Standard Library

The Tolvex standard library is implemented as a set of Rust crates under stdlib/.

Modules

• tolvex_data: healthcare data structures and ingestion helpers
• tolvex_stats: basic statistics (mean, stddev, Welch t-test)
• tolvex_compliance: HIPAA/compliance primitives
• tolvex_ai: minimal risk scorer and utilities
• tolvex_model: model interfaces and stubs

API overview (Rust host)

The following examples show how the current stdlib crates are consumed from Rust. This mirrors how Tolvex's runtime and compiler integrate today.

tolvex_data

• HL7 ingestion: tolvex_data::hl7_fhir::{hl7_to_fhir_patient_minimal, hl7_to_medical_event_minimal}
• Observation ingestion: tolvex_data::hl7_fhir_obx::hl7_to_observations_minimal
• Sanitization: tolvex_data::sanitize::{trim_patient_names, normalize_patient_birth_date, canonize_observation_code, normalize_observation_unit}
• Validation: tolvex_data::validate::{validate_patient, validate_observation, validate_medical_event}
• Query builder: tolvex_data::fhir_query("...")

use tolvex_data::hl7_fhir::hl7_to_fhir_patient_minimal;
use tolvex_data::hl7_fhir_obx::hl7_to_observations_minimal;
use tolvex_data::sanitize::{canonize_observation_code, normalize_observation_unit, trim_patient_names};
use tolvex_data::validate::{validate_observation, validate_patient};

let mut patient = hl7_to_fhir_patient_minimal(hl7_msg)?;
trim_patient_names(&mut patient);
validate_patient(&patient)?;

let mut observations = hl7_to_observations_minimal(hl7_msg)?;
for o in &mut observations {
    canonize_observation_code(o);
    normalize_observation_unit(o);
    validate_observation(o)?;
}

tolvex_stats

• Descriptive stats: mean, median, stddev_sample
• Inferential: t_test_welch

use tolvex_stats::{mean, stddev_sample, t_test_welch};

let a = [1.0, 2.0, 3.0, 4.0];
let b = [2.0, 3.0, 4.0, 5.0];
let mu_a = mean(&a);
let sd_a = stddev_sample(&a);
let res = t_test_welch(&a, &b);
println!("mean={mu_a} sd={sd_a} t={:.3} df={:.1}", res.t_stat, res.df);

tolvex_compliance

• Keyword checks: HipaaKeywordRule, run_hipaa_bundle
• Summaries: build_compliance_report_summary

use tolvex_compliance::{
    build_compliance_report_summary, run_hipaa_bundle, ComplianceProfile, HipaaKeywordRule,
    ReportKind, RuleSeverity,
};

let rules = vec![HipaaKeywordRule {
    id: "k1".into(),
    description: "Detect SSN".into(),
    keywords: vec!["ssn".into(), "social security".into()],
    severity: RuleSeverity::Error,
}];

let report = run_hipaa_bundle(
    "Patient SSN: 123-45-6789",
    &rules,
    None,
    None,
    None,
    Some(400),
    &[],
);

let summary = build_compliance_report_summary(
    ReportKind::Fda21Cfr11,
    Some(ComplianceProfile::Hipaa),
    &report.keyword_results,
);
println!("passed={}", summary.summary.passed);

tolvex_ai

• Scoring: RiskScorer::predict
• End-to-end helpers: predict_risk, stratify_risk

use tolvex_ai::{predict_risk, stratify_risk, RiskScorer};
use serde_json::json;

let model = RiskScorer {
    weights: vec![0.2, 0.5, 0.3],
    bias: 0.1,
    model_name: "demo".into(),
};
let score = model.predict(&[0.4, 0.2, 0.9]);
println!("score={score}");

let patient = json!({"age": 65, "bmi": 33.1, "a1c": 7.8});
let pred = predict_risk(&patient, "diabetes", "5y", &["age".into(), "bmi".into(), "a1c".into()]);
let stratum = stratify_risk(&pred);
println!("risk={:.3} stratum={:?}", pred.score, stratum);

tolvex_model

• Registry: ModelRegistry
• Metadata: ModelMetadata

use tolvex_model::{ModelMetadata, ModelRegistry};

let mut reg = ModelRegistry::default();
reg.register(ModelMetadata {
    id: "risk_demo".into(),
    name: "Risk Demo".into(),
    version: "0.1.0".into(),
    task: "risk".into(),
    input_schema: None,
    output_schema: None,
});
assert!(reg.get("risk_demo").is_some());

Running tests

cargo test -p tolvex_data
cargo test -p tolvex_stats
cargo test -p tolvex_compliance
cargo test -p tolvex_ai
cargo test -p tolvex_model

Benchmarks

Some stdlib crates include Criterion benchmarks.

cargo bench -p tolvex_data
cargo bench -p tolvex_compliance