Rename recursive fitter and binary model

examples/abcdef.rs

@@ -3,7 +3,7 @@ extern crate blas_src;
 use kickscore as ks;
 
 fn main() {
-    let mut model = ks::BinaryModel::new(ks::BinaryModelObservation::Probit);
+    let mut model = ks::model::Binary::new(ks::model::binary::Observation::Probit);
 
     for player in &["A", "B", "C", "D", "E", "F"] {
         let kernel: Vec<Box<dyn ks::Kernel>> = vec![

examples/kickscore-basics.rs

@@ -3,7 +3,7 @@ extern crate blas_src;
 use kickscore as ks;
 
 fn main() {
-    let mut model = ks::BinaryModel::new(ks::BinaryModelObservation::Probit);
+    let mut model = ks::model::Binary::new(ks::model::binary::Observation::Probit);
 
     // Spike's skill does not change over time.
     let k_spike = ks::kernel::Constant::new(0.5);

examples/nba-history.rs

@@ -54,7 +54,7 @@ fn main() -> Result<(), Box<dyn std::error::Error>> {
 
     let seconds_in_year = 365.25 * 24.0 * 60.0 * 60.0;
 
-    let mut model = ks::BinaryModel::new(ks::BinaryModelObservation::Probit);
+    let mut model = ks::model::Binary::new(ks::model::binary::Observation::Probit);
 
     for team in teams {
         let kernel: Vec<Box<dyn ks::Kernel>> = vec![

src/fitter.rs

@@ -1,6 +1,6 @@
 mod recursive;
 
-pub use recursive::RecursiveFitter;
+pub use recursive::Recursive;
 
 pub trait Fitter {
     fn add_sample(&mut self, t: f64) -> usize;

src/fitter/recursive.rs

@@ -7,7 +7,7 @@ use crate::kernel::Kernel;
 
 use super::Fitter;
 
-pub struct RecursiveFitter {
+pub struct Recursive {
     ts_new: Vec<f64>,
     kernel: Box<dyn Kernel>,
     ts: Vec<f64>,
@@ -28,7 +28,7 @@ pub struct RecursiveFitter {
     p_s: Vec<Array2<f64>>,
 }
 
-impl RecursiveFitter {
+impl Recursive {
     pub fn new(kernel: Box<dyn Kernel>) -> Self {
         let m = kernel.order();
         let h = kernel.measurement_vector();
@@ -56,7 +56,7 @@ impl RecursiveFitter {
     }
 }
 
-impl fmt::Debug for RecursiveFitter {
+impl fmt::Debug for Recursive {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         f.debug_struct("RecursiveFitter")
             .field("ts_new", &self.ts_new)
@@ -80,7 +80,7 @@ impl fmt::Debug for RecursiveFitter {
     }
 }
 
-impl Fitter for RecursiveFitter {
+impl Fitter for Recursive {
     fn add_sample(&mut self, t: f64) -> usize {
         let idx = self.ts.len() + self.ts_new.len();
 

src/lib.rs

@@ -5,10 +5,10 @@ mod item;
 pub mod kernel;
 mod linalg;
 mod math;
-mod model;
+pub mod model;
 pub mod observation;
 mod storage;
 mod utils;
 
 pub use kernel::Kernel;
-pub use model::*;
+// pub use model::*;

src/model.rs

@@ -1,7 +1,7 @@
-mod binary;
+pub mod binary;
 mod difference;
 mod ternary;
 
-pub use binary::*;
+pub use binary::Binary;
 pub use difference::*;
 pub use ternary::*;

src/model/binary.rs

@@ -1,6 +1,6 @@
 use std::f64;
 
-use crate::fitter::RecursiveFitter;
+use crate::fitter::Recursive;
 use crate::item::Item;
 use crate::kernel::Kernel;
 use crate::observation::*;
@@ -12,28 +12,28 @@ trait BinaryObeservation: Observation {
 }
 */
 
-pub enum BinaryModelObservation {
+pub enum Observation {
     Probit,
     Logit,
 }
 
 #[derive(Clone, Copy)]
-pub enum BinaryModelFitMethod {
+pub enum FitMethod {
     Ep,
     Kl,
 }
 
-pub struct BinaryModel {
+pub struct Binary {
     pub storage: Storage,
     last_t: f64,
-    win_obs: BinaryModelObservation,
+    win_obs: Observation,
-    observations: Vec<Box<dyn Observation>>,
+    observations: Vec<Box<dyn crate::observation::Observation>>,
-    last_method: Option<BinaryModelFitMethod>,
+    last_method: Option<FitMethod>,
 }
 
-impl BinaryModel {
+impl Binary {
-    pub fn new(win_obs: BinaryModelObservation) -> Self {
+    pub fn new(win_obs: Observation) -> Self {
-        BinaryModel {
+        Binary {
             storage: Storage::default(),
             last_t: f64::NEG_INFINITY,
             win_obs,
@@ -49,7 +49,7 @@ impl BinaryModel {
 
         self.storage.insert(
             name.to_string(),
-            Item::new(Box::new(RecursiveFitter::new(kernel))),
+            Item::new(Box::new(Recursive::new(kernel))),
         );
     }
 
@@ -72,11 +72,11 @@ impl BinaryModel {
         let mut elems = self.process_items(winners, 1.0);
         elems.extend(self.process_items(losers, -1.0));
 
-        let obs: Box<dyn Observation> = match self.win_obs {
+        let obs: Box<dyn crate::observation::Observation> = match self.win_obs {
-            BinaryModelObservation::Probit => {
+            Observation::Probit => {
                 Box::new(ProbitWinObservation::new(&mut self.storage, &elems, t, 0.0))
             }
-            BinaryModelObservation::Logit => {
+            Observation::Logit => {
                 Box::new(LogitWinObservation::new(&mut self.storage, &elems, t, 0.0))
             }
         };
@@ -95,7 +95,7 @@ impl BinaryModel {
     pub fn fit(&mut self) -> bool {
         // method="ep", lr=1.0, tol=1e-3, max_iter=100, verbose=False):
 
-        let method = BinaryModelFitMethod::Ep;
+        let method = FitMethod::Ep;
         let lr = 1.0;
         let tol = 1e-3;
         let max_iter = 100;
@@ -112,8 +112,8 @@ impl BinaryModel {
 
             for obs in &mut self.observations {
                 let diff = match method {
-                    BinaryModelFitMethod::Ep => obs.ep_update(&mut self.storage, lr),
+                    FitMethod::Ep => obs.ep_update(&mut self.storage, lr),
-                    BinaryModelFitMethod::Kl => obs.kl_update(&mut self.storage, lr),
+                    FitMethod::Kl => obs.kl_update(&mut self.storage, lr),
                 };
 
                 if diff > max_diff {
@@ -142,10 +142,8 @@ impl BinaryModel {
         elems.extend(self.process_items(team_2, -1.0));
 
         let prob = match self.win_obs {
-            BinaryModelObservation::Probit => {
+            Observation::Probit => ProbitWinObservation::probability(&self.storage, &elems, t, 0.0),
-                ProbitWinObservation::probability(&self.storage, &elems, t, 0.0)
+            Observation::Logit => todo!(),
-            }
-            BinaryModelObservation::Logit => todo!(),
         };
 
         (prob, 1.0 - prob)

src/model/difference.rs

@@ -1,6 +1,6 @@
 use std::f64;
 
-use crate::fitter::RecursiveFitter;
+use crate::fitter::Recursive;
 use crate::item::Item;
 use crate::kernel::Kernel;
 use crate::observation::*;
@@ -38,7 +38,7 @@ impl DifferenceModel {
 
         self.storage.insert(
             name.to_string(),
-            Item::new(Box::new(RecursiveFitter::new(kernel))),
+            Item::new(Box::new(Recursive::new(kernel))),
         );
     }
 

src/model/ternary.rs

@@ -1,6 +1,6 @@
 use std::f64;
 
-use crate::fitter::RecursiveFitter;
+use crate::fitter::Recursive;
 use crate::item::Item;
 use crate::kernel::Kernel;
 use crate::observation::*;
@@ -46,7 +46,7 @@ impl TernaryModel {
 
         self.storage.insert(
             name.to_string(),
-            Item::new(Box::new(RecursiveFitter::new(kernel))),
+            Item::new(Box::new(Recursive::new(kernel))),
         );
     }
 

src/observation/ordinal.rs

@@ -1,9 +1,7 @@
-use std::f64::consts::{SQRT_2, TAU};
+use std::f64::consts::TAU;
 
 use crate::storage::Storage;
-use crate::utils::{
+use crate::utils::{logphi, normcdf, normpdf, ROOTS_HERMITENORM_WS, ROOTS_HERMITENORM_XS};
-    logphi, logsumexp2, normcdf, normpdf, ROOTS_HERMITENORM_WS, ROOTS_HERMITENORM_XS,
-};
 
 use super::{f_params, Core, Observation};
 
@@ -54,52 +52,6 @@ fn ll_probit_tie(x: f64, margin: f64) -> f64 {
     }
 }
 
-fn lambdas() -> [f64; 5] {
-    [
-        0.44 * SQRT_2,
-        0.41 * SQRT_2,
-        0.40 * SQRT_2,
-        0.39 * SQRT_2,
-        0.36 * SQRT_2,
-    ]
-}
-
-const CS: [f64; 5] = [
-    1.146480988574439e+02,
-    -1.508871030070582e+03,
-    2.676085036831241e+03,
-    -1.356294962039222e+03,
-    7.543285642111850e+01,
-];
-
-fn mm_logit_win(mean_cav: f64, cov_cav: f64) {
-    let mut arr1 = [0.0; 5];
-    let mut arr2 = [0.0; 5];
-    let mut arr3 = [0.0; 5];
-
-    for (i, x) in lambdas().iter().enumerate() {
-        let (a, b, c) = mm_probit_win(x * mean_cav, x * x * cov_cav);
-
-        arr1[i] = a;
-        arr2[i] = b;
-        arr3[i] = c;
-    }
-
-    let logpart1 = logsumexp2(arr1, CS);
-
-    /*
-    dlogpart1 = (np.dot(np.exp(arr1) * arr2, CS * LAMBDAS)
-                 / np.dot(np.exp(arr1), CS))
-    d2logpart1 = (np.dot(np.exp(arr1) * (arr2 * arr2 + arr3),
-                         CS * LAMBDAS * LAMBDAS)
-                  / np.dot(np.exp(arr1), CS)) - (dlogpart1 * dlogpart1)
-    */
-}
-
-fn mm_logit_tie(x: f64, margin: f64) {
-    //
-}
-
 fn ll_logit_win(x: f64, margin: f64) -> f64 {
     let z = x - margin;
 

src/utils.rs

@@ -89,7 +89,7 @@ pub fn logphi(z: f64) -> (f64, f64) {
     }
 }
 
-pub fn logsumexp2(xs: [f64; 5], bs: [f64; 5]) -> f64 {
+pub fn logsumexp2(xs: &[f64], bs: &[f64]) -> f64 {
     let a = xs.iter().max_by(|a, b| a.partial_cmp(b).unwrap()).unwrap();
 
     a + bs

tests/binary-1.rs

@@ -4,7 +4,7 @@ use kickscore as ks;
 
 #[test]
 fn binary_1() {
-    let mut model = ks::BinaryModel::new(ks::BinaryModelObservation::Probit);
+    let mut model = ks::model::Binary::new(ks::model::binary::Observation::Probit);
 
     let k_audrey = ks::kernel::Matern52::new(1.0, 2.0);
     let k_benjamin = ks::kernel::Matern52::new(1.0, 2.0);

tests/kickscore-basics.rs

@@ -5,7 +5,7 @@ use kickscore as ks;
 
 #[test]
 fn kickscore_basic() {
-    let mut model = ks::BinaryModel::new(ks::BinaryModelObservation::Probit);
+    let mut model = ks::model::Binary::new(ks::model::binary::Observation::Probit);
 
     let k_spike = ks::kernel::Constant::new(0.5);
 

tests/nba-history.rs

@@ -58,7 +58,7 @@ fn nba_history() {
 
     let seconds_in_year = 365.25 * 24.0 * 60.0 * 60.0;
 
-    let mut model = ks::BinaryModel::new(ks::BinaryModelObservation::Probit);
+    let mut model = ks::model::Binary::new(ks::model::binary::Observation::Probit);
 
     for team in teams {
         let kernel: Vec<Box<dyn ks::Kernel>> = vec![