Python copy.deepcopy() Examples

def money_trader_action(agent):
    dic1 = copy.deepcopy(agent["goods"])
    ret = seek_a_trade(agent)
    dic2 = copy.deepcopy(agent["goods"])
    diff = {x: (dic1[x][AMT_AVAIL] - dic2[x][AMT_AVAIL])
            for x in dic1 if x in dic2}
    for good in diff:
        # updated due to change in durability calculation
        # decayed_amt = dic1[good][DUR_DECR] * dic1[good][AMT_AVAIL]
        # if (diff[good] != decayed_amt and diff[good] != 0):
        if diff[good] != 0:
            incr_trade_count(good)
    # print("TRADE COUNT")
    # for good in natures_goods:
    #     print(good, " is traded ",
    #           natures_goods[good]["trade_count"], " times")
    # good_decay(agent["goods"])
    return ret 

def OptimizeTrace(self, check_box):
        self.undo_stack.append(deepcopy(self.trace))
        self.last_cb = check_box
        optimization = self.opti_map[check_box.text()]
        if check_box.isChecked():
            self.order.append(optimization)
            self.trace = optimization(self.trace)
        else:
            try:
                self.order.remove(optimization)
            except:
                pass
            self.trace = deepcopy(self.orig_trace)
            for optimization in self.order:
                self.trace = optimization(self.trace)
        self.FoldRegs() 

def test_MutateAST_visit_compare(idx, mut_op, lineno, compare_file, compare_expected_locs):
    """Test mutation of the == to != in the compare op."""
    tree = Genome(compare_file).ast

    # apply the mutation to the original tree copy
    testing_tree = deepcopy(tree)
    mutated_tree = MutateAST(target_idx=compare_expected_locs[idx], mutation=mut_op).visit(
        testing_tree
    )

    # revisit in read-only mode to gather the locations of the new nodes
    mast = MutateAST(readonly=True)
    mast.visit(mutated_tree)

    assert len(mast.locs) == 3

    # check that the lineno marked for mutation is changed, otherwise original ops should
    # still be present without modification
    for loc in mast.locs:
        if loc.lineno == lineno and loc.col_offset == 11:
            assert loc.op_type == mut_op
        else:
            assert loc.op_type in {ast.Eq, ast.Is, ast.In}  # based on compare_file fixture 

def test_MutateAST_visit_index_neg(
    i_order, lineno, col_offset, mut, index_file, index_expected_locs
):
    """Test mutation for Index: i[0], i[1], i[-1]."""
    tree = Genome(index_file).ast
    test_mutation = mut

    testing_tree = deepcopy(tree)
    mutated_tree = MutateAST(target_idx=index_expected_locs[i_order], mutation=test_mutation).visit(
        testing_tree
    )

    mast = MutateAST(readonly=True)
    mast.visit(mutated_tree)

    assert len(mast.locs) == 4

    for loc in mast.locs:
        # spot check on mutation from Index_NumNeg to Index_NumPos
        if loc.lineno == lineno and loc.col_offset == col_offset:
            assert loc.op_type == test_mutation

        # spot check on not-mutated location still being None
        if loc.lineno == 4 and loc.col_offset == 23:
            assert loc.op_type == "Index_NumPos" 

def test_MutateAST_visit_boolop(boolop_file, boolop_expected_loc):
    """Test mutation of AND to OR in the boolop."""
    tree = Genome(boolop_file).ast
    test_mutation = ast.Or

    # apply the mutation to the original tree copy
    testing_tree = deepcopy(tree)
    mutated_tree = MutateAST(target_idx=boolop_expected_loc, mutation=test_mutation).visit(
        testing_tree
    )

    # revisit in read-only mode to gather the locations of the new nodes
    mast = MutateAST(readonly=True)
    mast.visit(mutated_tree)

    # four locations from the binary operations in binop_file
    assert len(mast.locs) == 1

    # there will only be one loc, but this still works
    # basedon the col and line offset in the fixture for compare_expected_loc
    for loc in mast.locs:
        if loc.lineno == 2 and loc.col_offset == 11:
            assert loc.op_type == test_mutation 

def test_MutateAST_visit_nameconst(nameconst_file, nameconst_expected_locs):
    """Test mutation for nameconst: True, False, None."""
    tree = Genome(nameconst_file).ast
    test_mutation = False

    testing_tree = deepcopy(tree)
    mutated_tree = MutateAST(target_idx=nameconst_expected_locs[0], mutation=test_mutation).visit(
        testing_tree
    )

    mast = MutateAST(readonly=True)
    mast.visit(mutated_tree)

    # if statement is included with this file that will be picked up
    nc_locs = [loc for loc in mast.locs if loc.ast_class == "NameConstant"]
    assert len(nc_locs) == 4

    for loc in nc_locs:
        # spot check on mutation from True to False
        if loc.lineno == 1 and loc.col_offset == 14:
            assert loc.op_type == test_mutation

        # spot check on not-mutated location still being None
        if loc.lineno == 7 and loc.col_offset == 22:
            assert loc.op_type is None 

def create_entr(name, i, props=None):
    """
    Create an agent.
    """
    starting_cash = DEF_ENTR_CASH
    if props is not None:
        starting_cash = get_prop('entr_starting_cash',
                                 DEF_ENTR_CASH)

    resources = copy.deepcopy(DEF_CAP_WANTED)
    if props is not None:
        total_resources = get_prop('entr_want_resource_total',
                                   DEF_TOTAL_RESOURCES_ENTR_WANT)
        num_resources = len(resources)
        for k in resources.keys():
            resources[k] = int((total_resources * 2)
                               * (random.random() / num_resources))

    return Agent(name + str(i), action=entr_action,
                 attrs={"cash": starting_cash,
                        "wants": resources,
                        "have": {}}) 

def test_MutateAST_visit_read_only(binop_file):
    """Read only test to ensure locations are aggregated."""
    tree = Genome(binop_file).ast
    mast = MutateAST(readonly=True)
    testing_tree = deepcopy(tree)
    mast.visit(testing_tree)

    # four locations from the binary operations in binop_file
    assert len(mast.locs) == 4

    # tree should be unmodified
    assert ast.dump(tree) == ast.dump(testing_tree)


####################################################################################################
# GENERIC TRANSFORMER NODE TESTS
# These represent the basic pattern for visiting a node in the MutateAST class and applying a
# mutation without running the full test suite against the cached files.
#################################################################################################### 

def test_MutateAST_visit_augassign(augassign_file, augassign_expected_locs):
    """Test mutation for AugAssign: +=, -=, /=, *=."""
    tree = Genome(augassign_file).ast
    test_mutation = "AugAssign_Div"

    testing_tree = deepcopy(tree)
    mutated_tree = MutateAST(target_idx=augassign_expected_locs[0], mutation=test_mutation).visit(
        testing_tree
    )

    mast = MutateAST(readonly=True)
    mast.visit(mutated_tree)

    assert len(mast.locs) == 4

    for loc in mast.locs:
        # spot check on mutation from Add tp Div
        if loc.lineno == 1 and loc.col_offset == 4:
            assert loc.op_type == test_mutation

        # spot check on not-mutated location still being Mult
        if loc.lineno == 5 and loc.col_offset == 4:
            assert loc.op_type == "AugAssign_Mult" 

def __new__(mcs, name, bases, attrs):
        for b in bases:
            if not hasattr(b, "_fields"):
                continue

            for k, v in b.__dict__.items():
                if k in attrs:
                    continue
                if isinstance(v, FieldDescriptor):
                    attrs[k] = copy.deepcopy(v.field)

        mcs = super().__new__(mcs, name, bases, attrs)
        mcs._fields = {}

        for name, field in mcs.__dict__.items():
            if isinstance(field, BaseField):
                field.add_to_class(mcs, name)
        return mcs 

def __init__(self, args: dict={}):
        self.meta = {}
        self.args = deepcopy(args)

        # since input-patterns can be used in many different kinds of io sources, grab it too.
        self.input_patterns = args.get('input-patterns', '**/*')
        if isinstance(self.input_patterns, str):
            self.input_patterns = self.input_patterns.split(",")
        self.samples_dir = args.get('samples-dir', 'samples')
        self.cache_dir = args.get('cache-dir', '.cache')
        self.random_seed = args.get('random-seed', 42)
        self.trainings_dir = args.get('trainings-dir', './trainings')

        self._cached_file_state = None

        spltype, splval = parse_split(args.get('val-split', '10%'))
        self.val_dir = splval if spltype == 'dir' else None
        self.val_num = splval if spltype == 'num' else None
        self.val_perc = splval if spltype == 'perc' else None

        spltype, splval = parse_split(args.get('test-split', '10%'))
        self.test_dir = splval if spltype == 'dir' else None
        self.test_num = splval if spltype == 'num' else None
        self.test_perc = splval if spltype == 'perc' else None 

def OptimizeTrace(self, check_box):
        self.undo_stack.append(deepcopy(self.trace))
        self.last_cb = check_box
        optimization = self.opti_map[check_box.text()]
        if check_box.isChecked():
            self.order.append(optimization)
            self.trace = optimization(self.trace)
        else:
            try:
                self.order.remove(optimization)
            except:
                pass
            self.trace = deepcopy(self.orig_trace)
            for optimization in self.order:
                self.trace = optimization(self.trace)
        self.FoldRegs() 

def turn(self):
        """Turn"""
        r = (4, 5, 5, 5)
        nt = copy.deepcopy(self.table)
        for z in range(self.depth):
            for y in range(self.height):
                for x in range(self.width):
                    neighbours = self.liveNeighbours(z, y, x)
                    if self.table[z][y][x] == 0 and (neighbours > r[0] and neighbours <= r[1]):
                        nt[z][y][x] = 1
                    elif self.table[z][y][x] == 1 and (neighbours > r[2] and neighbours < r[3]):
                        nt[z][y][x] = 0

        self._oldStates.append(self.table)
        if len(self._oldStates) > 3:
            self._oldStates.popleft()

        self.table = nt 

def turn(self):
        """Turn"""
        nt = copy.deepcopy(self.table)
        for y in range(0, self.height):
            for x in range(0, self.width):
                neighbours = self.liveNeighbours(y, x)
                if self.table[y][x] == 0:
                    if neighbours == 3:
                        nt[y][x] = 1
                else:
                    if (neighbours < 2) or (neighbours > 3):
                        nt[y][x] = 0

        self._oldStates.append(self.table)
        if len(self._oldStates) > 3:
            self._oldStates.popleft()

        self.table = nt 

def update_generation_losses(losses, nums, micro, macro, bs, length, loss):
    # Update Losses
    losses[micro] += \
        [copy.deepcopy(losses[micro][-1])]
    losses[macro] += \
        [copy.deepcopy(losses[macro][-1])]

    losses[micro][-1] *= nums[micro]
    losses[macro][-1] *= nums[macro]

    nums[macro] += bs

    if isinstance(length, int):
        update_indiv_generation_losses(
            losses, nums, micro, macro, bs, length, loss)
    else:
        update_tensor_generation_losses(
            losses, nums, micro, macro, bs, length, loss) 

def make_model(d_vocab, N, d_model, d_ff=1024, h=4, dropout=0.1):
    """Helper: Construct a model from hyperparameters."""
    c = copy.deepcopy
    attn = MultiHeadedAttention(h, d_model)
    ff = PositionwiseFeedForward(d_model, d_ff, dropout)
    position = PositionalEncoding(d_model, dropout)
    model = EncoderDecoder(
        Encoder(EncoderLayer(d_model, c(attn), c(ff), dropout), N),
        Decoder(DecoderLayer(d_model, c(attn), c(attn), c(ff), dropout), N),
        nn.GRU(d_model, d_model, 1),
        nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        Generator(d_model, d_vocab),
        d_model
    )
    # This was important from their code.
    # Initialize parameters with Glorot / fan_avg.
    for p in model.parameters():
        if p.dim() > 1:
            nn.init.xavier_uniform_(p)
    return model 

def merge_dicts(child_dict, parent_dict):
        if child_dict is None:
            return parent_dict

        if parent_dict is None:
            return child_dict

        effective_dict = {}

        try:
            # Probably should handle non-string keys
            use_keys = filter(lambda x: ("!" + x) not in child_dict.keys(),
                              parent_dict)

            for k in use_keys:
                effective_dict[k] = deepcopy(parent_dict[k])

            use_keys = filter(lambda x: not x.startswith("!"), child_dict)

            for k in use_keys:
                effective_dict[k] = deepcopy(child_dict[k])
        except TypeError:
            raise TypeError("Error iterating dict argument")

        return effective_dict 

def make_model(d_vocab, N, d_model, latent_size, d_ff=1024, h=4, dropout=0.1):
    """Helper: Construct a model from hyperparameters."""
    c = copy.deepcopy
    attn = MultiHeadedAttention(h, d_model)
    ff = PositionwiseFeedForward(d_model, d_ff, dropout)
    position = PositionalEncoding(d_model, dropout)
    share_embedding = Embeddings(d_model, d_vocab)
    model = EncoderDecoder(
        Encoder(EncoderLayer(d_model, c(attn), c(ff), dropout), N),
        Decoder(DecoderLayer(d_model, c(attn), c(attn), c(ff), dropout), N),
        # nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        # nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        nn.Sequential(share_embedding, c(position)),
        nn.Sequential(share_embedding, c(position)),
        Generator(d_model, d_vocab),
        c(position),
        d_model,
        latent_size,
    )
    # This was important from their code.
    # Initialize parameters with Glorot / fan_avg.
    for p in model.parameters():
        if p.dim() > 1:
            nn.init.xavier_uniform_(p)
    return model 

def make_model(d_vocab, N, d_model, latent_size, d_ff=1024, h=4, dropout=0.1):
    """Helper: Construct a model from hyperparameters."""
    c = copy.deepcopy
    attn = MultiHeadedAttention(h, d_model)
    ff = PositionwiseFeedForward(d_model, d_ff, dropout)
    position = PositionalEncoding(d_model, dropout)
    share_embedding = Embeddings(d_model, d_vocab)
    model = EncoderDecoder(
        Encoder(EncoderLayer(d_model, c(attn), c(ff), dropout), N),
        Decoder(DecoderLayer(d_model, c(attn), c(attn), c(ff), dropout), N),
        # nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        # nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        nn.Sequential(share_embedding, c(position)),
        nn.Sequential(share_embedding, c(position)),
        Generator(d_model, d_vocab),
        c(position),
        d_model,
        latent_size,
    )
    # This was important from their code.
    # Initialize parameters with Glorot / fan_avg.
    for p in model.parameters():
        if p.dim() > 1:
            nn.init.xavier_uniform_(p)
    return model 

def make_model(d_vocab, N, d_model, d_ff=1024, h=4, dropout=0.1):
    """Helper: Construct a model from hyperparameters."""
    c = copy.deepcopy
    attn = MultiHeadedAttention(h, d_model)
    ff = PositionwiseFeedForward(d_model, d_ff, dropout)
    position = PositionalEncoding(d_model, dropout)
    model = EncoderDecoder(
        Encoder(EncoderLayer(d_model, c(attn), c(ff), dropout), N),
        Decoder(DecoderLayer(d_model, c(attn), c(attn), c(ff), dropout), N),
        nn.GRU(d_model, d_model, 1),
        nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        nn.Sequential(Embeddings(d_model, d_vocab), c(position)),
        Generator(d_model, d_vocab),
        d_model
    )
    # This was important from their code.
    # Initialize parameters with Glorot / fan_avg.
    for p in model.parameters():
        if p.dim() > 1:
            nn.init.xavier_uniform_(p)
    return model 

def clones(module, N):
    """Produce N identical layers."""
    return nn.ModuleList([copy.deepcopy(module) for _ in range(N)]) 

def find_output(trace, manual=False, update=None):
    """
    Find output operands to the vm_function.
    :param trace: instruction trace
    :param manual: console output y/n
    :return: set of output operands
    """
    if update is not None:
        update.pbar_update(20)
    ex_trace, vmp_seg_start, vmp_seg_end = extract_vm_segment(deepcopy(trace))
    ex_trace.reverse()
    if update is not None:
        update.pbar_update(20)
    pop_lines = []
    lastline = ''
    ctx = {}
    for line in ex_trace:
        if line.disasm[0].startswith('ret'):
            ctx = line.ctx
            lastline = line
            break
        elif line.disasm[0].startswith('pop'):
            ctx = line.ctx
            lastline = line
            break
    if update is not None:
        update.pbar_update(40)
    if manual:
        print ''.join('%s:%s\n' % (c, ctx[c]) for c in ctx.keys() if get_reg_class(c) is not None)
    return set([ctx[get_reg(reg, trace.ctx_reg_size)].upper() for reg in ctx if get_reg_class(reg) is not None]) 

def to_dict(self):
        """Serializes this instance to a Python dictionary."""
        output = copy.deepcopy(self.__dict__)
        return output 

def clones(module, N):
    """Produce N identical layers."""
    return nn.ModuleList([copy.deepcopy(module) for _ in range(N)]) 

def run(mpfile, nmax=None):
    # print json.dumps(mpfile.document, indent=4)
    print mpfile.document["_hdata"].keys()
    #    datasource = mpfile.document['general'].pop('Datasource')
    datasource = mpfile.document["_hdata"]["general"].pop("input_file")
    subdir = os.path.abspath(
        os.path.join(datasource["work_dir"], datasource["directory"])
    )

    # TODO Potentially we have to insert a preprocessing step, probably in msp
    scandata_f = msp.read_scans(subdir, datacounter="Counter 1")
    scan_groups = scandata_f.groupby(datasource["group_by"].split())
    process_template = mpfile.document["general"].pop("process_template")
    translate = get_translate(datasource["work_dir"])
    keys = scan_groups.groups.keys()
    keys.sort()

    for i, g in enumerate(tqdm(keys, leave=True)):
        # TODO: Group information is saved into the output. Rethink?
        comp, sx, sy = translate(g)
        composition = normalize_root_level(comp)[1]
        process_template_copy = copy.deepcopy(process_template)
        process_template_copy["position"] = {"x": sx, "y": sy}
        mpfile.document.rec_update(
            nest_dict(process_template_copy, [composition, "process_chain"])
        )
        sg = scan_groups.get_group(g)
        for process_chain_name in process_template.keys():
            scan_params = mpfile.document[composition]["process_chain"][
                process_chain_name
            ]
            xmcd_frame = treat_xmcd(sg, scan_params, xas_process.process_dict)
            mpfile.add_data_table(
                composition,
                xmcd_frame[["Energy", "XAS", "XMCD"]],
                "_".join(["data", process_chain_name]),
            )
        if nmax is not None and i > nmax:
            break 

def clones(module, N):
    """Produce N identical layers."""
    return nn.ModuleList([copy.deepcopy(module) for _ in range(N)]) 

def clones(module, N):
    """Produce N identical layers."""
    return nn.ModuleList([copy.deepcopy(module) for _ in range(N)]) 

def ClusterRemoval(self):
        threshold = AskLong(1, 'How many most common clusters do you want removed?')
        self.undo_stack.append(deepcopy(self.trace))
        self.trace = cluster_removal(deepcopy(self.trace), threshold=threshold)
        self.PopulateModel() 

def clones(module, N):
    """Produce N identical layers."""
    return nn.ModuleList([copy.deepcopy(module) for _ in range(N)]) 

def clones(module, N):
    """Produce N identical layers."""
    return nn.ModuleList([copy.deepcopy(module) for _ in range(N)])