implement open-ended reasoning #14
Conversation
bddap
changed the title
| } | ||
| } | ||
|
|
||
| /// Attempt to forward translate a quad. |
There was a problem hiding this comment.
would be really useful for future maintainers to have a high level comment of what forward means
(even if it seems redundant from translator.rs)
btw does forward mean find the index of each of s,p,o, and g ?
There was a problem hiding this comment.
Translator is a bi-directional dictionary from one language to another. One language is T and the other is usize.
forall t in T
if trans.forward(t) == Some(u)
then trans.back(u) == Some(t)
btw does forward mean find the index of each of s,p,o, and g ?
If you want to call members of the second language (usize) indices then sure.
The fact that they are indices is an implementation detail. Translator currently looks like this:
pub struct Translator<T> {
widdershins: Box<[T]>,
}But a less efficient version of Translator might look like this:
pub struct Translator<T> {
widdershins: BtreeMap<usize, T>,
deosil: BtreeMap<T, usize>,
}
src/common.rs
Outdated
| ) | ||
| } | ||
|
|
||
| /// Reverse of forward. |
There was a problem hiding this comment.
same as above
and does this mean replace each index with the actual corresponding item ?
There was a problem hiding this comment.
and does this mean replace each index with the actual corresponding item ?
yes? I think? See above comment.
| @@ -19,6 +27,87 @@ where | |||
| ret | |||
| } | |||
|
|
|||
| #[derive(Clone, Debug, PartialEq, Eq, PartialOrd, Ord)] | |||
There was a problem hiding this comment.
I realize I should have pointed that out in the previous PR,
Could you add doc-comments for both highlighting the difference between Rule, LowRule, LowRuleApplication, and what it means to lower()/raise() a rule and why it's used ?
There was a problem hiding this comment.
On it.
Unrelated observation, lower is kinda doing two separate things here. It both changes the way the rule is represented, and it translates elements from T to usize. Maybe lower simplified in the future by separating the two steps. The refactor would be a little tricky though.
There was a problem hiding this comment.
Still need to add docs for LowRuleApplication. Will do.
| @@ -184,4 +186,75 @@ mod tests { | |||
| let expected: [[&str; 4]; 0] = []; | |||
| assert_eq!(&inferences, &expected); | |||
| } | |||
|
|
|||
| #[test] | |||
| fn sum_of_consecutive_ints_is_odd() { | |||
There was a problem hiding this comment.
Thanks for adding this.
|
Thanks for the explanations and comments. |
|
Tagged and pushed changes to https://www.npmjs.com/package/rify and https://crates.io/crates/rify |
This pr adds a new function.
inferinferdoesn't have a goal in mind likeprovedoes.Unlike
validate,inferthinks for itself. It generates its own conclusions.inferis an open minded individual who won't stop until it groks its universe in full.:)