Enable outputs to access all inputs

[skrawcz]

Is your feature request related to a problem? Please describe.
You cannot currently request inputs that are not referenced in the DAG.

Describe the solution you'd like
Allow one to pass in arbitrary inputs not used in the DAG and then be able to request them as outputs.

To recreate

# functions.py - declare and link your transformations as functions....
import pandas as pd

def a(input: pd.Series) -> pd.Series:
    return input % 7

def b(a: pd.Series) -> pd.Series:
    return a * 2

def c(a: pd.Series, b: pd.Series) -> pd.Series:
    return a * 3 + b * 2

def d(c: pd.Series) -> pd.Series:
    return c ** 3

# And run them!
import functions
from hamilton import driver
dr = driver.Driver({}, functions)
result = dr.execute(
   ['c','d', 'foo'],  # want to request foo -- but it's not referenced anywhere
   inputs={'input': pd.Series([1, 2, 3, 4, 5]), 'foo': 1}
)
print(result)
dr.display_all_functions(
   "graph.dot", orient="TB", show_legend=False)

Traceback (most recent call last):
ValueError: Unknown nodes [foo] requested. Check for typos?

Describe alternatives you've considered
You write a dummy function to contain it as input.

[zilto]

Should the values appear in the viz? Or that isn't part of the requirements?

[elijahbenizzy]

If we want to support this we can just add this to the final result computation (E.G. if it's included in the output we just wire it from the input. Seems decidedly reasonable.

That said, what's the use-case? This should work if these are inputs are referenced, but why not just join later if they're not? Or is it to leverage the result builder?