diff options
| author | Ellie Huxtable <ellie@elliehuxtable.com> | 2024-06-26 12:40:17 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-06-26 12:40:17 +0100 |
| commit | 1201caee5c7b3020e1c879e527feb934fd1d8023 (patch) | |
| tree | 5af6ecc78075401e459c97b65f74689e141d5137 /crates/atuin-history | |
| parent | feat: add several other GitHub access token patterns (#2200) (diff) | |
| download | atuin-1201caee5c7b3020e1c879e527feb934fd1d8023.zip | |
perf(search): benchmark smart sort (#2202)
Diffstat (limited to 'crates/atuin-history')
| -rw-r--r-- | crates/atuin-history/Cargo.toml | 9 | ||||
| -rw-r--r-- | crates/atuin-history/benches/smart_sort.rs | 35 | ||||
| -rw-r--r-- | crates/atuin-history/src/lib.rs | 1 | ||||
| -rw-r--r-- | crates/atuin-history/src/sort.rs | 46 |
4 files changed, 91 insertions, 0 deletions
diff --git a/crates/atuin-history/Cargo.toml b/crates/atuin-history/Cargo.toml index 4836cf74..2ed0e1a5 100644 --- a/crates/atuin-history/Cargo.toml +++ b/crates/atuin-history/Cargo.toml @@ -39,3 +39,12 @@ tracing = "0.1" uuid = { workspace = true } unicode-segmentation = "1.11.0" sysinfo = "0.30.7" + +[dev-dependencies] +tracing-tree = "0.3" +divan = "0.1.14" +rand = { workspace = true } + +[[bench]] +name = "smart_sort" +harness = false diff --git a/crates/atuin-history/benches/smart_sort.rs b/crates/atuin-history/benches/smart_sort.rs new file mode 100644 index 00000000..a78064de --- /dev/null +++ b/crates/atuin-history/benches/smart_sort.rs @@ -0,0 +1,35 @@ +use atuin_client::history::History; +use atuin_history::sort::sort; + +use rand::Rng; + +fn main() { + // Run registered benchmarks. + divan::main(); +} + +// Smart sort usually runs on 200 entries, test on a few sizes +#[divan::bench(args=[100, 200, 400, 800, 1600, 10000])] +fn smart_sort(lines: usize) { + // benchmark a few different sizes of "history" + // first we need to generate some history. This will use a whole bunch of memory, sorry + let mut rng = rand::thread_rng(); + let now = time::OffsetDateTime::now_utc().unix_timestamp(); + + let possible_commands = ["echo", "ls", "cd", "grep", "atuin", "curl"]; + let mut commands = Vec::<History>::with_capacity(lines); + + for _ in 0..lines { + let command = possible_commands[rng.gen_range(0..possible_commands.len())]; + + let command = History::import() + .command(command) + .timestamp(time::OffsetDateTime::from_unix_timestamp(rng.gen_range(0..now)).unwrap()) + .build() + .into(); + + commands.push(command); + } + + let _ = sort("curl", commands); +} diff --git a/crates/atuin-history/src/lib.rs b/crates/atuin-history/src/lib.rs index 9d34677f..e7b33916 100644 --- a/crates/atuin-history/src/lib.rs +++ b/crates/atuin-history/src/lib.rs @@ -1 +1,2 @@ +pub mod sort; pub mod stats; diff --git a/crates/atuin-history/src/sort.rs b/crates/atuin-history/src/sort.rs new file mode 100644 index 00000000..4465a142 --- /dev/null +++ b/crates/atuin-history/src/sort.rs @@ -0,0 +1,46 @@ +use atuin_client::history::History; + +type ScoredHistory = (f64, History); + +// Fuzzy search already comes sorted by minspan +// This sorting should be applicable to all search modes, and solve the more "obvious" issues +// first. +// Later on, we can pass in context and do some boosts there too. +pub fn sort(query: &str, input: Vec<History>) -> Vec<History> { + // This can totally be extended. We need to be _careful_ that it's not slow. + // We also need to balance sorting db-side with sorting here. SQLite can do a lot, + // but some things are just much easier/more doable in Rust. + + let mut scored = input + .into_iter() + .map(|h| { + // If history is _prefixed_ with the query, score it more highly + let score = if h.command.starts_with(query) { + 2.0 + } else if h.command.contains(query) { + 1.75 + } else { + 1.0 + }; + + // calculate how long ago the history was, in seconds + let now = time::OffsetDateTime::now_utc().unix_timestamp(); + let time = h.timestamp.unix_timestamp(); + let diff = std::cmp::max(1, now - time); // no /0 please + + // prefer newer history, but not hugely so as to offset the other scoring + // the numbers will get super small over time, but I don't want time to overpower other + // scoring + #[allow(clippy::cast_precision_loss)] + let time_score = 1.0 + (1.0 / diff as f64); + let score = score * time_score; + + (score, h) + }) + .collect::<Vec<ScoredHistory>>(); + + scored.sort_by(|a, b| a.0.partial_cmp(&b.0).unwrap().reverse()); + + // Remove the scores and return the history + scored.into_iter().map(|(_, h)| h).collect::<Vec<History>>() +} |
