Python copy.copy方法代碼示例

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def __copy__(self):
        """
        Shallow copy of a WrapLinker.

        Returns
        -------
        object
            A copy of self, where each of the linkers in self.linkers
            have been shallow-copied.

        It is useful because in FunctionMaker, copy.copy is called on the
        Mode's linker, so that it is not modified inplace when linker.accept()
        is called. In this case, we want the wrapped linkers to be copied too.

        """
        other = self.__class__(
            linkers=[copy(l) for l in self.linkers],
            wrapper=self.wrapper)
        return other 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def clone(self, allow_gc=undef):
        new = copy(self)
        if allow_gc is not undef:
            new.allow_gc = allow_gc
        return new 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def test_tracking_step(lattice, parametr, update_ref_values=False):
    """Tracking step function test
    :parametr=0 - tracking with MethodTM() - params[Undulator] = UndulatorTestTM
    :parametr=1 - tracking with default MethodTM()
    """
    
    p = Particle(x=0.001, y=0.002)
    p.E = 2.5

    navi = Navigator(lattice)
    dz = 0.01

    P1 = []
    for iii in range(int(lattice[parametr].totalLen/dz)):
        tracking_step(lattice[parametr], [p], dz=dz, navi=navi)
        P1.append(copy.copy(p))

    tracking_step(lattice[parametr], p, dz=dz, navi=navi)
    
    P1 = obj2dict(P1)
    
    if update_ref_values:
        return P1

    p_ref = json_read(REF_RES_DIR + sys._getframe().f_code.co_name + str(parametr) +'.json')

    #assert check_dict(P1, p_ref, TOL)
    result = check_dict(P1, p_ref, TOL, assert_info=' P1 - ')
    assert check_result(result) 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def construct( s ):
      @update
      def upblk():
        u = copy(42, 10) 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def changesettings(self, case_id, *args, **kw ):
        import copy
        
        newsettings = copy.copy(self.settings)
        for name,val in kw.iteritems():
            newsettings.set( name, val )
        storepath = self.db.get_case_storepath(case_id)
        cfgpath = os.path.join( storepath, Includes.CASE_SETTINGSFILENAME )
        newsettings.set_storefile(cfgpath)
        newsettings.save()
        self.settings = newsettings
        return json.dumps({"success":1}) 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def copy(self):
        """
        Construct a new IntervalTree using shallow copies of the
        intervals in the source tree.

        Completes in O(n*log n) time.
        :rtype: IntervalTree
        """
        return IntervalTree(iv.copy() for iv in self) 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def iou3d(corners_3d_b1, corners_3d_b2, vol):

    corners_3d_b1 = copy.copy(corners_3d_b1)
    corners_3d_b2 = copy.copy(corners_3d_b2)

    corners_3d_b1 = corners_3d_b1.T
    corners_3d_b2 = corners_3d_b2.T

    y_min_b1 = np.min(corners_3d_b1[:, 1])
    y_max_b1 = np.max(corners_3d_b1[:, 1])
    y_min_b2 = np.min(corners_3d_b2[:, 1])
    y_max_b2 = np.max(corners_3d_b2[:, 1])
    y_intersect = np.max([0, np.min([y_max_b1, y_max_b2]) - np.max([y_min_b1, y_min_b2])])

    # set Z as Y
    corners_3d_b1[:, 1] = corners_3d_b1[:, 2]
    corners_3d_b2[:, 1] = corners_3d_b2[:, 2]

    polygon_order = [7, 2, 3, 6, 7]
    box_b1_bev = Polygon([list(corners_3d_b1[i][0:2]) for i in polygon_order])
    box_b2_bev = Polygon([list(corners_3d_b2[i][0:2]) for i in polygon_order])

    intersect_bev = box_b2_bev.intersection(box_b1_bev).area
    intersect_3d = y_intersect * intersect_bev

    iou_bev = intersect_bev / (box_b2_bev.area + box_b1_bev.area - intersect_bev)
    iou_3d = intersect_3d / (vol - intersect_3d)

    return iou_bev, iou_3d 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def load_weights(model, path, remove_module=False):
    """
    Simply loads a pytorch models weights from a given path.
    """
    dst_weights = model.state_dict()
    src_weights = torch.load(path)

    dst_keys = list(dst_weights.keys())
    src_keys = list(src_weights.keys())

    if remove_module:

        # copy keys without module
        for key in src_keys:
            src_weights[key.replace('module.', '')] = src_weights[key]
            del src_weights[key]
        src_keys = list(src_weights.keys())

        # remove keys not in dst
        for key in src_keys:
            if key not in dst_keys: del src_weights[key]

    else:

        # remove keys not in dst
        for key in src_keys:
            if key not in dst_keys: del src_weights[key]

        # add keys not in src
        for key in dst_keys:
            if key not in src_keys: src_weights[key] = dst_weights[key]

    model.load_state_dict(src_weights) 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def create_surface(self, vertices, simplices, fn, alpha=.8):
        """
        Method to create the polydata that define the surfaces

        Args:
            vertices (numpy.array): 2D array (XYZ) with the coordinates of the points
            simplices (numpy.array): 2D array with the value of the vertices that form every single triangle
            fn (int): id
            alpha (float): Opacity

        Returns:
            vtk.vtkActor, vtk.vtkPolyDataMapper, vtk.vtkPolyData
        """
        vertices_c = copy.deepcopy(vertices)
        simplices_c = copy.deepcopy(simplices)

        surf_polydata = vtk.vtkPolyData()

        surf_polydata.SetPoints(self.create_surface_points(vertices_c))
        surf_polydata.SetPolys(self.create_surface_triangles(simplices_c))
        surf_polydata.Modified()

        surf_mapper = vtk.vtkPolyDataMapper()
        surf_mapper.SetInputData(surf_polydata)
        surf_mapper.Update()

        surf_actor = vtk.vtkActor()
        surf_actor.SetMapper(surf_mapper)
        surf_actor.GetProperty().SetColor(mcolors.hex2color(self.geo_model._surfaces.df.set_index('id')['color'][fn]))#self.C_LOT[fn])
        surf_actor.GetProperty().SetOpacity(alpha)

        return surf_actor, surf_mapper, surf_polydata 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def cut_finite_fault_surfaces(geo_model, ver:dict, sim:dict):
    """Cut vertices and simplices for finite fault surfaces to finite fault ellipse

    Args:
        geo_model (gempy.core.model.Project): gempy geo_model object
        ver (dict): Dictionary with surfaces as keys and vertices ndarray as values.
        sim (dict): Dictionary with surfaces as keys and simplices ndarray as values.

    Returns:
        ver, sim (dict, dict): Updated vertices and simplices with finite fault
            surfaces cut to ellipses.
    """
    from scipy.spatial import Delaunay
    from copy import copy

    finite_ver = copy(ver)
    finite_sim = copy(sim)

    finite_fault_series = list(geo_model._faults.df[geo_model._faults.df["isFinite"] == True].index)
    finite_fault_surfaces = list(
        geo_model._surfaces.df[geo_model._surfaces.df._stack == finite_fault_series].surface.unique())

    for fault in finite_fault_surfaces:
        U, fpoints_rot, fctr_rot, a, b = get_fault_rotation_objects(geo_model, "Fault 1")
        rpoints = np.dot(ver[fault], U[0])
        # rpoints = np.dot(rpoints, U[-1])
        r = (rpoints[:, 0] - fctr_rot[0]) ** 2 / a ** 2 + (rpoints[:, 1] - fctr_rot[1]) ** 2 / b ** 2

        finite_ver[fault] = finite_ver[fault][r < 1]
        delaunay = Delaunay(finite_ver[fault])
        finite_sim[fault] = delaunay.simplices
        # finite_sim[fault] = finite_sim[fault][np.isin(sim[fault], np.argwhere(r<0.33))]

    return finite_ver, finite_sim 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def make_thunk(self, **kwargs):
        no_recycling = self.no_recycling

        make_all = [self.linkers[0].make_all(**kwargs)]
        kwargs.pop('input_storage', None)
        make_all += [l.make_all(**kwargs) for l in self.linkers[1:]]

        fns, input_lists, output_lists, thunk_lists, order_lists \
            = zip(*make_all)

        order_list0 = order_lists[0]
        for order_list in order_lists[1:]:
            if not order_list0 == order_list:
                raise Exception(
                    "All linkers to WrapLinker should execute operations in the same order.")

        inputs0 = input_lists[0]
        outputs0 = output_lists[0]

        thunk_groups = list(zip(*thunk_lists))
        order = [x[0] for x in zip(*order_lists)]

        to_reset = []
        for thunks, node in izip(thunk_groups, order):
            for j, output in enumerate(node.outputs):
                if output in no_recycling:
                    for thunk in thunks:
                        to_reset.append(thunk.outputs[j])

        wrapper = self.wrapper
        pre = self.pre

        def f():
            for inputs in input_lists[1:]:
                for input1, input2 in izip(inputs0, inputs):
                    input2.storage[0] = copy(input1.storage[0])
            for x in to_reset:
                x[0] = None
            pre(self, [input.data for input in input_lists[0]],
                order, thunk_groups)
            for i, (thunks, node) in enumerate(izip(thunk_groups, order)):
                try:
                    wrapper(i, node, *thunks)
                except Exception:
                    raise_with_op(node, *thunks)
        f.thunk_groups = thunk_groups

        return f, inputs0, outputs0 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def onDepth(self, ws_data):
        """
        5檔深度行情數據回報
        :param ws_data:
        :return:
        """
        channel = ws_data.get('channel')
        data = ws_data.get('data', {})
        # 檢查channel/data
        if channel is None and len(data) == 0:
            return

        symbol = self.channelSymbolMap.get(channel)
        if symbol == None:
            return

        # 更新tick
        if symbol not in self.tickDict:
            tick = VtTickData()
            tick.symbol = symbol
            tick.vtSymbol = symbol
            tick.gatewayName = self.gatewayName
            self.tickDict[symbol] = tick
        else:
            tick = self.tickDict[symbol]

        tick.bidPrice1, tick.bidVolume1 = data['bids'][0]
        tick.bidPrice2, tick.bidVolume2 = data['bids'][1]
        tick.bidPrice3, tick.bidVolume3 = data['bids'][2]
        tick.bidPrice4, tick.bidVolume4 = data['bids'][3]
        tick.bidPrice5, tick.bidVolume5 = data['bids'][4]

        tick.askPrice1, tick.askVolume1 = data['asks'][-1]
        tick.askPrice2, tick.askVolume2 = data['asks'][-2]
        tick.askPrice3, tick.askVolume3 = data['asks'][-3]
        tick.askPrice4, tick.askVolume4 = data['asks'][-4]
        tick.askPrice5, tick.askVolume5 = data['asks'][-5]

        tick.bidPrice1, tick.bidVolume1 = float(tick.bidPrice1), float(tick.bidVolume1)
        tick.bidPrice2, tick.bidVolume2 = float(tick.bidPrice2), float(tick.bidVolume2)
        tick.bidPrice3, tick.bidVolume3 = float(tick.bidPrice3), float(tick.bidVolume3)
        tick.bidPrice4, tick.bidVolume4 = float(tick.bidPrice4), float(tick.bidVolume4)
        tick.bidPrice5, tick.bidVolume5 = float(tick.bidPrice5), float(tick.bidVolume5)
        tick.askPrice1, tick.askVolume1 = float(tick.askPrice1), float(tick.askVolume1)
        tick.askPrice2, tick.askVolume2 = float(tick.askPrice2), float(tick.askVolume2)
        tick.askPrice3, tick.askVolume3 = float(tick.askPrice3), float(tick.askVolume3)
        tick.askPrice4, tick.askVolume4 = float(tick.askPrice4), float(tick.askVolume4)
        tick.askPrice5, tick.askVolume5 = float(tick.askPrice5), float(tick.askVolume5)

        tick.date, tick.time,tick.datetime = self.generateDateTime(data['timestamp'])
        # print "Depth", tick.date, tick.time

        # 推送tick事件
        newtick = copy(tick)
        self.gateway.onTick(newtick) 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def merge_equals(self, data_reducer=None, data_initializer=None):
        """
        Finds all intervals with equal ranges and merges them
        into a single interval. If provided, uses data_reducer and
        data_initializer with similar semantics to Python's built-in
        reduce(reducer_func[, initializer]), as follows:

        If data_reducer is set to a function, combines the data
        fields of the Intervals with
            current_reduced_data = data_reducer(current_reduced_data, new_data)
        If data_reducer is None, the merged Interval's data
        field will be set to None, ignoring all the data fields
        of the merged Intervals.

        On encountering the first Interval to merge, if
        data_initializer is None (default), uses the first
        Interval's data field as the first value for
        current_reduced_data. If data_initializer is not None,
        current_reduced_data is set to a shallow copy of
        data_initiazer created with
            copy.copy(data_initializer).

        Completes in O(n*logn).
        """
        if not self:
            return

        sorted_intervals = sorted(self.all_intervals)  # get sorted intervals
        merged = []
        # use mutable object to allow new_series() to modify it
        current_reduced = [None]
        higher = None  # iterating variable, which new_series() needs access to

        def new_series():
            if data_initializer is None:
                current_reduced[0] = higher.data
                merged.append(higher)
                return
            else:  # data_initializer is not None
                current_reduced[0] = copy(data_initializer)
                current_reduced[0] = data_reducer(current_reduced[0], higher.data)
                merged.append(Interval(higher.begin, higher.end, current_reduced[0]))

        for higher in sorted_intervals:
            if merged:  # series already begun
                lower = merged[-1]
                if higher.range_matches(lower):  # should merge
                    upper_bound = max(lower.end, higher.end)
                    if data_reducer is not None:
                        current_reduced[0] = data_reducer(current_reduced[0], higher.data)
                    else:  # annihilate the data, since we don't know how to merge it
                        current_reduced[0] = None
                    merged[-1] = Interval(lower.begin, upper_bound, current_reduced[0])
                else:
                    new_series()
            else:  # not merged; is first of Intervals to merge
                new_series()

        self.__init__(merged) 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def init_training_model(conf, cache_folder, conf_name):
    """
    This function is meant to load the training model and optimizer, which expects
    ./model/<conf.model>.py to be the pytorch model file.

    The function copies the model file into the cache BEFORE loading, for easy reproducibility.
    """

    src_path = os.path.join('.', 'models', conf.model + '.py')
    dst_path = os.path.join(cache_folder, conf.model + '.py')

    # (re-) copy the pytorch model file
    if os.path.exists(dst_path): os.remove(dst_path)
    shutil.copyfile(src_path, dst_path)
    shutil.copyfile(os.path.join('.', 'scripts', 'config', conf_name + '.py'), os.path.join(cache_folder, conf_name + '.py'))

    # load and build
    network = absolute_import(dst_path)
    network = network.build(conf, 'train')

    # multi-gpu
    network = torch.nn.DataParallel(network)

    # load SGD
    if conf.solver_type.lower() == 'sgd':

        lr = conf.lr
        mo = conf.momentum
        wd = conf.weight_decay

        optimizer = torch.optim.SGD(network.parameters(), lr=lr, momentum=mo, weight_decay=wd)

    # load adam
    elif conf.solver_type.lower() == 'adam':

        lr = conf.lr
        wd = conf.weight_decay

        optimizer = torch.optim.Adam(network.parameters(), lr=lr, weight_decay=wd)

    # load adamax
    elif conf.solver_type.lower() == 'adamax':

        lr = conf.lr
        wd = conf.weight_decay

        optimizer = torch.optim.Adamax(network.parameters(), lr=lr, weight_decay=wd)

    scheduler = None
    if 'cosine' in conf.lr_policy:
        scheduler = torch.optim.lr_scheduler.CosineAnnealingWarmRestarts(optimizer, conf.restart_iters, T_mult=1)  # 5000 iter restart
    elif 'onecycle' in conf.lr_policy:
        scheduler = OneCycleLR(optimizer, max_lr=lr, total_steps=conf.max_iter)
    return network, optimizer, scheduler 

# 需要導入模塊: from copy import copy [as 別名]
# 或者: from copy.copy import copy [as 別名]
def init_training_model(conf, cache_folder):
    """
    This function is meant to load the training model and optimizer, which expects
    ./model/<conf.model>.py to be the pytorch model file.

    The function copies the model file into the cache BEFORE loading, for easy reproducibility.
    """

    src_path = os.path.join('.', 'models', conf.model + '.py')
    dst_path = os.path.join(cache_folder, conf.model + '.py')

    # (re-) copy the pytorch model file
    if os.path.exists(dst_path): os.remove(dst_path)
    shutil.copyfile(src_path, dst_path)

    # load and build
    network = absolute_import(dst_path)
    network = network.build(conf, 'train')

    # multi-gpu
    network = torch.nn.DataParallel(network)

    # load SGD
    if conf.solver_type.lower() == 'sgd':

        lr = conf.lr
        mo = conf.momentum
        wd = conf.weight_decay

        optimizer = torch.optim.SGD(network.parameters(), lr=lr, momentum=mo, weight_decay=wd)

    # load adam
    elif conf.solver_type.lower() == 'adam':

        lr = conf.lr
        wd = conf.weight_decay

        optimizer = torch.optim.Adam(network.parameters(), lr=lr, weight_decay=wd)

    # load adamax
    elif conf.solver_type.lower() == 'adamax':

        lr = conf.lr
        wd = conf.weight_decay

        optimizer = torch.optim.Adamax(network.parameters(), lr=lr, weight_decay=wd)


    return network, optimizer