2012-12-10 19:32:48 -06:00
|
|
|
// Copyright 2012 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.
|
|
|
|
|
2012-11-15 16:52:38 -06:00
|
|
|
// xfail-pretty (extra blank line is inserted in vec::mapi call)
|
2012-07-06 17:15:52 -05:00
|
|
|
// multi tasking k-nucleotide
|
|
|
|
|
2012-09-11 19:46:20 -05:00
|
|
|
extern mod std;
|
2012-09-05 14:32:05 -05:00
|
|
|
use std::sort;
|
2013-04-03 08:28:36 -05:00
|
|
|
use core::hashmap::HashMap;
|
2013-03-01 17:55:31 -06:00
|
|
|
use core::io::ReaderUtil;
|
|
|
|
use core::comm::{stream, Port, Chan};
|
|
|
|
use core::cmp::Ord;
|
2012-07-06 17:15:52 -05:00
|
|
|
|
|
|
|
// given a map, print a sorted version of it
|
2013-04-03 08:28:36 -05:00
|
|
|
fn sort_and_fmt(mm: &HashMap<~[u8], uint>, total: uint) -> ~str {
|
2012-07-06 17:15:52 -05:00
|
|
|
fn pct(xx: uint, yy: uint) -> float {
|
2012-08-01 19:30:05 -05:00
|
|
|
return (xx as float) * 100f / (yy as float);
|
2012-07-06 17:15:52 -05:00
|
|
|
}
|
|
|
|
|
2013-03-22 13:23:21 -05:00
|
|
|
fn le_by_val<TT:Copy,UU:Copy + Ord>(kv0: &(TT,UU),
|
2012-08-02 17:42:56 -05:00
|
|
|
kv1: &(TT,UU)) -> bool {
|
|
|
|
let (_, v0) = *kv0;
|
|
|
|
let (_, v1) = *kv1;
|
2012-08-01 19:30:05 -05:00
|
|
|
return v0 >= v1;
|
2012-07-06 17:15:52 -05:00
|
|
|
}
|
|
|
|
|
2013-03-22 13:23:21 -05:00
|
|
|
fn le_by_key<TT:Copy + Ord,UU:Copy>(kv0: &(TT,UU),
|
2012-08-02 17:42:56 -05:00
|
|
|
kv1: &(TT,UU)) -> bool {
|
|
|
|
let (k0, _) = *kv0;
|
|
|
|
let (k1, _) = *kv1;
|
2012-08-01 19:30:05 -05:00
|
|
|
return k0 <= k1;
|
2012-07-06 17:15:52 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
// sort by key, then by value
|
2013-02-20 19:07:17 -06:00
|
|
|
fn sortKV<TT:Copy + Ord,UU:Copy + Ord>(orig: ~[(TT,UU)]) -> ~[(TT,UU)] {
|
2012-09-27 20:44:31 -05:00
|
|
|
return sort::merge_sort(sort::merge_sort(orig, le_by_key), le_by_val);
|
2012-07-06 17:15:52 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
let mut pairs = ~[];
|
|
|
|
|
|
|
|
// map -> [(k,%)]
|
2013-04-06 10:22:36 -05:00
|
|
|
for mm.each |&key, &val| {
|
2012-09-26 19:33:34 -05:00
|
|
|
pairs.push((key, pct(val, total)));
|
2013-02-02 13:47:41 -06:00
|
|
|
}
|
2012-07-06 17:15:52 -05:00
|
|
|
|
|
|
|
let pairs_sorted = sortKV(pairs);
|
2012-09-19 18:55:01 -05:00
|
|
|
|
2012-07-14 00:57:48 -05:00
|
|
|
let mut buffer = ~"";
|
2012-07-06 17:15:52 -05:00
|
|
|
|
2012-09-19 18:55:01 -05:00
|
|
|
for pairs_sorted.each |kv| {
|
2012-12-14 17:57:59 -06:00
|
|
|
let (k,v) = copy *kv;
|
2012-09-19 18:55:01 -05:00
|
|
|
unsafe {
|
2013-04-23 11:20:55 -05:00
|
|
|
let b = str::raw::from_bytes(k);
|
|
|
|
// FIXME: #4318 Instead of to_ascii and to_str_ascii, could use
|
|
|
|
// to_ascii_consume and to_str_consume to not do a unnecessary copy.
|
|
|
|
buffer += (fmt!("%s %0.3f\n", b.to_ascii().to_upper().to_str_ascii(), v));
|
2012-09-19 18:55:01 -05:00
|
|
|
}
|
|
|
|
}
|
2012-07-06 17:15:52 -05:00
|
|
|
|
2012-08-01 19:30:05 -05:00
|
|
|
return buffer;
|
2012-07-06 17:15:52 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
// given a map, search for the frequency of a pattern
|
2013-04-03 08:28:36 -05:00
|
|
|
fn find(mm: &HashMap<~[u8], uint>, key: ~str) -> uint {
|
2013-04-23 11:20:55 -05:00
|
|
|
// FIXME: #4318 Instead of to_ascii and to_str_ascii, could use
|
|
|
|
// to_ascii_consume and to_str_consume to not do a unnecessary copy.
|
|
|
|
match mm.find(&str::to_bytes(key.to_ascii().to_lower().to_str_ascii())) {
|
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; }
|
2012-07-06 17:15:52 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// given a map, increment the counter for a key
|
2013-04-03 08:28:36 -05:00
|
|
|
fn update_freq(mm: &mut HashMap<~[u8], uint>, key: &[u8]) {
|
2013-02-08 13:28:20 -06:00
|
|
|
let key = vec::slice(key, 0, key.len()).to_vec();
|
2013-03-23 20:22:00 -05:00
|
|
|
let newval = match mm.pop(&key) {
|
|
|
|
Some(v) => v + 1,
|
|
|
|
None => 1
|
|
|
|
};
|
|
|
|
mm.insert(key, newval);
|
2012-07-06 17:15:52 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
// given a ~[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"
|
2012-07-26 16:40:00 -05:00
|
|
|
fn windows_with_carry(bb: &[u8], nn: uint,
|
2013-03-07 16:38:38 -06:00
|
|
|
it: &fn(window: &[u8])) -> ~[u8] {
|
2012-07-06 17:15:52 -05:00
|
|
|
let mut ii = 0u;
|
|
|
|
|
|
|
|
let len = vec::len(bb);
|
|
|
|
while ii < len - (nn - 1u) {
|
2013-02-08 13:28:20 -06:00
|
|
|
it(vec::slice(bb, ii, ii+nn));
|
2012-07-06 17:15:52 -05:00
|
|
|
ii += 1u;
|
|
|
|
}
|
|
|
|
|
2013-02-08 13:28:20 -06:00
|
|
|
return vec::slice(bb, len - (nn - 1u), len).to_vec();
|
2012-07-06 17:15:52 -05:00
|
|
|
}
|
|
|
|
|
2013-04-22 23:19:58 -05:00
|
|
|
fn make_sequence_processor(sz: uint,
|
|
|
|
from_parent: comm::Port<~[u8]>,
|
2013-02-02 05:10:12 -06:00
|
|
|
to_parent: comm::Chan<~str>) {
|
2013-04-03 08:28:36 -05:00
|
|
|
let mut freqs: HashMap<~[u8], uint> = HashMap::new();
|
2012-07-06 17:15:52 -05:00
|
|
|
let mut carry: ~[u8] = ~[];
|
|
|
|
let mut total: uint = 0u;
|
|
|
|
|
|
|
|
let mut line: ~[u8];
|
|
|
|
|
|
|
|
loop {
|
|
|
|
|
|
|
|
line = from_parent.recv();
|
|
|
|
if line == ~[] { break; }
|
|
|
|
|
|
|
|
carry = windows_with_carry(carry + line, sz, |window| {
|
2013-03-23 20:22:00 -05:00
|
|
|
update_freq(&mut freqs, window);
|
2012-07-06 17:15:52 -05:00
|
|
|
total += 1u;
|
|
|
|
});
|
|
|
|
}
|
|
|
|
|
2013-02-01 01:13:36 -06:00
|
|
|
let buffer = match sz {
|
2013-03-23 20:22:00 -05:00
|
|
|
1u => { sort_and_fmt(&freqs, total) }
|
|
|
|
2u => { sort_and_fmt(&freqs, total) }
|
|
|
|
3u => { fmt!("%u\t%s", find(&freqs, ~"GGT"), ~"GGT") }
|
|
|
|
4u => { fmt!("%u\t%s", find(&freqs, ~"GGTA"), ~"GGTA") }
|
|
|
|
6u => { fmt!("%u\t%s", find(&freqs, ~"GGTATT"), ~"GGTATT") }
|
|
|
|
12u => { fmt!("%u\t%s", find(&freqs, ~"GGTATTTTAATT"), ~"GGTATTTTAATT") }
|
|
|
|
18u => { fmt!("%u\t%s", find(&freqs, ~"GGTATTTTAATTTATAGT"), ~"GGTATTTTAATTTATAGT") }
|
2012-08-03 21:59:04 -05:00
|
|
|
_ => { ~"" }
|
2012-07-06 17:15:52 -05:00
|
|
|
};
|
|
|
|
|
2013-02-15 04:44:18 -06:00
|
|
|
to_parent.send(buffer);
|
2012-07-06 17:15:52 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
// given a FASTA file on stdin, process sequence THREE
|
2012-10-03 21:16:27 -05:00
|
|
|
fn main() {
|
|
|
|
let args = os::args();
|
2012-07-14 00:57:48 -05:00
|
|
|
let rdr = if os::getenv(~"RUST_BENCH").is_some() {
|
2012-07-06 17:15:52 -05:00
|
|
|
// FIXME: Using this compile-time env variable is a crummy way to
|
2012-11-07 15:41:02 -06:00
|
|
|
// get to this massive data set, but include_bin! chokes on it (#2598)
|
2012-08-24 17:28:43 -05:00
|
|
|
let path = Path(env!("CFG_SRC_DIR"))
|
|
|
|
.push_rel(&Path("src/test/bench/shootout-k-nucleotide.data"));
|
2012-09-25 18:23:04 -05:00
|
|
|
result::get(&io::file_reader(&path))
|
2012-07-06 17:15:52 -05:00
|
|
|
} else {
|
|
|
|
io::stdin()
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
// initialize each sequence sorter
|
2012-09-19 00:44:34 -05:00
|
|
|
let sizes = ~[1,2,3,4,6,12,18];
|
2012-08-20 14:23:37 -05:00
|
|
|
let streams = vec::map(sizes, |_sz| Some(stream()));
|
2013-02-15 04:44:18 -06:00
|
|
|
let mut streams = streams;
|
2012-07-06 17:15:52 -05:00
|
|
|
let mut from_child = ~[];
|
|
|
|
let to_child = vec::mapi(sizes, |ii, sz| {
|
2012-09-21 20:43:30 -05:00
|
|
|
let sz = *sz;
|
2012-08-20 14:23:37 -05:00
|
|
|
let mut stream = None;
|
2012-07-06 17:15:52 -05:00
|
|
|
stream <-> streams[ii];
|
2013-03-16 14:49:12 -05:00
|
|
|
let (from_child_, to_parent_) = stream.unwrap();
|
2012-07-06 17:15:52 -05:00
|
|
|
|
2013-02-15 04:44:18 -06:00
|
|
|
from_child.push(from_child_);
|
2012-07-06 17:15:52 -05:00
|
|
|
|
2013-02-02 05:10:12 -06:00
|
|
|
let (from_parent, to_child) = comm::stream();
|
2012-07-06 17:15:52 -05:00
|
|
|
|
2013-02-15 04:44:18 -06:00
|
|
|
do task::spawn_with(from_parent) |from_parent| {
|
2012-07-06 17:15:52 -05:00
|
|
|
make_sequence_processor(sz, from_parent, to_parent_);
|
|
|
|
};
|
2013-02-01 01:13:36 -06:00
|
|
|
|
2013-02-15 04:44:18 -06:00
|
|
|
to_child
|
2012-07-06 17:15:52 -05:00
|
|
|
});
|
2013-02-01 01:13:36 -06:00
|
|
|
|
|
|
|
|
2012-07-06 17:15:52 -05:00
|
|
|
// latch stores true after we've started
|
|
|
|
// reading the sequence of interest
|
|
|
|
let mut proc_mode = false;
|
|
|
|
|
|
|
|
while !rdr.eof() {
|
2012-07-14 00:57:48 -05:00
|
|
|
let line: ~str = rdr.read_line();
|
2012-07-06 17:15:52 -05:00
|
|
|
|
2012-09-07 17:32:04 -05:00
|
|
|
if str::len(line) == 0u { loop; }
|
2012-07-06 17:15:52 -05:00
|
|
|
|
2012-08-06 14:34:08 -05:00
|
|
|
match (line[0], proc_mode) {
|
2012-07-06 17:15:52 -05:00
|
|
|
|
|
|
|
// start processing if this is the one
|
2012-08-03 21:59:04 -05:00
|
|
|
('>' as u8, false) => {
|
2012-08-06 14:34:08 -05:00
|
|
|
match str::find_str_from(line, ~"THREE", 1u) {
|
2012-08-20 14:23:37 -05:00
|
|
|
option::Some(_) => { proc_mode = true; }
|
|
|
|
option::None => { }
|
2012-07-06 17:15:52 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// break our processing
|
2012-08-03 21:59:04 -05:00
|
|
|
('>' as u8, true) => { break; }
|
2012-07-06 17:15:52 -05:00
|
|
|
|
|
|
|
// process the sequence for k-mers
|
2012-08-03 21:59:04 -05:00
|
|
|
(_, true) => {
|
2012-08-23 17:44:57 -05:00
|
|
|
let line_bytes = str::to_bytes(line);
|
2012-07-06 17:15:52 -05:00
|
|
|
|
|
|
|
for sizes.eachi |ii, _sz| {
|
2012-12-14 17:57:59 -06:00
|
|
|
let mut lb = copy line_bytes;
|
2012-07-06 17:15:52 -05:00
|
|
|
to_child[ii].send(lb);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// whatever
|
2012-08-03 21:59:04 -05:00
|
|
|
_ => { }
|
2012-07-06 17:15:52 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// finish...
|
|
|
|
for sizes.eachi |ii, _sz| {
|
|
|
|
to_child[ii].send(~[]);
|
|
|
|
}
|
|
|
|
|
|
|
|
// now fetch and print result messages
|
|
|
|
for sizes.eachi |ii, _sz| {
|
|
|
|
io::println(from_child[ii].recv());
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|