Include lowercase name fields in Artist,Album,Song

[fsktom]

Noticed it doing #33 that I'm calling .to_lowercase() a lot (bc of summing across albums and ignoring capitalization). Maybe it'd be nice if each SongEntry and song/album/artist contained an Rc<str> with its name but in lowercase. Would take up more space but I think it could speed up the computations.

[fsktom]

Some benchmarks @ a4f0bf2 with only endsong_0.json before doing anything

parsing:

parse, sum and filter   time:   [24.457 ms 24.565 ms 24.719 ms]
Found 8 outliers among 100 measurements (8.00%)
  2 (2.00%) high mild
  6 (6.00%) high severe

gathering songs summed across albums:

gather songs summed across albums
                        time:   [16.655 ms 16.965 ms 17.356 ms]
Found 17 outliers among 100 measurements (17.00%)
  4 (4.00%) high mild
  13 (13.00%) high severe

[fsktom]

hmm
parsing (and even summing different capitalizations after making it use the intrinsic _lowercase fields) is 25% slower

parse, sum and filter   time:   [30.462 ms 30.833 ms 31.556 ms]
                        change: [+23.504% +25.515% +29.583%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 5 outliers among 100 measurements (5.00%)
  4 (4.00%) high mild
  1 (1.00%) high severe

[fsktom]

huh

gather songs summed across albums
                        time:   [16.378 ms 16.578 ms 16.952 ms]
                        change: [-5.0224% -2.2867% +0.8282%] (p = 0.14 > 0.05)
                        No change in performance detected.
Found 15 outliers among 100 measurements (15.00%)
  3 (3.00%) low mild
  4 (4.00%) high mild
  8 (8.00%) high severe

didn't really make a difference
XD

[fsktom]

Hmm maybe I should think about using a crate called unicase for no allocation string case-insensitive comparisons?

https://www.reddit.com/r/rust/comments/6wbru2/performance_issue_can_i_avoid_of_using_the_slow/dm6u0as/
https://crates.io/crates/unicase

But how would I do sum_different_capitalization or gather::songs_summed_across_albums with it?

[fsktom]

I've done quite a bit of rewriting, benchmarking and profiling. Using UniCase in sum_different_capitalization doesn't make a difference really compared to .to_lowercase()

[fsktom]

hey but at least find is 10x faster now :)

find v1                 time:   [1.0716 ms 1.0754 ms 1.0802 ms]
Found 11 outliers among 100 measurements (11.00%)
  2 (2.00%) high mild
  9 (9.00%) high severe

find v2                 time:   [11.730 ms 11.774 ms 11.831 ms]
Found 13 outliers among 100 measurements (13.00%)
  1 (1.00%) low mild
  1 (1.00%) high mild
  11 (11.00%) high severe

[fsktom]

XDDD find song ignore album went from 7ms to 200μs XD

find song               time:   [6.9057 ms 7.0559 ms 7.2482 ms]
Found 7 outliers among 100 measurements (7.00%)
  1 (1.00%) high mild
  6 (6.00%) high severe

to

find song               time:   [207.80 µs 209.54 µs 211.58 µs]
                        change: [-97.110% -97.023% -96.946%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 10 outliers among 100 measurements (10.00%)
  4 (4.00%) low mild
  2 (2.00%) high mild
  4 (4.00%) high severe

 pub fn song(entries: &[SongEntry], song_name: &str, artist_name: &str) -> Option<Vec<Song>> {
-    let (song_name, artist_name) = (song_name.to_lowercase(), artist_name.to_lowercase());
+    let song_bogus_album = Song::new(song_name, "", artist_name);

     let song_versions = entries
         .iter()
-        .filter(|entry| {
-            entry.track.to_lowercase() == song_name && entry.artist.to_lowercase() == artist_name
-        })
+        .filter(|entry| song_bogus_album.is_entry_lowercase_ignore_album(entry))
         .map(Song::from)
         .unique()
         .collect_vec();

and

pub fn is_entry_lowercase_ignore_album(&self, entry: &SongEntry) -> bool {
    self.album.artist.is_entry(entry) && UniCase::new(&entry.track) == UniCase::new(&self.name)
}

[fsktom]

So all in all
after MANY hours of trying things out and testing
decided not to include new fields
but instead use UniCase for case-insensitive comparisons

[fsktom]

gather songs summed across albums
time: [16.378 ms 16.578 ms 16.952 ms]
change: [-5.0224% -2.2867% +0.8282%] (p = 0.14 > 0.05)
No change in performance detected.
Found 15 outliers among 100 measurements (15.00%)
3 (3.00%) low mild
4 (4.00%) high mild
8 (8.00%) high severe

improved performance of this in fefc4f9
from

gather songs summed across albums
                        time:   [16.684 ms 17.729 ms 19.667 ms]
Found 16 outliers among 100 measurements (16.00%)
  4 (4.00%) high mild
  12 (12.00%) high severe

in b764ae0

to

gather songs summed across albums
                        time:   [1.8376 ms 1.8443 ms 1.8536 ms]
                        change: [-90.573% -89.523% -88.853%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 12 outliers among 100 measurements (12.00%)
  3 (3.00%) high mild
  9 (9.00%) high severe

because I'm doing it far smarter now and even fixed a bug heh...