Allow ability to extend a sweep with a prior default value #5
Conversation
Simplifies the logic in several functions by allowing chaining Nullable
codepaths. Helps keep logic relatively flat instead of nested ifs.
```
if x is None:
if y is None:
# do a
else:
# do b
else:
# do c
```
becomes (something like)
```
outcome = (
Maybe(x)
.or(y)
)
```
| @@ -7,7 +7,7 @@ version_files = ["pyproject.toml"] | |||
|
|
|||
| [tool.ruff.lint] | |||
| select = ['F', 'E', 'W', 'B'] | |||
| ignore = ['E501', 'E701'] | |||
| ignore = ['E501', 'E701', 'B023'] | |||
There was a problem hiding this comment.
This new warning is for unbound variable captures in a for loop. Python really isn't a great language for functional programming.
Something like this:
obj = {}
for i in range(10):
call_later(lambda: obj['a'] = i)depending on when that call_later function runs (e.g. immediately, or some other time in another thread) the result of the loop might be different.
With the Maybe monad, we use exactly this pattern. However, we know that function executions always happen synchronously and are blocking, so this is a safe pattern. So we'll just tell ruff to ignore it.
There was a problem hiding this comment.
This is a Maybe monad. It's an algebraic type, meaning that transformations (e.g. via map and otherwise) are type-preserving --- they also return a Maybe. The transformation may alter the interior type, i.e. Maybe[str] might become a Maybe[int] in the following:
maybe_str = get_user_input()
maybe_int = maybe_str.map(parseInt)Maybe monads are quite helpful when you have complex "nullable" logic where you might go down one of many code paths depending on if any of those code paths return a None.
Ideally, we would refactor a lot of the internal code to use Maybe[T] instead of T | None for consistency and to reduce several if x is None: ... guards.
There was a problem hiding this comment.
Thanks! I will work on adding maybe monad's to other parts
Consider the QRC algorithm implementation where we assume
beta=1.0by default. We execute an experiment like:then later we decide to sweep
beta:A problem occurs where every result is now considered invalidated, despite the fact that we already have results for the
beta=1.0case from the first experiment execution.This PR addresses exactly this issue. The PR is structured in a sequence of 3 commits, it might be useful to go commit-by-commit (i.e. in order to see where the Maybe monad fits in).
You can test this PR with the following (run once to generate the v0 table, uncomment and generate the v1):
We expect the
beta=1.0rows to have ids0, 1, 2, 3and the full range of the v1 table to be[0, 11]. The original implementation would have ids[4, 15].