import String
    >>> pub = publisher('/my_topic', String)
    >>> pub.send("Hello World!")
    >>> # At this point you would receive "Hello World!" in /my_topic

    See: rospy.publisher
    """
    pub = rospy.Publisher(topic, msg_type, *args, **kwargs)
    while True:
        pub.publish((yield))


@coroutine
def logger(logger_, prefix='', suffix=''):
    """
    Call logger_ on incoming data.

    :param callable logger_: Logger function that prints logging messages
    :param str prefix: (Default: '') Prefix to append to incoming data.
    :param str suffix: (Default: '') Suffix to append to incoming data.


    Example:

    >>> err_logger = logger(rospy.logerr, prefix="ERROR: ", suffix="!!!")
    >>> err_logger.send("This is an error message")
    "ERROR: This is an error message!!!"
    """
    while True:
        try:
            string = str((yield))
            logger_(''.join([prefix, string, suffix]))
        except StopIteration:
            pass


@coroutine
def printer(prefix='', suffix=''):
    """
    Print incoming data.

    :param str prefix: (Default: '') Prefix to append to incoming data.
    :param str suffix: (Default: '') Suffix to append to incoming data.

    Example:

    >>> p = printer()
    >>> p.send("Hello World")
    'Hello World'
    >>> p = printer(prefix='You said: ', suffix='!')
    >>> p.send("Hello World")
    'You said: Hello World"'
    """
    while True:
        try:
            item = (yield)
            print(''.join([prefix, item, suffix]))
        except StopIteration:
            pass


###############################################################################
# Data producers
###############################################################################
class PipedSubscriber(object):

    """
    Subscriber to a ROS topic to send/receive msgs to a coroutine or a pipe.

    A wrapper of the `rospy.Subscriber` class that connects the Subscriber
    directly with a coroutine (or a pipe) that processes the incoming msgs.
    In short it has the same api as `rospy.subscriber` but, instead of a
    callback you pass it a coroutine or a `pipe` to it.

    :param str topic_name: Name of the topic to Subscriber
    :param type msg_type: Type of the messages of `topic_name`
    :param generator target: The coroutine or
        ..mod:pipe where the incoming messages will be sent

    Here you can see an example of use:

    >>> ### This node subscribes to the 'my_topic' topic,
    >>> ### transforms incoming strings to uppercase and
    >>> ### publishes them into rospy.logerr
    >>> import rospy
    >>> pipe = pipe([transformer(lambda x: str(x).upper()),
                     logger(rospy.logwarn, prefix='Got: ')])
    >>> rospy.init_node('my_node')
    >>> rospy.loginfo("Node {} started".format(rospy.get_name()))
    >>> PipedSubscriber('my_topic', String, pipe)
    >>> # rospy.spin()
    """

    def __init__(self, topic_name, msg_type, target, *args, **kwargs):
        """ Class constructor. """
        rospy.Subscriber(topic_name, msg_type, target.send, *args, **kwargs)


###############################################################################
# Utilities
###############################################################################
def pipe(coroutines):
    """
    Chain several coroutines to create a data processing pipe.

    Chains several coroutines together and returns the first coroutine
    of the pipe so you can send messages through the whole pipe.

    Note: The pipe establishes the connections between coroutines,
    therefore, you do not need to establish the targets.

    Params
    :param list coroutines: list of coroutines to pipe
    :return: The first coroutine of the pipe
    :rtype: coroutine

    Example:

    >>> coroutines = (transformer(lambda x: x+1),
                      filterer(lambda x: x%2==0),
                      printer())
    >>> p = pipe(coroutines)
    >>> p.send(1)
    2
    >>> p.send(4)    # No output
    >>> p.send(-1)
    0

    Here you can see a more useful example where we calculate the mean
    of the last 3 received messages and print them on screen only if they
    meet certain conditions:

    Example

    >>> coroutines = [sliding_window(3),
                      transformer(np.mean),
                      filterer(lambda x: 0<= x <= 1),
                      printer(prefix="Result: ")]
    >>> pipe = pipe(coroutines)
    >>> pipe.send(3)    # No output since mean <= 1
    >>> pipe.send(1)    # No output since mean <= 1
    >>> pipe.send(1)    # No output since mean <= 1
    >>> pipe.send(1)
    1.0
    >>> pipe.send(1)
    1.0
    >>> pipe.send(0.1)
    0.7
    >>> pipe.send(0.1)
    0.4
    """
    _coroutines = list(reversed(coroutines))
    pairs = list(zip(_coroutines[:], _coroutines[1:]))
    for p in pairs:
        p[1].send(p[0])
    return coroutines[0]

PipedSubscriber.init() A
last analyzed 2016-06-03 10:45 UTC

Complexity

Size

Duplication

Importance

1. Missing Dependencies

2. Missing init.py files

UC3MSocialRobots / rospy_utils

PipedSubscriber.__init__() A last analyzed 2016-06-03 10:45 UTC

Complexity

Size

Duplication

Importance

1. Missing Dependencies

2. Missing __init__.py files

Duplication Side-by-Side

Filter issues like

PipedSubscriber.init() A
last analyzed 2016-06-03 10:45 UTC

2. Missing init.py files
