feature: support IUPAC bases - attempt 2 #54
Conversation
| @@ -0,0 +1,571 @@ | |||
| // Copyright 2014-2016 Johannes Köster. | |||
There was a problem hiding this comment.
Pulled directly from rust-bio so we don't have to have it as a full dependency.
| self.len == 0 | ||
| } | ||
|
|
||
| /// Calculate the Hamming distance between this and another bitencoded sequence. |
| } | ||
|
|
||
| #[test] | ||
| fn test_hamming() { |
nh13
force-pushed
the
feature/iupac-support-bio
branch
2 times, most recently
from
4dc2274 to
877f21c
Compare
src/lib/mod.rs
Outdated
| pub mod samples; | ||
|
|
||
| use crate::bitenc::BitEnc; | ||
| use lazy_static::lazy_static; |
There was a problem hiding this comment.
It is no longer necessary to use lazy_static thanks to the addition of std::sync::LazyLock. If this was legacy I'd say just add it as a todo, but since it's new I'd say do it in this PR to avoid introducing a new unnecessary dependency.
| @@ -1,14 +1,97 @@ | |||
| pub mod barcode_matching; | |||
| pub mod bitenc; | |||
There was a problem hiding this comment.
Is there a reason this needs to be exposed externally?
| /// N, it will not match anything but an N, and if the other base is an R, it | ||
| /// will match R, V, D, and N, since the latter IUPAC codes allow both A and G. | ||
| pub fn hamming(&self, other: &BitEnc, max_mismatches: u32) -> u32 { | ||
| assert!(self.len == other.len, "Both bitenc sequences must have the same length"); |
There was a problem hiding this comment.
If this module were internal only I wouldn't care as much about this, but you're exporting it so panicking here might lead to unpleasant surprises. At a minimum, you should document this potential panic. However, I would strongly recommend to instead return a Result with a custom error type.
| /// will match R, V, D, and N, since the latter IUPAC codes allow both A and G. | ||
| pub fn hamming(&self, other: &BitEnc, max_mismatches: u32) -> u32 { | ||
| assert!(self.len == other.len, "Both bitenc sequences must have the same length"); | ||
| assert!(self.width == other.width, "Both bitenc sequences must have the same width"); |
src/lib/bitenc.rs
Outdated
| let values_per_block = self.usable_bits_per_block / self.width; | ||
| for block_index in 0..self.nr_blocks() { | ||
| let intersection = self.storage[block_index] & other.storage[block_index]; | ||
| if intersection != self.storage[block_index] { |
There was a problem hiding this comment.
This seems overly complicated. I think what you want is self.storage[block_index] & !other.storage[block_index].
| assert!(self.width == other.width, "Both bitenc sequences must have the same width"); | ||
| let mut count: u32 = 0; | ||
| let values_per_block = self.usable_bits_per_block / self.width; | ||
| for block_index in 0..self.nr_blocks() { |
There was a problem hiding this comment.
Nit: this could be written using (0..self.nr_blocks()).into_iter().scan(..) to avoid the need for let mut.
nh13
force-pushed
the
feature/iupac-support-bio
branch
from
baefa86 to
eb401f2
Compare
nh13
force-pushed
the
feature/iupac-support-bio
branch
from
eb401f2 to
4c0de2a
Compare
Alternative to #53
I found that
bio-seqwas significantly slower than usingbitencfrombiowhen there were A LOT of barcodes that didn't match. Rather than pulling in all of rust-bio, I copiedbitencfrom rust-bio, then added thehammingmethod (that's slightly faster than our current method) and associated tests.