rust/src/test/bench/shootout-k-nucleotide-pipes.rs

223 lines
6.3 KiB
Rust
Raw Normal View History

// Copyright 2012-2014 The Rust Project Developers. See the COPYRIGHT
// file at the top-level directory of this distribution and at
// http://rust-lang.org/COPYRIGHT.
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.
// ignore-android: FIXME(#10393)
// ignore-pretty the `let to_child` line gets an extra newline
// multi tasking k-nucleotide
extern crate collections;
use collections::HashMap;
use std::mem::replace;
use std::option;
use std::os;
use std::strbuf::StrBuf;
fn f64_cmp(x: f64, y: f64) -> Ordering {
// arbitrarily decide that NaNs are larger than everything.
if y.is_nan() {
Less
} else if x.is_nan() {
Greater
} else if x < y {
Less
} else if x == y {
Equal
} else {
Greater
}
}
// given a map, print a sorted version of it
fn sort_and_fmt(mm: &HashMap<Vec<u8> , uint>, total: uint) -> ~str {
fn pct(xx: uint, yy: uint) -> f64 {
return (xx as f64) * 100.0 / (yy as f64);
}
// sort by key, then by value
fn sortKV(mut orig: Vec<(Vec<u8> ,f64)> ) -> Vec<(Vec<u8> ,f64)> {
orig.sort_by(|&(ref a, _), &(ref b, _)| a.cmp(b));
orig.sort_by(|&(_, a), &(_, b)| f64_cmp(b, a));
orig
}
let mut pairs = Vec::new();
// map -> [(k,%)]
for (key, &val) in mm.iter() {
pairs.push(((*key).clone(), pct(val, total)));
2013-02-02 13:47:41 -06:00
}
let pairs_sorted = sortKV(pairs);
let mut buffer = StrBuf::new();
for &(ref k, v) in pairs_sorted.iter() {
buffer.push_str(format!("{} {:0.3f}\n",
k.as_slice()
.to_ascii()
.to_upper()
.into_str(), v));
}
return buffer.into_owned();
}
// given a map, search for the frequency of a pattern
fn find(mm: &HashMap<Vec<u8> , uint>, key: ~str) -> uint {
let key = key.into_ascii().to_lower().into_str();
match mm.find_equiv(&key.as_bytes()) {
2012-08-20 14:23:37 -05:00
option::None => { return 0u; }
2013-03-23 20:22:00 -05:00
option::Some(&num) => { return num; }
}
}
// given a map, increment the counter for a key
fn update_freq(mm: &mut HashMap<Vec<u8> , uint>, key: &[u8]) {
let key = Vec::from_slice(key);
2013-03-23 20:22:00 -05:00
let newval = match mm.pop(&key) {
Some(v) => v + 1,
None => 1
};
mm.insert(key, newval);
}
// given a Vec<u8>, for each window call a function
// i.e., for "hello" and windows of size four,
// run it("hell") and it("ello"), then return "llo"
fn windows_with_carry(bb: &[u8], nn: uint, it: |window: &[u8]|) -> Vec<u8> {
let mut ii = 0u;
2013-05-14 04:52:12 -05:00
let len = bb.len();
while ii < len - (nn - 1u) {
it(bb.slice(ii, ii+nn));
ii += 1u;
}
return Vec::from_slice(bb.slice(len - (nn - 1u), len));
}
fn make_sequence_processor(sz: uint,
from_parent: &Receiver<Vec<u8>>,
to_parent: &Sender<~str>) {
let mut freqs: HashMap<Vec<u8>, uint> = HashMap::new();
let mut carry = Vec::new();
let mut total: uint = 0u;
let mut line: Vec<u8>;
loop {
line = from_parent.recv();
if line == Vec::new() { break; }
carry = windows_with_carry(carry.append(line.as_slice()).as_slice(),
sz,
|window| {
2013-03-23 20:22:00 -05:00
update_freq(&mut freqs, window);
total += 1u;
});
}
let buffer = match sz {
2013-03-23 20:22:00 -05:00
1u => { sort_and_fmt(&freqs, total) }
2u => { sort_and_fmt(&freqs, total) }
2013-09-30 01:13:47 -05:00
3u => { format!("{}\t{}", find(&freqs, ~"GGT"), "GGT") }
4u => { format!("{}\t{}", find(&freqs, ~"GGTA"), "GGTA") }
6u => { format!("{}\t{}", find(&freqs, ~"GGTATT"), "GGTATT") }
12u => { format!("{}\t{}", find(&freqs, ~"GGTATTTTAATT"), "GGTATTTTAATT") }
18u => { format!("{}\t{}", find(&freqs, ~"GGTATTTTAATTTATAGT"), "GGTATTTTAATTTATAGT") }
2012-08-03 21:59:04 -05:00
_ => { ~"" }
};
2013-02-15 04:44:18 -06:00
to_parent.send(buffer);
}
// given a FASTA file on stdin, process sequence THREE
fn main() {
use std::io::{stdio, MemReader, BufferedReader};
2013-10-06 18:08:56 -05:00
let rdr = if os::getenv("RUST_BENCH").is_some() {
let foo = include_bin!("shootout-k-nucleotide.data");
2014-03-27 00:46:25 -05:00
~MemReader::new(Vec::from_slice(foo)) as ~Reader
2013-10-06 18:08:56 -05:00
} else {
~stdio::stdin() as ~Reader
};
let mut rdr = BufferedReader::new(rdr);
// initialize each sequence sorter
let sizes = vec!(1u,2,3,4,6,12,18);
let mut streams = Vec::from_fn(sizes.len(), |_| Some(channel::<~str>()));
let mut from_child = Vec::new();
let to_child = sizes.iter().zip(streams.mut_iter()).map(|(sz, stream_ref)| {
let sz = *sz;
let stream = replace(stream_ref, None);
let (to_parent_, from_child_) = stream.unwrap();
2013-02-15 04:44:18 -06:00
from_child.push(from_child_);
let (to_child, from_parent) = channel();
2014-01-27 17:29:50 -06:00
spawn(proc() {
2013-04-26 16:04:39 -05:00
make_sequence_processor(sz, &from_parent, &to_parent_);
2014-01-27 17:29:50 -06:00
});
2013-02-15 04:44:18 -06:00
to_child
}).collect::<Vec<Sender<Vec<u8> >> >();
// latch stores true after we've started
// reading the sequence of interest
let mut proc_mode = false;
for line in rdr.lines() {
let line = line.unwrap().trim().to_owned();
2013-10-17 23:08:48 -05:00
if line.len() == 0u { continue; }
match (line[0] as char, proc_mode) {
// start processing if this is the one
('>', false) => {
match line.slice_from(1).find_str("THREE") {
option::Some(_) => { proc_mode = true; }
option::None => { }
}
}
// break our processing
('>', true) => { break; }
// process the sequence for k-mers
(_, true) => {
let line_bytes = line.as_bytes();
for (ii, _sz) in sizes.iter().enumerate() {
let lb = Vec::from_slice(line_bytes);
to_child.get(ii).send(lb);
2013-10-17 23:08:48 -05:00
}
}
// whatever
_ => { }
}
}
// finish...
2013-10-17 23:08:48 -05:00
for (ii, _sz) in sizes.iter().enumerate() {
to_child.get(ii).send(Vec::new());
}
// now fetch and print result messages
2013-10-17 23:08:48 -05:00
for (ii, _sz) in sizes.iter().enumerate() {
println!("{}", from_child.get(ii).recv());
}
}