def get_dataflow():
df = DataFlow()
vid1 = df.add_vertex()
pid10 = df.add_in_port(vid1, "in")
pid11 = df.add_out_port(vid1, "out")
vid2 = df.add_vertex()
pid21 = df.add_out_port(vid2, "out")
vid3 = df.add_vertex()
pid31 = df.add_in_port(vid3, "in1")
pid32 = df.add_in_port(vid3, "in2")
pid33 = df.add_out_port(vid3, "res")
vid4 = df.add_vertex()
pid41 = df.add_in_port(vid4, "in")
pid42 = df.add_out_port(vid4, "out")
df.connect(pid11, pid31)
df.connect(pid21, pid32)
df.connect(pid33, pid41)
df.set_actor(vid1, FuncNode({}, {}, int))
df.set_actor(vid2, FuncNode({}, {}, fixed_function))
df.set_actor(vid3, FuncNode({}, {}, operator.add))
df.set_actor(vid4, FuncNode({}, {}, print))
return df, (pid10, pid42)
def test_dataflow_evaluation_no_input_two_outputs():
df = DataFlow()
vid = df.add_vertex()
pid0 = df.add_out_port(vid, "out1")
df.set_actor(vid, FuncNode({}, {}, double_fixed_function))
dfs = DataflowState(df)
env = 0
algo = BruteEvaluation(df)
algo.eval(env, dfs, vid)
assert tuple(dfs.get_data(pid0)) == (5, 5)
algo.clear()
dfs.reinit()
pid1 = df.add_out_port(vid, "out2")
algo.eval(env, dfs, vid)
assert dfs.get_data(pid0) == 5
assert dfs.get_data(pid1) == 5
algo.clear()
dfs.reinit()
pid2 = df.add_out_port(vid, "out3")
assert_raises(UserWarning, lambda: algo.eval(env, dfs, vid))
def __init__(self, auto_caption=None, *args, **kargs):
"""
:Input:
same input as FuncNode:
`inputs`, `outputs` and `func`
auto_caption:
- if None, no automatic update of node caption when called
- Otherwise:
- 'all': automatically add all inputs to caption when called
- a number 'n': add only the nth inputs to caption
"""
FuncNode.__init__(self, *args, **kargs)
if hasattr(self.func, 'set_wrapper_node'):
self.func.set_wrapper_node(self)
if auto_caption is True: auto_caption = slice(None)
self._auto_caption = auto_caption
def __init__(self, auto_caption=None, *args, **kargs):
"""
:Input:
same input as FuncNode:
`inputs`, `outputs` and `func`
auto_caption:
- if None, no automatic update of node caption when called
- Otherwise:
- 'all': automatically add all inputs to caption when called
- a number 'n': add only the nth inputs to caption
"""
FuncNode.__init__(self, *args,**kargs)
if hasattr(self.func,'set_wrapper_node'):
self.func.set_wrapper_node(self)
if auto_caption is True: auto_caption = slice(None)
self._auto_caption = auto_caption
def test_dataflow_evaluation_single_input_no_output():
df = DataFlow()
vid = df.add_vertex()
pid0 = df.add_in_port(vid, "in")
df.set_actor(vid, FuncNode({}, {}, print))
dfs = DataflowState(df)
env = 0
algo = BruteEvaluation(df)
dfs.set_data(pid0, 1)
algo.eval(env, dfs, vid)
assert dfs.get_data(pid0) == 1
df.set_actor(vid, FuncNode({}, {}, int))
dfs.reinit()
algo.clear()
assert_raises(UserWarning, lambda: algo.eval(env, dfs, vid))
def test_dataflow_evaluation_eval_no_vid2():
df = DataFlow()
vid0 = df.add_vertex()
pid0 = df.add_in_port(vid0, "in")
df.add_out_port(vid0, "out")
vid1 = df.add_vertex()
pid1 = df.add_in_port(vid1, "in")
df.add_out_port(vid1, "out")
df.set_actor(vid0, FuncNode({}, {}, int))
df.set_actor(vid1, FuncNode({}, {}, int))
dfs = DataflowState(df)
env = 0
algo = BruteEvaluation(df)
dfs.set_data(pid0, 0)
dfs.set_data(pid1, 1)
assert not dfs.is_valid()
algo.eval(env, dfs)
assert dfs.is_valid()
def test_dataflow_evaluation_single_input_single_output():
df = DataFlow()
vid = df.add_vertex()
pid0 = df.add_in_port(vid, "in")
pid1 = df.add_in_port(vid, "in")
pid2 = df.add_out_port(vid, "out")
df.set_actor(vid, FuncNode({}, {}, operator.add))
dfs = DataflowState(df)
env = 0
algo = BruteEvaluation(df)
dfs.set_data(pid0, 1)
dfs.set_data(pid1, 2)
algo.eval(env, dfs, vid)
assert dfs.get_data(pid0) == 1
assert dfs.get_data(pid1) == 2
assert dfs.get_data(pid2) == 3