feat: Non-compartmental analysis by Siel · Pull Request #208 · LAPKB/pharmsol

Siel · 2026-02-13T19:49:59Z

No description provided.

This commit adds two related features: ## NormalizedRow API (parser/) - New struct for format-agnostic data parsing - Decouples column mapping from event creation logic - Full ADDL/II expansion support (both positive and negative directions) - Refactors pmetrics.rs to use NormalizedRow internally - Enables external tools (like vial) to reuse parsing logic without reimplementing ADDL expansion ## ResidualErrorModel (data/) - New for parametric algorithms (SAEM, FOCE) - Uses prediction-based sigma (vs observation-based in ErrorModel) - Adds and functions - Documentation clarifying ErrorModel vs ResidualErrorModel usage Both features are independent but included together to avoid merge conflicts.

* nca * wip: current version * feat: nca * clenup * chore: documentation * chore: cleanup * chore: cleanup * chore: deprecating ErrorModel in favor of AssayErrorModel, subdividing the likelihood module and deprecating linear space likelihood calculation functions * feat: the Data parsing is centraliced to NormalizedRow * feat: the ErrorModel -> AssayErrorModel * feat: validation * chore: cleanup * chore: cleanup

Co-authored-by: Markus Hovd <markushh@uio.no>

… require extra size

* chore: Rename modules and structures * Update src/error/mod.rs Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Update src/data/parser/mod.rs Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Name changes --------- Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

Copilot

Pull request overview

This is a major feature addition introducing Non-Compartmental Analysis (NCA) capabilities to pharmsol. The PR adds comprehensive NCA functionality for calculating standard pharmacokinetic parameters from concentration-time data, alongside refactoring error models into observation-based (assay) and prediction-based (residual) variants, and introducing a JSON model definition system with code generation.

Changes:

Adds complete NCA module with support for λz estimation, AUC calculations, clearance parameters, and steady-state analysis
Refactors error models into AssayErrorModels (observation-based, for NPAG/NPOD) and ResidualErrorModels (prediction-based, for SAEM/FOCE)
Introduces JSON model definition system with validation and code generation for analytical, ODE, and SDE models
Adds observation data processing utilities and dose introspection methods

Reviewed changes

Copilot reviewed 76 out of 83 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
tests/ode_optimizations.rs	Updates likelihood calculations to use log-space and new AssayErrorModels
tests/nca/*.rs	New NCA integration tests covering terminal phase, quality metrics, parameters, and AUC
src/simulator/mod.rs	Updates documentation to use inline code formatting for macro references
src/simulator/likelihood/*.rs	New modules for subject-level predictions and log-likelihood calculations
src/simulator/equation/*.rs	Updates to use AssayErrorModels and log-likelihood calculations
src/optimize/*.rs	Updates to use new log-likelihood matrix API
src/nca/*.rs	New NCA module with analysis, types, traits, and utilities
src/lib.rs	Updates public API exports to include NCA and refactored error models
src/json/*.rs	New JSON model definition system with validation and code generation
src/data/*.rs	Adds observation processing, dose introspection, and data row parsing
src/error/mod.rs	Updates error types to use DataError instead of PmetricsError
examples/*.rs	New NCA and JSON compilation examples, updates to existing examples

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-02-13T19:51:20Z

src/nca/mod.rs

+pub use bioavailability::{
+    bioavailability, bioequivalence, compare, metabolite_parent_ratio, BioavailabilityResult,
+    BioequivalenceResult,


The public API exports bioavailability functions but there are no corresponding module definitions or implementations visible in the diff. This suggests missing module files or incomplete implementation that could cause compilation errors.

Copilot · 2026-02-13T19:51:20Z

src/json/library/models/pk_1cmt_iv.json

+  "output": "x[0] / V",
+  "neqs": [1, 1],


The output equation x[0] / V references state variable index 0, but for a one-compartment IV model this should reference the concentration in the central compartment. Verify that this indexing is correct for the analytical function being used.

Copilot · 2026-02-13T19:51:20Z

src/data/parser/pmetrics.rs

                                .value()
                                .map_or_else(|| ".".to_string(), |v| v.to_string());
-                            let outeq = (obs.outeq() + 1).to_string();
+                            let outeq = obs.outeq().to_string();


The outeq is now written directly without adding 1, changing from 1-indexed to 0-indexed output. This is a breaking change in the file format that should be documented in a migration guide or changelog to help users update their data files.

Suggested change

let outeq = obs.outeq().to_string();

let outeq = (obs.outeq() + 1).to_string();

Copilot · 2026-02-13T19:51:21Z

CHANGELOG.md

 ### Other

- *(Exa)* when installing Vial on MacOs, the environment varaibles are not completly shared to the sandbox in which Vial is running, this changes are meant to provide vial a better way to approach finding the rust binary ([#181](https://github.com/LAPKB/pharmsol/pull/181))
+- _(Exa)_ when installing Papir on MacOs, the environment varaibles are not completly shared to the sandbox in which Papir is running, this changes are meant to provide papir a better way to approach finding the rust binary ([#181](https://github.com/LAPKB/pharmsol/pull/181))


Corrected spelling of 'varaibles' to 'variables' and 'completly' to 'completely'.

Suggested change

- _(Exa)_ when installing Papir on MacOs, the environment varaibles are not completly shared to the sandbox in which Papir is running, this changes are meant to provide papir a better way to approach finding the rust binary ([#181](https://github.com/LAPKB/pharmsol/pull/181))

- _(Exa)_ when installing Papir on MacOs, the environment variables are not completely shared to the sandbox in which Papir is running, this changes are meant to provide papir a better way to approach finding the rust binary ([#181](https://github.com/LAPKB/pharmsol/pull/181))

Copilot · 2026-02-13T19:51:21Z

README.md

+let results = subject.nca(&NCAOptions::default(), 0);
+let result = results[0].as_ref().expect("NCA failed");


The example shows calling .nca() with two arguments, but according to the trait definition in src/nca/traits.rs, the method signature is .nca(&self, options: &NCAOptions) -> Result<NCAResult, NCAError> which only takes one argument. This example code will not compile.

Suggested change

let results = subject.nca(&NCAOptions::default(), 0);

let result = results[0].as_ref().expect("NCA failed");

let result = subject.nca(&NCAOptions::default()).expect("NCA failed");

Copilot · 2026-02-13T19:51:21Z

src/simulator/likelihood/prediction.rs

+        };
+        write!(
+            f,
+            "Time: {:.2}\tObs: {:.4}\tPred: {:.4}\tOuteq: {:.2}",


The format specifier for outeq is {:.2} which is for floating point numbers, but outeq is a usize. This should use {} instead for integer formatting.

Suggested change

"Time: {:.2}\tObs: {:.4}\tPred: {:.4}\tOuteq: {:.2}",

"Time: {:.2}\tObs: {:.4}\tPred: {:.4}\tOuteq: {}",

github-actions · 2026-02-13T20:01:22Z

Bencher Report

Branch	feat/new-nca
Testbed	mhovd-pgx

Click to view all benchmark results

Benchmark	Latency	Benchmark Result nanoseconds (ns) (Result Δ%)	Upper Boundary nanoseconds (ns) (Limit %)
Analytical vs ODE/One-compartment IV/Analytical	📈 view plot 🚷 view threshold	837.66 ns (-2.10%) Baseline: 855.62 ns	981.77 ns (85.32%)
Analytical vs ODE/One-compartment IV/ODE	📈 view plot 🚷 view threshold	17,553.00 ns (-0.69%) Baseline: 17,675.00 ns	18,168.42 ns (96.61%)
Analytical vs ODE/One-compartment oral/Analytical	📈 view plot 🚷 view threshold	846.46 ns (-0.59%) Baseline: 851.47 ns	1,025.00 ns (82.58%)
Analytical vs ODE/One-compartment oral/ODE	📈 view plot 🚷 view threshold	25,811.00 ns (-0.03%) Baseline: 25,817.50 ns	26,350.92 ns (97.95%)
Analytical vs ODE/Two-compartment IV/Analytical	📈 view plot 🚷 view threshold	876.97 ns (-5.38%) Baseline: 926.82 ns	1,097.98 ns (79.87%)
Analytical vs ODE/Two-compartment IV/ODE	📈 view plot 🚷 view threshold	26,218.00 ns (-0.16%) Baseline: 26,259.50 ns	26,659.27 ns (98.34%)
Analytical vs ODE/Two-compartment oral/Analytical	📈 view plot 🚷 view threshold	983.02 ns (-0.48%) Baseline: 987.80 ns	1,162.04 ns (84.59%)
Analytical vs ODE/Two-compartment oral/ODE	📈 view plot 🚷 view threshold	29,373.00 ns (+0.04%) Baseline: 29,360.00 ns	30,352.71 ns (96.77%)
Conditional dose modification	📈 view plot 🚷 view threshold	1,194.80 ns (-0.04%) Baseline: 1,195.30 ns	1,218.58 ns (98.05%)
Create large dataset (100 subjects)	📈 view plot 🚷 view threshold	56,271.00 ns (+3.42%) Baseline: 54,410.00 ns	59,323.17 ns (94.86%)
Data expand complex (1h intervals)	📈 view plot 🚷 view threshold	27,160.00 ns (-0.85%) Baseline: 27,391.50 ns	29,191.42 ns (93.04%)
Data expand simple (1h intervals)	📈 view plot 🚷 view threshold	484.58 ns (-0.37%) Baseline: 486.40 ns	503.10 ns (96.32%)
Data expand with additional time	📈 view plot 🚷 view threshold	37,664.00 ns (-0.89%) Baseline: 38,003.25 ns	40,190.82 ns (93.71%)
Filter exclude subjects	📈 view plot 🚷 view threshold	30,876.00 ns (-0.56%) Baseline: 31,049.25 ns	31,675.99 ns (97.47%)
Filter include subjects	📈 view plot 🚷 view threshold	8,206.80 ns (+2.38%) Baseline: 8,015.70 ns	8,661.63 ns (94.75%)
Modify all bolus doses	📈 view plot 🚷 view threshold	1,154.00 ns (-0.60%) Baseline: 1,161.00 ns	1,180.80 ns (97.73%)
Modify all infusion doses	📈 view plot 🚷 view threshold	1,195.60 ns (-0.84%) Baseline: 1,205.70 ns	1,258.39 ns (95.01%)
SubjectBuilder multi-occasion	📈 view plot 🚷 view threshold	259.75 ns (-0.39%) Baseline: 260.75 ns	264.12 ns (98.35%)
SubjectBuilder simple	📈 view plot 🚷 view threshold	102.58 ns (-0.65%) Baseline: 103.25 ns	105.40 ns (97.33%)
SubjectBuilder with covariates	📈 view plot 🚷 view threshold	250.91 ns (-5.57%) Baseline: 265.70 ns	304.55 ns (82.39%)
nca_auc_cmax_metrics	📈 view plot 🚷 view threshold	587.16 ns
nca_lambda_z_candidates	📈 view plot 🚷 view threshold	628.59 ns
nca_population/10	📈 view plot 🚷 view threshold	46,348.00 ns
nca_population/100	📈 view plot 🚷 view threshold	143,860.00 ns
nca_population/500	📈 view plot 🚷 view threshold	318,660.00 ns
nca_single_subject	📈 view plot 🚷 view threshold	995.94 ns
one_compartment	📈 view plot 🚷 view threshold	19,436.00 ns (-1.04%) Baseline: 19,640.00 ns	20,373.66 ns (95.40%)
one_compartment_covariates	📈 view plot 🚷 view threshold	26,793.00 ns (+1.29%) Baseline: 26,452.25 ns	28,856.21 ns (92.85%)
readme 20	📈 view plot 🚷 view threshold	303,060.00 ns (-0.64%) Baseline: 305,007.50 ns	324,290.40 ns (93.45%)
two_compartment	📈 view plot 🚷 view threshold	21,615.00 ns (-0.77%) Baseline: 21,782.50 ns	22,384.73 ns (96.56%)

🐰 View full continuous benchmarking report in Bencher

mhovd · 2026-02-13T20:52:36Z

schemas/model-v1.json

@@ -0,0 +1,792 @@
+{


I feel like we shold be able to define these schemas in Rust using a struc, which is converted to this scheme. There has to be some way of doing that, at least I hope.

The JSON itself can be defined on any language, but the Schema itself (to my knowledge) is always defined in JSON

mhovd · 2026-02-13T20:53:12Z

src/data/auc.rs

+//! # Design
+//!
+//! All functions in this module are **pure math** — no dependency on data structures,
+//! no BLQ filtering, no error types beyond what the caller can check. They accept
+//! raw slices and an [`AUCMethod`], and return `f64`.


This documentation is a little out of place, and seems very LLM-y

mhovd · 2026-02-13T20:55:15Z

src/data/auc.rs

+/// Linear trapezoidal AUC for a single segment
+#[inline]
+fn auc_linear(c1: f64, c2: f64, dt: f64) -> f64 {
+    (c1 + c2) / 2.0 * dt
+}
+
+/// Log-linear AUC for a single segment (assumes c1 > c2 > 0)
+#[inline]
+fn auc_log(c1: f64, c2: f64, dt: f64) -> f64 {
+    (c1 - c2) * dt / (c1 / c2).ln()
+}


Maybe it will be apparent as I dig through the code, but I think it would make more sense to use the times directly instead of the dt. But I may change my mind as I go through the PR.

mhovd · 2026-02-13T20:56:54Z

src/data/auc.rs

+/// calculation cannot know Tmax context. Use [`auc`] or
+/// [`auc_segment_with_tmax`] for proper LinLog handling.
+#[inline]
+pub fn auc_segment(t1: f64, c1: f64, t2: f64, c2: f64, method: &AUCMethod) -> f64 {


This implements Copy, so I don't think it has to be a reference?

Even if the struct implements Copy, passing it by reference is cheaper

mhovd · 2026-02-13T20:57:35Z

src/data/auc.rs

+    let dt = t2 - t1;
+    if dt <= 0.0 {
+        return 0.0;
+    }


Instead of silently returning zero, perhaps return an error?

mhovd · 2026-02-13T21:08:09Z

src/data/builder.rs

+    /// Add an extravascular bolus dose (oral, SC, IM, etc.)
+    ///
+    /// Convenience alias for `.bolus(time, amount, 0)` — targets the depot compartment.
+    ///
+    /// # Arguments
+    ///
+    /// * `time` - Time of the bolus dose
+    /// * `amount` - Amount of drug administered
+    pub fn bolus_ev(self, time: f64, amount: f64) -> Self {
+        self.bolus(time, amount, 0)
+    }
+
+    /// Add an intravenous bolus dose
+    ///
+    /// Convenience alias for `.bolus(time, amount, 1)` — targets the central compartment.
+    ///
+    /// # Arguments
+    ///
+    /// * `time` - Time of the bolus dose
+    /// * `amount` - Amount of drug administered
+    pub fn bolus_iv(self, time: f64, amount: f64) -> Self {
+        self.bolus(time, amount, 1)
+    }
+
+    /// Add an intravenous infusion
+    ///
+    /// Convenience alias for `.infusion(time, amount, 1, duration)` — targets the central compartment.
+    ///
+    /// # Arguments
+    ///
+    /// * `time` - Start time of the infusion
+    /// * `amount` - Total amount of drug to be administered
+    /// * `duration` - Duration of the infusion in time units
+    pub fn infusion_iv(self, time: f64, amount: f64, duration: f64) -> Self {
+        self.infusion(time, amount, 1, duration)
+    }


I don't see the use for these - and they only seem to be used in an example. Consider removing.

A copy before this change is at #209

mhovd · 2026-02-15T14:53:34Z

examples/nca.rs

+        .build();
+
+    let options = NCAOptions::default().with_tau(12.0); // 12-hour dosing interval
+    let result = subject.nca(&options).expect("NCA analysis failed");


I like this API a lot

Siel and others added 19 commits December 30, 2025 14:37

Merge branch 'main' into feat/normalized-row-api

4d3109c

chore: cleanup

1e0c4ed

chore: fmt

f24d525

feat: more validation scenarios

0229425

feat: Json

932150b

chore: wire up the tests, and update them to the newer API

eb99b43

chore: fmt

49e5fc6

Update src/optimize/effect.rs

8f4b155

Co-authored-by: Markus Hovd <markushh@uio.no>

feat: all datasets are now 0 index, Pmetrics is supported but vectors…

07ef9fe

… require extra size

Merge branch 'feat/zero-index' into feat/normalized-row-api

feedeec

Don't use deprecated method

7abf761

Merge branch 'feat/normalized-row-api' into feat/json

86b82e6

making sure both 0-index and 1-index data are supported

f364d76

fix: vector size missmatch

26246a8

wip: adding missing functionality and re-defining base structs

2415f1c

total rewrite of the NCA module

4528d8d

Copilot AI review requested due to automatic review settings February 13, 2026 19:50

Copilot AI reviewed Feb 13, 2026

View reviewed changes

Siel added 2 commits February 13, 2026 19:55

Merge branch 'main' into feat/new-nca

7dcf028

chore: fmt

7ff63ce

mhovd changed the title ~~Feat/new nca~~ feat: Non-compartmental analysis Feb 13, 2026

mhovd reviewed Feb 13, 2026

View reviewed changes

mhovd added 2 commits February 15, 2026 15:45

Remove JSON from this PR

e7376b2

A copy before this change is at #209

Delete json_exa.rs

c36e797

mhovd reviewed Feb 15, 2026

View reviewed changes

Move tests (#210)

7832f67

	let outeq = obs.outeq().to_string();
	let outeq = (obs.outeq() + 1).to_string();

	- _(Exa)_ when installing Papir on MacOs, the environment varaibles are not completly shared to the sandbox in which Papir is running, this changes are meant to provide papir a better way to approach finding the rust binary ([#181](https://github.com/LAPKB/pharmsol/pull/181))
	- _(Exa)_ when installing Papir on MacOs, the environment variables are not completely shared to the sandbox in which Papir is running, this changes are meant to provide papir a better way to approach finding the rust binary ([#181](https://github.com/LAPKB/pharmsol/pull/181))

		let results = subject.nca(&NCAOptions::default(), 0);
		let result = results[0].as_ref().expect("NCA failed");

	let results = subject.nca(&NCAOptions::default(), 0);
	let result = results[0].as_ref().expect("NCA failed");
	let result = subject.nca(&NCAOptions::default()).expect("NCA failed");

	"Time: {:.2}\tObs: {:.4}\tPred: {:.4}\tOuteq: {:.2}",
	"Time: {:.2}\tObs: {:.4}\tPred: {:.4}\tOuteq: {}",

Conversation

Siel commented Feb 13, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Feb 13, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 13, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 13, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 13, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 13, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 13, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Feb 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

github-actions bot commented Feb 13, 2026 •

edited

Loading