Пример #1
0
    def __init__(self, **kwargs):
        """
        :Parameters:
            kwargs : dict
                Keyword arguments, see Keywords.
        :Keywords:
            debug : bool
                Debug output toggle.
                Default=False
            externals : list
                A list of SimExternalDelegate instances. Internal state changes
                will be propagated to the delegates.
                Default=False
            sample_rate : float
                Sample rate.
            frame : int
                Frame to start at.
            frame_size : int
                Frame size.
            cfg : str
                Path to a config file, readable by a ConfigParser instance.
        """

        # private property members
        self._cfg = None
        self._externals = []
        self._frame = None
        self._frame_size = None
        self._sample_rate = None
        self._status = None

        # public members
        self.cls_dyn = ClusterDynamics()
        self.io_man = SimIOManager()
        self.neuron_data = NeuronDataContainer()
        self.debug = kwargs.get('debug', False)

        # externals
        for ext in kwargs.get('externals', []):
            if not isinstance(ext, SimExternalDelegate):
                continue
            self._externals.append(ext)
Пример #2
0
    def __init__(self, **kwargs):
        """
        :Parameters:
            kwargs : dict
                Keyword arguments, see Keywords.
        :Keywords:
            debug : bool
                Debug output toggle.
                Default=False
            externals : list
                A list of SimExternalDelegate instances. Internal state changes
                will be propagated to the delegates.
                Default=False
            sample_rate : float
                Sample rate.
            frame : int
                Frame to start at.
            frame_size : int
                Frame size.
            cfg : str
                Path to a config file, readable by a ConfigParser instance.
        """

        # private property members
        self._cfg = None
        self._externals = []
        self._frame = None
        self._frame_size = None
        self._sample_rate = None
        self._status = None

        # public members
        self.cls_dyn = ClusterDynamics()
        self.io_man = SimIOManager()
        self.neuron_data = NeuronDataContainer()
        self.debug = kwargs.get("debug", False)

        # externals
        for ext in kwargs.get("externals", []):
            if not isinstance(ext, SimExternalDelegate):
                continue
            self._externals.append(ext)
Пример #3
0
class BaseSimulation(dict):
    """simulation class controlling the scene"""

    ## constructor

    def __init__(self, **kwargs):
        """
        :Parameters:
            kwargs : dict
                Keyword arguments, see Keywords.
        :Keywords:
            debug : bool
                Debug output toggle.
                Default=False
            externals : list
                A list of SimExternalDelegate instances. Internal state changes
                will be propagated to the delegates.
                Default=False
            sample_rate : float
                Sample rate.
            frame : int
                Frame to start at.
            frame_size : int
                Frame size.
            cfg : str
                Path to a config file, readable by a ConfigParser instance.
        """

        # private property members
        self._cfg = None
        self._externals = []
        self._frame = None
        self._frame_size = None
        self._sample_rate = None
        self._status = None

        # public members
        self.cls_dyn = ClusterDynamics()
        self.io_man = SimIOManager()
        self.neuron_data = NeuronDataContainer()
        self.debug = kwargs.get("debug", False)

        # externals
        for ext in kwargs.get("externals", []):
            if not isinstance(ext, SimExternalDelegate):
                continue
            self._externals.append(ext)

    def initialize(self, **kwargs):
        """initialize the simulation

        :Keywords:
            debug : bool
                Debug flag, enables verbose output.
                Default=False
            frame : long
                Frame id to start at.
                Default=0
            frame_size : int
                The frame size.
                Default=16
            sample_rate : float
                Sample rate to operate.
                Default=16000.0
        """

        self.clear()

        # reset private members
        self.sample_rate = kwargs.get("sample_rate", 16000.0)
        self.frame = kwargs.get("frame", 0)
        self.frame_size = kwargs.get("frame_size", 1024)
        self.status

        # reset pubic members
        self.cls_dyn.clear()
        self.io_man.initialize()
        self.neuron_data.clear()

    def finalize(self):
        """finalize the simulation"""

        self.clear()

        # reset pubic members
        self.cls_dyn.clear()
        self.io_man.finalize()
        self.neuron_data.clear()

    ## properties

    def get_frame(self):
        return self._frame

    def set_frame(self, value):
        self._frame = long(value)
        for ext in self._externals:
            ext.frame(self._frame)

    frame = property(get_frame, set_frame)

    def get_frame_size(self):
        return self._frame_size

    def set_frame_size(self, value):
        self._frame_size = int(value)
        for ext in self._externals:
            ext.frame_size(self._frame_size)
        self.status

    frame_size = property(get_frame_size, set_frame_size)

    def get_sample_rate(self):
        return self._sample_rate

    def set_sample_rate(self, value):
        self._sample_rate = float(value)
        self.cls_dyn.sample_rate = self._sample_rate
        for ext in self._externals:
            ext.sample_rate(self._sample_rate)
        self.status

    sample_rate = property(get_sample_rate, set_sample_rate)

    def get_status(self):
        self._status = {
            "frame_size": self.frame_size,
            "sample_rate": self.sample_rate,
            "neurons": self.neuron_keys,
            "recorders": self.recorder_keys,
        }
        if self.io_man.is_initialized:
            self.io_man.status = self._status
        return self._status

    status = property(get_status)

    def get_neuron_keys(self):
        return [idx for idx in self if isinstance(self[idx], Neuron)]

    neuron_keys = property(get_neuron_keys)

    def get_recorder_keys(self):
        return [idx for idx in self if isinstance(self[idx], Recorder)]

    recorder_keys = property(get_recorder_keys)

    ## simulation control methods

    def simulate(self):
        """advance the simulation by one frame"""

        # inc frame counter
        self.frame += 1

        # process events
        self._simulate_io_tick()

        # process units
        self._simulate_neuron_tick()

        # record for recorders
        self._simulate_recorder_tick()

    def _simulate_io_tick(self):
        """process io loop for the current frame

        This will tick the SimIOManager and process all queued events.
        """

        # get events
        events = self.io_man.tick()

        while len(events) > 0:

            pkg = events.pop(0)
            print pkg

            log_str = ">>> "

            if pkg.ident in self:

                # recorder event
                if isinstance(self[pkg.ident], Recorder):

                    log_str += "R[%s]:" % pkg.ident

                    # position event
                    if pkg.tid == SimPkg.T_POS:

                        # position request
                        if pkg.nitems == 0:
                            log_str += "MOVE[%s] - request" % (self[pkg.ident].name)

                        # reposition request
                        elif pkg.nitems == 1:
                            pos, vel = pkg.cont[0].cont[:2]
                            log_str += "MOVE: %s, %s" % (pos, vel)
                            self[pkg.ident].trajectory_pos = pos
                            # TODO: implementation of the velocity component

                        # weird position event
                        else:
                            print "weird event, was T_POS with:", pkg.cont
                            continue

                        # send position acknowledgement
                        self.io_man.send_package(SimPkg.T_POS, pkg.ident, self._frame, self[pkg.ident].trajectory_pos)

                # neuron event
                elif isinstance(self[pkg.ident], Neuron):
                    log_str += "N:[%s] neuron event" % pkg.ident
                else:
                    log_str += "ANY:[%s] unknown" % pkg.ident

            # other event
            else:

                log_str += "O[%s]:" % pkg.ident
                if pkg.tid == SimPkg.T_CON:
                    log_str += "CONNECT from %s" % str(pkg.cont)
                elif pkg.tid == SimPkg.T_END:
                    log_str += "DISCONNECT from %s" % str(pkg.cont)
                else:
                    log_str += "unknown"
            # log
            self.log(log_str)

    def _simulate_neuron_tick(self):
        """process neurons for current frame

        This will generate spike trains and configure the neuronal firing
        behavior for the current frame.
        """

        # generate spike trains for the scene
        self.cls_dyn.generate(self.frame_size)

        # propagate spike trains to neurons
        for nrn_k in self.neuron_keys:
            self[nrn_k].simulate(frame_size=self.frame_size, firing_times=self.cls_dyn.get_spike_train(nrn_k))

    def _simulate_recorder_tick(self):
        """process recorders for the current frame

        This will record waveforms and grountruth for the current frame.
        """

        # list of all neurons
        nlist = [self[nrn_k] for nrn_k in self.neuron_keys]

        # record per recorder
        for rec_k in self.recorder_keys:
            self.io_man.send_package(
                SimPkg.T_REC, rec_k, self._frame, self[rec_k].simulate(nlist=nlist, frame_size=self.frame_size)
            )

    ## methods logging

    def log(self, log_str):
        """log a string"""

        for ext in self._externals:
            ext.log("[#%09d] %s" % (self.frame, log_str))

    def log_d(self, log_str):
        """log a string"""

        if self.debug is True:
            for ext in self._externals:
                ext.log("[DEBUG] %s" % log_str)

    ## methods object management

    def register_neuron(self, **kwargs):
        """register a neuron to the simulation

        :Keywords:
            neuron_data : NeuronData or path or None
                Reference to a NeuronData object or a path to the HDF5 archive
                where the data can be load from.
                CAUTION! str != QString etc.. use str(neuron_data)!
            position : arraylike
                Position in the scene (x,y,z).
                Default=[0,0,0]
            orientation : arraylike
                Orientation of the object. If ndarray it is interpreted as
                direction of orientation relative to the scene's positive
                z-axis. If its a list or tuple, it is interpreted as a triple of
                euler angles relative to the scene's positive z-axis. If it is
                True a random rotation is created.
                Default=False
            rate_of_fire : float
                Rate of fire in Hz.
                Default=50.0
            amplitude : float
                Amplitude of the waveform.
                Default=1.0
            cluster : int
                The cluster idx
        :Raises:
            some error .. mostly ValueErrors for invalid parameters.
        :Returns:
            The string representation of the registered Neuron.
        """

        # check neuron data
        try:
            neuron_data = kwargs.pop("neuron_data")
        except:
            raise ValueError('No "neuron_data" given!')
        if neuron_data in self.neuron_data:
            ndata = self.neuron_data[neuron_data]
        else:
            if not self.neuron_data.insert(neuron_data):
                raise ValueError("Unknown neuron_data: %s" % str(neuron_data))
            else:
                ndata = self.neuron_data[neuron_data]
        kwargs.update(neuron_data=ndata)

        # build neuron
        neuron = Neuron(**kwargs)
        self[id(neuron)] = neuron

        # register in cluster dynamics
        cls_idx = kwargs.get("cluster", None)
        if cls_idx is not None:
            cls_idx = int(cls_idx)
        self.cls_dyn.add_neuron(neuron, cls_idx=cls_idx)

        # log and return
        self.log(">> %s created!" % neuron)
        self.status
        return str(neuron)

    def register_recorder(self, **kwargs):
        """register a recorder for simulation

        :Keywords:
            position : array like
                Position in scene (x,y,z).
                Default=[0,0,0]
            orientation : arraylike or bool
                Orientation of the object. If ndarray it is interpreted as
                direction of orientation relative to the scene's positive
                z-axis. If its a list or tuple, it is interpreted as a triple of
                euler angles relative to the scene's positive z-axis. If it is
                True a random rotation is created.
                Default=False
            scale : float
                Scale factor for the Tetrahedron height. The default uses values
                from the Thomas Recording GmbH website, with a tetrode of about
                40µm height and 20µm for the basis triangle. Note that negative
                values invert the tetrode direction with respect to
                self.position.
                Default=1.0
            snr : float
                Signal to Noise Ratio (SNR) for the noise process.
                Default=1.0
                TODO: fix this value to be congruent with the neuron amplitudes.
            noise_params : list
                The noise generator parameters.
                Default=None
        :Raises:
            some error ..mostly ValueError for invalid parameters.
        :Returns:
            The string representation of the registered Recorder.
        """

        # build tetrode
        tetrode = Tetrode(**kwargs)
        self[id(tetrode)] = tetrode

        # connect and return
        self.log(">> %s created!" % tetrode)
        self.status
        return str(tetrode)

    def remove_object(self, key):
        """remove the SimObject with objectName 'key'

        :Parameters:
            key : SimObject or int/long
                A SimObject or the id of a SimObject attached (as int/long/str).
        :Returns:
            True on successful removal, False else.
        """

        # convert the key
        if isinstance(key, SimObject):
            lookup = id(key)
        elif isinstance(key, (str, unicode, int, long)):
            try:
                lookup = long(key)
            except:
                try:
                    lookup = long(key, 16)
                except:
                    return False
        else:
            return False

        # remove item
        try:
            item = self.pop(lookup)
            self.log(">> %s destroyed!" % item)
            self.status
            return True
        except:
            return False

    def scene_config_load(self, fname):
        """load a scene configuration file

        :Parameters:
            fname : str
                Path to the scene configuration file.
        """

        # checks and inits
        cfg = ConfigParser()
        cfg_check = cfg.read(fname)
        if fname not in cfg_check:
            raise IOError("could not load scene from %s" % fname)

        # check for config
        ndata_paths = cfg.get("CONFIG", "neuron_data_dir")
        ndata_paths = ndata_paths.strip().split("\n")

        # read per section
        for sec in cfg.sections():

            # check section
            sec_str = sec.split()
            cls = sec_str[0]
            if cls not in ["Neuron", "Tetrode"]:
                continue
            kwargs = {}
            if len(sec_str) > 1:
                kwargs["name"] = " ".join(sec_str[1:])

            # elaborate class and keyword arguments
            bad_ndata = False
            for k, v in cfg.items(sec):

                # read items
                if k is None or v is None or k == "" or v == "":
                    continue
                elif k in ["position", "orientation", "trajectory"]:
                    if v == "False":
                        kwargs[k] = False
                    elif v == "Truse":
                        kwargs[k] = True
                    else:
                        kwargs[k] = map(float, v.split())
                elif k in ["neuron_data"]:
                    kwargs[k] = v
                    ndata_path_list = [osp.join(path, v) for path in ndata_paths]
                    added_ndata = self.neuron_data.insert(ndata_path_list)
                    if added_ndata == 0:
                        bad_ndata = True
                else:
                    kwargs[k] = v

            # delegate action
            if bad_ndata:
                continue
            elif cls == "Neuron":
                self.register_neuron(**kwargs)
            elif cls == "Tetrode":
                self.register_recorder(**kwargs)

    def scene_config_save(self, fname):
        """save the current scene configuration to a file

        :Parameters:
            fname : str
                Path to save tha cfg to.
        """

        # checks and inits
        cfg = ConfigParser()

        # write CONFIG section
        cfg.add_section("CONFIG")
        cfg.set("CONFIG", "frame_size", self.frame_size)
        cfg.set("CONFIG", "sample_rate", self.sample_rate)
        ndata_paths = "\t\n".join(self.neuron_data.paths)
        cfg.set("CONFIG", "neuron_data_dir", ndata_paths)

        # helper functions
        npy2cfg = lambda x: " ".join(map(str, x.tolist()))

        # add per SimObject
        for obj in self.values():

            if isinstance(obj, Neuron):

                name = "Neuron %s" % obj.name
                cfg.add_section(name)
                cfg.set(name, "cluster", str(self.cls_dyn.get_cls_for_nrn(obj)))
                cfg.set(name, "position", npy2cfg(obj.position))
                if obj.orientation is True or obj.orientation is False:
                    cfg.set(name, "orientation", obj.orientation)
                else:
                    cfg.set(name, "orientation", npy2cfg(obj.orientation))
                cfg.set(name, "rate_of_fire", str(obj.rate_of_fire))
                cfg.set(name, "amplitude", str(obj.amplitude))
                cfg.set(name, "neuron_data", str(obj._neuron_data.filename))

            elif isinstance(obj, Recorder):

                name = "Tetrode %s" % obj.name
                cfg.add_section(name)
                cfg.set(name, "position", npy2cfg(obj.position))
                cfg.set(name, "orientation", npy2cfg(obj._trajectory))
                cfg.set(name, "snr", str(obj.snr))

        # save
        save_file = open(fname, "w")
        cfg.write(save_file)
        save_file.close()

    ## methods config loading

    def load_config(self, cfg_path=None):
        """loads initialization values from file

        :Parameters:
            cfg_file : str
                Path to the config file. Should be readable by a ConfigParser.
        """

        # TODO: implement config file handling

    ## special methods

    def __len__(self):
        return len(filter(lambda x: isinstance(x, Neuron), self.values()))

    def __str__(self):
        return "BaseSimulation :: %d items" % len(self)
Пример #4
0
class BaseSimulation(dict):
    """simulation class controlling the scene"""

    ## constructor

    def __init__(self, **kwargs):
        """
        :Parameters:
            kwargs : dict
                Keyword arguments, see Keywords.
        :Keywords:
            debug : bool
                Debug output toggle.
                Default=False
            externals : list
                A list of SimExternalDelegate instances. Internal state changes
                will be propagated to the delegates.
                Default=False
            sample_rate : float
                Sample rate.
            frame : int
                Frame to start at.
            frame_size : int
                Frame size.
            cfg : str
                Path to a config file, readable by a ConfigParser instance.
        """

        # private property members
        self._cfg = None
        self._externals = []
        self._frame = None
        self._frame_size = None
        self._sample_rate = None
        self._status = None

        # public members
        self.cls_dyn = ClusterDynamics()
        self.io_man = SimIOManager()
        self.neuron_data = NeuronDataContainer()
        self.debug = kwargs.get('debug', False)

        # externals
        for ext in kwargs.get('externals', []):
            if not isinstance(ext, SimExternalDelegate):
                continue
            self._externals.append(ext)

    def initialize(self, **kwargs):
        """initialize the simulation

        :Keywords:
            debug : bool
                Debug flag, enables verbose output.
                Default=False
            frame : long
                Frame id to start at.
                Default=0
            frame_size : int
                The frame size.
                Default=16
            sample_rate : float
                Sample rate to operate.
                Default=16000.0
        """

        self.clear()

        # reset private members
        self.sample_rate = kwargs.get('sample_rate', 16000.0)
        self.frame = kwargs.get('frame', 0)
        self.frame_size = kwargs.get('frame_size', 1024)
        self.status

        # reset pubic members
        self.cls_dyn.clear()
        self.io_man.initialize()
        self.neuron_data.clear()

    def finalize(self):
        """finalize the simulation"""

        self.clear()

        # reset pubic members
        self.cls_dyn.clear()
        self.io_man.finalize()
        self.neuron_data.clear()

    ## properties

    def get_frame(self):
        return self._frame

    def set_frame(self, value):
        self._frame = long(value)
        for ext in self._externals:
            ext.frame(self._frame)

    frame = property(get_frame, set_frame)

    def get_frame_size(self):
        return self._frame_size

    def set_frame_size(self, value):
        self._frame_size = int(value)
        for ext in self._externals:
            ext.frame_size(self._frame_size)
        self.status

    frame_size = property(get_frame_size, set_frame_size)

    def get_sample_rate(self):
        return self._sample_rate

    def set_sample_rate(self, value):
        self._sample_rate = float(value)
        self.cls_dyn.sample_rate = self._sample_rate
        for ext in self._externals:
            ext.sample_rate(self._sample_rate)
        self.status

    sample_rate = property(get_sample_rate, set_sample_rate)

    def get_status(self):
        self._status = {
            'frame_size': self.frame_size,
            'sample_rate': self.sample_rate,
            'neurons': self.neuron_keys,
            'recorders': self.recorder_keys,
        }
        if self.io_man.is_initialized:
            self.io_man.status = self._status
        return self._status

    status = property(get_status)

    def get_neuron_keys(self):
        return [idx for idx in self if isinstance(self[idx], Neuron)]

    neuron_keys = property(get_neuron_keys)

    def get_recorder_keys(self):
        return [idx for idx in self if isinstance(self[idx], Recorder)]

    recorder_keys = property(get_recorder_keys)

    ## simulation control methods

    def simulate(self):
        """advance the simulation by one frame"""

        # inc frame counter
        self.frame += 1

        # process events
        self._simulate_io_tick()

        # process units
        self._simulate_neuron_tick()

        # record for recorders
        self._simulate_recorder_tick()

    def _simulate_io_tick(self):
        """process io loop for the current frame

        This will tick the SimIOManager and process all queued events.
        """

        # get events
        events = self.io_man.tick()

        while len(events) > 0:

            pkg = events.pop(0)
            print pkg

            log_str = '>>> '

            if pkg.ident in self:

                # recorder event
                if isinstance(self[pkg.ident], Recorder):

                    log_str += 'R[%s]:' % pkg.ident

                    # position event
                    if pkg.tid == SimPkg.T_POS:

                        # position request
                        if pkg.nitems == 0:
                            log_str += 'MOVE[%s] - request' % (
                                self[pkg.ident].name)

                        # reposition request
                        elif pkg.nitems == 1:
                            pos, vel = pkg.cont[0].cont[:2]
                            log_str += 'MOVE: %s, %s' % (pos, vel)
                            self[pkg.ident].trajectory_pos = pos
                            # TODO: implementation of the velocity component

                        # weird position event
                        else:
                            print 'weird event, was T_POS with:', pkg.cont
                            continue

                        # send position acknowledgement
                        self.io_man.send_package(
                            SimPkg.T_POS, pkg.ident, self._frame,
                            self[pkg.ident].trajectory_pos)

                # neuron event
                elif isinstance(self[pkg.ident], Neuron):
                    log_str += 'N:[%s] neuron event' % pkg.ident
                else:
                    log_str += 'ANY:[%s] unknown' % pkg.ident

            # other event
            else:

                log_str += 'O[%s]:' % pkg.ident
                if pkg.tid == SimPkg.T_CON:
                    log_str += 'CONNECT from %s' % str(pkg.cont)
                elif pkg.tid == SimPkg.T_END:
                    log_str += 'DISCONNECT from %s' % str(pkg.cont)
                else:
                    log_str += 'unknown'
            # log
            self.log(log_str)

    def _simulate_neuron_tick(self):
        """process neurons for current frame

        This will generate spike trains and configure the neuronal firing
        behavior for the current frame.
        """

        # generate spike trains for the scene
        self.cls_dyn.generate(self.frame_size)

        # propagate spike trains to neurons
        for nrn_k in self.neuron_keys:
            self[nrn_k].simulate(
                frame_size=self.frame_size,
                firing_times=self.cls_dyn.get_spike_train(nrn_k))

    def _simulate_recorder_tick(self):
        """process recorders for the current frame

        This will record waveforms and grountruth for the current frame.
        """

        # list of all neurons
        nlist = [self[nrn_k] for nrn_k in self.neuron_keys]

        # record per recorder
        for rec_k in self.recorder_keys:
            self.io_man.send_package(
                SimPkg.T_REC, rec_k, self._frame,
                self[rec_k].simulate(nlist=nlist, frame_size=self.frame_size))

    ## methods logging

    def log(self, log_str):
        """log a string"""

        for ext in self._externals:
            ext.log('[#%09d] %s' % (self.frame, log_str))

    def log_d(self, log_str):
        """log a string"""

        if self.debug is True:
            for ext in self._externals:
                ext.log('[DEBUG] %s' % log_str)

    ## methods object management

    def register_neuron(self, **kwargs):
        """register a neuron to the simulation

        :Keywords:
            neuron_data : NeuronData or path or None
                Reference to a NeuronData object or a path to the HDF5 archive
                where the data can be load from.
                CAUTION! str != QString etc.. use str(neuron_data)!
            position : arraylike
                Position in the scene (x,y,z).
                Default=[0,0,0]
            orientation : arraylike
                Orientation of the object. If ndarray it is interpreted as
                direction of orientation relative to the scene's positive
                z-axis. If its a list or tuple, it is interpreted as a triple of
                euler angles relative to the scene's positive z-axis. If it is
                True a random rotation is created.
                Default=False
            rate_of_fire : float
                Rate of fire in Hz.
                Default=50.0
            amplitude : float
                Amplitude of the waveform.
                Default=1.0
            cluster : int
                The cluster idx
        :Raises:
            some error .. mostly ValueErrors for invalid parameters.
        :Returns:
            The string representation of the registered Neuron.
        """

        # check neuron data
        try:
            neuron_data = kwargs.pop('neuron_data')
        except:
            raise ValueError('No "neuron_data" given!')
        if neuron_data in self.neuron_data:
            ndata = self.neuron_data[neuron_data]
        else:
            if not self.neuron_data.insert(neuron_data):
                raise ValueError('Unknown neuron_data: %s' % str(neuron_data))
            else:
                ndata = self.neuron_data[neuron_data]
        kwargs.update(neuron_data=ndata)

        # build neuron
        neuron = Neuron(**kwargs)
        self[id(neuron)] = neuron

        # register in cluster dynamics
        cls_idx = kwargs.get('cluster', None)
        if cls_idx is not None:
            cls_idx = int(cls_idx)
        self.cls_dyn.add_neuron(neuron, cls_idx=cls_idx)

        # log and return
        self.log('>> %s created!' % neuron)
        self.status
        return str(neuron)

    def register_recorder(self, **kwargs):
        """register a recorder for simulation

        :Keywords:
            position : array like
                Position in scene (x,y,z).
                Default=[0,0,0]
            orientation : arraylike or bool
                Orientation of the object. If ndarray it is interpreted as
                direction of orientation relative to the scene's positive
                z-axis. If its a list or tuple, it is interpreted as a triple of
                euler angles relative to the scene's positive z-axis. If it is
                True a random rotation is created.
                Default=False
            scale : float
                Scale factor for the Tetrahedron height. The default uses values
                from the Thomas Recording GmbH website, with a tetrode of about
                40µm height and 20µm for the basis triangle. Note that negative
                values invert the tetrode direction with respect to
                self.position.
                Default=1.0
            snr : float
                Signal to Noise Ratio (SNR) for the noise process.
                Default=1.0
                TODO: fix this value to be congruent with the neuron amplitudes.
            noise_params : list
                The noise generator parameters.
                Default=None
        :Raises:
            some error ..mostly ValueError for invalid parameters.
        :Returns:
            The string representation of the registered Recorder.
        """

        # build tetrode
        tetrode = Tetrode(**kwargs)
        self[id(tetrode)] = tetrode

        # connect and return
        self.log('>> %s created!' % tetrode)
        self.status
        return str(tetrode)

    def remove_object(self, key):
        """remove the SimObject with objectName 'key'

        :Parameters:
            key : SimObject or int/long
                A SimObject or the id of a SimObject attached (as int/long/str).
        :Returns:
            True on successful removal, False else.
        """

        # convert the key
        if isinstance(key, SimObject):
            lookup = id(key)
        elif isinstance(key, (str, unicode, int, long)):
            try:
                lookup = long(key)
            except:
                try:
                    lookup = long(key, 16)
                except:
                    return False
        else:
            return False

        # remove item
        try:
            item = self.pop(lookup)
            self.log('>> %s destroyed!' % item)
            self.status
            return True
        except:
            return False

    def scene_config_load(self, fname):
        """load a scene configuration file

        :Parameters:
            fname : str
                Path to the scene configuration file.
        """

        # checks and inits
        cfg = ConfigParser()
        cfg_check = cfg.read(fname)
        if fname not in cfg_check:
            raise IOError('could not load scene from %s' % fname)

        # check for config
        ndata_paths = cfg.get('CONFIG', 'neuron_data_dir')
        ndata_paths = ndata_paths.strip().split('\n')

        # read per section
        for sec in cfg.sections():

            # check section
            sec_str = sec.split()
            cls = sec_str[0]
            if cls not in ['Neuron', 'Tetrode']:
                continue
            kwargs = {}
            if len(sec_str) > 1:
                kwargs['name'] = ' '.join(sec_str[1:])

            # elaborate class and keyword arguments
            bad_ndata = False
            for k, v in cfg.items(sec):

                # read items
                if k is None or v is None or k == '' or v == '':
                    continue
                elif k in ['position', 'orientation', 'trajectory']:
                    if v == 'False':
                        kwargs[k] = False
                    elif v == 'Truse':
                        kwargs[k] = True
                    else:
                        kwargs[k] = map(float, v.split())
                elif k in ['neuron_data']:
                    kwargs[k] = v
                    ndata_path_list = [
                        osp.join(path, v) for path in ndata_paths
                    ]
                    added_ndata = self.neuron_data.insert(ndata_path_list)
                    if added_ndata == 0:
                        bad_ndata = True
                else:
                    kwargs[k] = v

            # delegate action
            if bad_ndata:
                continue
            elif cls == 'Neuron':
                self.register_neuron(**kwargs)
            elif cls == 'Tetrode':
                self.register_recorder(**kwargs)

    def scene_config_save(self, fname):
        """save the current scene configuration to a file

        :Parameters:
            fname : str
                Path to save tha cfg to.
        """

        # checks and inits
        cfg = ConfigParser()

        # write CONFIG section
        cfg.add_section('CONFIG')
        cfg.set('CONFIG', 'frame_size', self.frame_size)
        cfg.set('CONFIG', 'sample_rate', self.sample_rate)
        ndata_paths = '\t\n'.join(self.neuron_data.paths)
        cfg.set('CONFIG', 'neuron_data_dir', ndata_paths)

        # helper functions
        npy2cfg = lambda x: ' '.join(map(str, x.tolist()))

        # add per SimObject
        for obj in self.values():

            if isinstance(obj, Neuron):

                name = 'Neuron %s' % obj.name
                cfg.add_section(name)
                cfg.set(name, 'cluster',
                        str(self.cls_dyn.get_cls_for_nrn(obj)))
                cfg.set(name, 'position', npy2cfg(obj.position))
                if obj.orientation is True or obj.orientation is False:
                    cfg.set(name, 'orientation', obj.orientation)
                else:
                    cfg.set(name, 'orientation', npy2cfg(obj.orientation))
                cfg.set(name, 'rate_of_fire', str(obj.rate_of_fire))
                cfg.set(name, 'amplitude', str(obj.amplitude))
                cfg.set(name, 'neuron_data', str(obj._neuron_data.filename))

            elif isinstance(obj, Recorder):

                name = 'Tetrode %s' % obj.name
                cfg.add_section(name)
                cfg.set(name, 'position', npy2cfg(obj.position))
                cfg.set(name, 'orientation', npy2cfg(obj._trajectory))
                cfg.set(name, 'snr', str(obj.snr))

        # save
        save_file = open(fname, 'w')
        cfg.write(save_file)
        save_file.close()

    ## methods config loading

    def load_config(self, cfg_path=None):
        """loads initialization values from file

        :Parameters:
            cfg_file : str
                Path to the config file. Should be readable by a ConfigParser.
        """

        # TODO: implement config file handling

    ## special methods

    def __len__(self):
        return len(filter(lambda x: isinstance(x, Neuron), self.values()))

    def __str__(self):
        return 'BaseSimulation :: %d items' % len(self)