Documentation

src/algorithms/npag.rs

@@ -19,8 +19,8 @@ use ndarray::{Array, Array1, Array2, Axis};
 use ndarray_stats::{DeviationExt, QuantileExt};
 use tokio::sync::mpsc::UnboundedSender;
 
-const THETA_E: f64 = 1e-4; //convergence Criteria
-const THETA_G: f64 = 1e-4; //objf stop criteria
+const THETA_E: f64 = 1e-4; // Convergence criteria
+const THETA_G: f64 = 1e-4; // Objective function convergence criteria
 const THETA_F: f64 = 1e-2;
 const THETA_D: f64 = 1e-4;
 

src/routines/datafile.rs

@@ -5,7 +5,8 @@ use std::process::exit;
 
 type Record = HashMap<String, String>;
 
-/// A Scenario is a collection of blocks that represent a single subject in the Datafile
+/// A Scenario is a collection of blocks that represent a single subject in the datafile
+/// Each block is a collection of events that represent a single dose, possibly followed by observations
 #[derive(Debug, Clone)]
 pub struct Scenario {
     pub id: String,
@@ -23,8 +24,8 @@ impl Scenario {
     }
 
     /// Adds "mock" events to a Scenario in order to generate predictions at those times
-    /// The interval is mapped to the `idelta`-setting in the configuration time
-    /// Time after dose (tad) will ensure that predictions are made until the last dose + tad
+    /// The interval is mapped to the `idelta`-setting in the configuration file
+    /// Time after dose (`tad`) will ensure that predictions are made until the last dose + tad
     pub fn add_event_interval(&self, interval: f64, tad: f64) -> Self {
         // Clone the underlying Event data instead of the references
         let all_events = self

src/routines/evaluation/ipm.rs

@@ -19,7 +19,7 @@ type OneDimArray = ArrayBase<OwnedRepr<f64>, ndarray::Dim<[usize; 1]>>;
 ///
 /// A `Result` containing a tuple with two elements:
 ///
-/// * `lam` - An Array1<f64> representing the solution of the optimization problem.
+/// * `lam` - An `Array1<f64>` representing the solution of the optimization problem.
 /// * `obj` - A f64 value representing the objective function value at the solution.
 ///
 /// # Errors

src/routines/expansion/adaptative_grid.rs

@@ -2,6 +2,13 @@ use ndarray::{Array, Array2};
 
 use crate::routines::condensation::prune::prune;
 
+/// Adaptive grid algorithm for support point expansion
+/// 
+/// For each support point, generate up to 2 new support points in each dimension
+/// 
+/// New support points are symmetrically placed around the original support point, at a distance of eps * (range_max - range_min)
+/// 
+/// If the new support point is too close to an existing support point, or it is outside the given range, it is discarded
 pub fn adaptative_grid(
     theta: &mut Array2<f64>,
     eps: f64,

src/routines/output.rs

@@ -290,6 +290,15 @@ impl CycleLog {
     }
 }
 
+/// Defines the result objects from a run
+/// An [NPResult] contains the necessary information to generate predictions and summary statistics
+/// It holds the following information:
+/// - `cycle`: The cycle number
+/// - `objf`: The objective function value
+/// - `gamlam`: The assay noise parameter, either gamma or lambda
+/// - `theta`: The support points and their associated probabilities
+/// - `nspp`: The number of support points
+/// - `delta_objf`: The change in objective function value from last cycle
 #[derive(Debug, Clone)]
 pub struct NPCycle {
     pub cycle: usize,

src/routines/simulation/predict.rs

@@ -160,16 +160,16 @@ pub fn get_ypred<S: Predict<'static> + Sync + Clone>(
 ///
 /// * `sim_eng` - A reference to the simulation engine implementing the `Predict` trait.
 ///
-/// * `scenarios` - A reference to a Vec<Scenario> containing information about different scenarios.
+/// * `scenarios` - A reference to a `Vec<Scenario>` containing information about different scenarios.
 ///
-/// * `support_points` - A 2D Array (Array2<f64>) representing the support points. Each row
+/// * `support_points` - A 2D Array `(Array2<f64>)` representing the support points. Each row
 ///                     corresponds to a different support point scenario.
 ///
 /// * `cache` - A boolean flag indicating whether to cache predicted values during simulation.
 ///
 /// # Returns
 ///
-/// A 2D Array (Array2<Array1<f64>>) where each element is an Array1<f64> representing the
+/// A 2D Array `(Array2<Array1<f64>>)` where each element is an `Array1<f64>` representing the
 /// simulated observations for a specific scenario and support point.
 ///
 /// # Example