allow slop in both directions #2020
Conversation
allow slop in both directions so "big wolf"~3 can also match "wolf big" This also fixes #1934, when the docsets were reordered by size and didn't match the terms.
| @@ -104,7 +104,8 @@ pub(crate) fn intersection_count(left: &[u32], right: &[u32]) -> usize { | |||
| /// resulting array in left. | |||
| /// | |||
| /// Returns the length of the intersection | |||
| fn intersection(left: &mut [u32], right: &[u32]) -> usize { | |||
| #[inline] | |||
| fn intersection(left: &mut Vec<u32>, right: &[u32]) -> usize { | |||
There was a problem hiding this comment.
Does this need to return count if left is truncated anyway so that the information is "returned" via left.len()?
There was a problem hiding this comment.
No, but it's clear for the function caller
There was a problem hiding this comment.
What is the motivation for this change? If it is to simplify the code, as @adamreichold noted, the new API is really bad.
I'd rather keep the previous logic. If you want to simplify this part, do this in a separate PR, as we will need to make sure that it does not hurt performance.
There was a problem hiding this comment.
It's simpler to use, you get the results in left, instead of left[..returned size]
There was a problem hiding this comment.
agreed. But currently the change is halfway, can we remove the returned value?
Also can we check it improves or does not hurt perf on phrase queries?
There was a problem hiding this comment.
agreed. But currently the change is halfway, can we remove the returned value?
We need to return the count, in case left is not updated
There was a problem hiding this comment.
the results looks good!
There was a problem hiding this comment.
We need to return the count, in case left is not updated
I don't understand.
left.len() is not always the same as the result of the function?
There was a problem hiding this comment.
sorry I thought was about intersection_count_with_slop
returned count is never used for intersection now, I removed it
Codecov Report
📣 This organization is not using Codecov’s GitHub App Integration. We recommend you install it so Codecov can continue to function properly for your repositories. Learn more @@ Coverage Diff @@
## main #2020 +/- ##
==========================================
- Coverage 94.42% 94.38% -0.04%
==========================================
Files 319 319
Lines 59216 59003 -213
==========================================
- Hits 55914 55691 -223
- Misses 3302 3312 +10
|
| if left_val < right_slop { | ||
| left_index += 1; | ||
| } else if right_slop <= left_val && left_val <= right_val { | ||
| let distance = left_val.abs_diff(right_val); |
allow slop in both directions
so "big wolf"~3 can also match "wolf big"
This also fixes #1934, when the docsets were reordered by size and didn't
match the terms.