alice_analytics/lib.rs
1//! ALICE-Analytics — High-Performance Telemetry & Statistical Estimation
2//!
3//! Probabilistic data structures for streaming analytics with mathematical
4//! error guarantees and minimal memory footprint.
5//!
6//! # Modules
7//!
8//! - [`sketch`]: `HyperLogLog` / `DDSketch` / Count-Min / Heavy Hitters
9//! - [`stats`]: Percentile rank, IQR, covariance, Welford streaming stats
10//! - [`window`]: Tumbling / sliding / hierarchical window aggregates
11//! - [`anomaly`]: MAD / Z-Score / EWMA / composite anomaly detectors
12//! - [`privacy`]: Laplace noise / Randomized Response / RAPPOR / privacy budget
13//! - [`pipeline`]: Lock-free metric aggregation pipeline
14//! - [`export`]: JSON / Prometheus export of metric snapshots
15//! - [`streaming_ops`]: Streaming aggregation operators
16//! - `law` (feature `law`): residual distribution of a fitted
17//! `alice_zip::law::SignalLaw`, summarised with a `DDSketch`
18//!
19//! # Example
20//!
21//! ```rust
22//! use alice_analytics::sketch::{DDSketch, HyperLogLog};
23//!
24//! // quantiles with a relative-error guarantee of 1 %
25//! let mut latency = DDSketch::new(0.01);
26//! for ms in 1..=1000 {
27//! latency.insert(f64::from(ms));
28//! }
29//! let p99 = latency.quantile(0.99);
30//! assert!((p99 - 990.0).abs() <= 0.01 * 990.0);
31//!
32//! // distinct count in 16 KiB of registers
33//! let mut users = HyperLogLog::new();
34//! for id in 0..10_000_u64 {
35//! users.insert(&id);
36//! }
37//! let n = users.cardinality();
38//! assert!((n - 10_000.0).abs() < 0.05 * 10_000.0);
39//! ```
40
41#![cfg_attr(not(feature = "std"), no_std)]
42
43pub(crate) mod math;
44
45pub mod anomaly;
46pub mod export;
47pub mod pipeline;
48pub mod privacy;
49pub mod sketch;
50pub mod stats;
51pub mod streaming_ops;
52pub mod window;
53
54#[cfg(feature = "law")]
55pub mod law;