Port Mapping

stringdale edges have a port mapping language that allows us to map subobjects of the output of one node, into the inputs of the functions of the other nodes.

In python, a function’s arguments are supplied as follows f(*args,**kwargs) where:

args is a list of positional arguments
kwargs is a dictionary that maps from the name of the keyword argument to its value
positional arguments can also be called by their name

In stringdale, we encode both positional and keyword arguments as a dictionary whose keys are:

Contiguous integers in case of positional arguments
strings in case of keyword arguments
each such key is called a port of the function

A port mapping is a sequence of terms of the form <port>=<access_pattern> Where an access pattern defines which part of the output to use.

Here are some examples to illustrate this:

from stringdale import V,E,Define

Let us make a nested object so we can see the power of port mapping in action.

Access patterns try to access object attributes and then object keys. Lets see both:

class TestObj:
    def __init__(self):
        self.a = 10
        self.b = {'x': 20}
        self.c = {'y': 30}
    def __str__(self):
        return f'TestObj(a={self.a},b={self.b},c={self.c})'
    def __repr__(self):
        return self.__str__()

nested_object = {
    'a': 1,
    'b': 2,
    'c':{
        'a': 3,
        'b': 4
    },
    'd':TestObj(),
}

Now lets define some simple diagrams, try to guess what the input of each node will be

with Define('Key Access') as D:
    E('Start->End(x=a,y=b)')

D.draw()

print(nested_object)
D().run_all(nested_object)

{'a': 1, 'b': 2, 'c': {'a': 3, 'b': 4}, 'd': TestObj(a=10,b={'x': 20},c={'y': 30})}

{'x': 1, 'y': 2}

with Define('Nested Key Access') as D:
    E('Start->End(x=d.b,y=d.b.x)')
D.draw()

print(nested_object)
D().run_all(nested_object)

{'a': 1, 'b': 2, 'c': {'a': 3, 'b': 4}, 'd': TestObj(a=10,b={'x': 20},c={'y': 30})}

{'x': {'x': 20}, 'y': 20}

with Define('Whole Object Access') as D:
    E('Start->End(a=.)')
D.draw()

print(nested_object)
D().run_all(nested_object)

{'a': 1, 'b': 2, 'c': {'a': 3, 'b': 4}, 'd': TestObj(a=10,b={'x': 20},c={'y': 30})}

{'a': {'a': 1,
  'b': 2,
  'c': {'a': 3, 'b': 4},
  'd': TestObj(a=10,b={'x': 20},c={'y': 30})}}

with Define('Positional Access') as D:
    E('Start->End(0=a,1=b,x=c)')
D.draw()

print(nested_object)
D().run_all(nested_object)

{'a': 1, 'b': 2, 'c': {'a': 3, 'b': 4}, 'd': TestObj(a=10,b={'x': 20},c={'y': 30})}

{0: 1, 1: 2, 'x': {'a': 3, 'b': 4}}

with Define('Nested Positional Access') as D:
    E('Start->End(0=c.a,1=d.c)')
D.draw()

print(nested_object)
D().run_all(nested_object)

{'a': 1, 'b': 2, 'c': {'a': 3, 'b': 4}, 'd': TestObj(a=10,b={'x': 20},c={'y': 30})}

{0: 3, 1: {'y': 30}}

with Define('Splat Access') as D:
    E('Start->End(**=.,x=a)')
D.draw()

print(nested_object)
D().run_all(nested_object)

{'a': 1, 'b': 2, 'c': {'a': 3, 'b': 4}, 'd': TestObj(a=10,b={'x': 20},c={'y': 30})}

{'a': 1,
 'b': 2,
 'c': {'a': 3, 'b': 4},
 'd': TestObj(a=10,b={'x': 20},c={'y': 30}),
 'x': 1}

with Define('Nested Splat Access') as D:
    E('Start->End(**=c)') # this only works for dict likes, and not for arbitrary objects
D.draw()

print(nested_object)
D().run_all(nested_object)

{'a': 1, 'b': 2, 'c': {'a': 3, 'b': 4}, 'd': TestObj(a=10,b={'x': 20},c={'y': 30})}

{'a': 3, 'b': 4}

with Define('Default Access') as D:
    E('Start->End') # this is identical to 'Start->End(0=.)'
D.draw()


print(nested_object)
# note that if we got a single object, we return obj, rather than {0:obj}
D().run_all(nested_object)

{'a': 1, 'b': 2, 'c': {'a': 3, 'b': 4}, 'd': TestObj(a=10,b={'x': 20},c={'y': 30})}

{'a': 1,
 'b': 2,
 'c': {'a': 3, 'b': 4},
 'd': TestObj(a=10,b={'x': 20},c={'y': 30})}

with Define('Default Access with other keys') as D:
    E('Start->End(0=., x=a)')
D.draw()


print(nested_object)
D().run_all(nested_object)

{'a': 1, 'b': 2, 'c': {'a': 3, 'b': 4}, 'd': TestObj(a=10,b={'x': 20},c={'y': 30})}

{0: {'a': 1,
  'b': 2,
  'c': {'a': 3, 'b': 4},
  'd': TestObj(a=10,b={'x': 20},c={'y': 30})},
 'x': 1}

with Define('ignore_key') as D:
    # this is used to ignore the output of a previous node,
    E('Start->End(_=.)') 
D.draw()

print(nested_object)
D().run_all(nested_object)

{'a': 1, 'b': 2, 'c': {'a': 3, 'b': 4}, 'd': TestObj(a=10,b={'x': 20},c={'y': 30})}

{}