@iyulab/u-insight 0.8.0 → 0.9.0

package/README.md

@@ -7,6 +7,19 @@
 
 A statistical analysis and data profiling engine in Rust with C FFI bindings.
 
+## What's New in 0.9.0
+
+- **Kendall tau-b correlation** added to `CorrelationMethod` (Pearson / Spearman / Kendall)
+- **Outlier fences exposed** on `OutlierResult` — `lower_fence`, `upper_fence`, `center`, `spread`
+- **`detect_outliers_slice(&[f64], method)`** helper for raw-slice input
+- **`vif_analysis()` and `condition_number()`** standalone multicollinearity diagnostics
+- **WASM**: `correlation_matrix` accepts optional `_method` field (`"pearson"` | `"spearman"` | `"kendall"`); new `detect_univariate_outliers`, `vif_diagnostic`, `condition_number_diagnostic`
+- **C#**: `CorrelationMethodKind` enum + `Correlate(data, method)` parameter
+- **BREAKING — Rust**: Non-finite numeric inputs now return `InsightError::DegenerateData` (was `NonNumericColumn`); audit covered 11 call sites
+- **BREAKING — Rust**: `OutlierResult` gained 4 new fields (exhaustive pattern matches must be updated)
+- **BREAKING — FFI**: `insight_correlation` signature gained a `method: u32` parameter (use `INSIGHT_CORR_PEARSON` = 0 to keep prior behaviour)
+- **BREAKING — C#**: `Correlate(...)` now takes an optional `CorrelationMethodKind` parameter (default `Pearson` keeps existing call-sites compiling)
+
 ## Overview
 
 u-insight transforms raw tabular data into actionable statistical insights. It operates in two distinct layers with **opposite assumptions about input data quality**:

package/package.json

@@ -5,7 +5,7 @@
     "iyulab"
   ],
   "description": "Statistical analysis and data profiling engine with C FFI bindings.",
-  "version": "0.8.0",
+  "version": "0.9.0",
   "license": "MIT",
   "repository": {
     "type": "git",

package/u_insight.d.ts

@@ -2,13 +2,33 @@
 /* eslint-disable */
 
 /**
- * Computes a Pearson correlation matrix for a column-major dataset.
+ * 2-norm condition number of the sample covariance matrix.
  *
  * # Input
  * ```json
- * { "col1": [1.0, 2.0, 3.0], "col2": [4.0, 5.0, 6.0] }
+ * { "col1": [...], "col2": [...] }
+ * ```
+ *
+ * # Output
+ * `{ condition_number, names }`
+ *
+ * Standard threshold: `cond > 30` indicates multicollinearity (Belsley 1991).
+ * Returns `Infinity` for numerically singular input.
+ */
+export function condition_number_diagnostic(data_json: any): any;
+
+/**
+ * Computes a correlation matrix for a column-major dataset.
+ *
+ * # Input
+ * ```json
+ * { "col1": [1.0, 2.0, 3.0], "col2": [4.0, 5.0, 6.0], "_method": "pearson" }
  * ```
  *
+ * `_method` ∈ `{"pearson", "spearman", "kendall"}` — optional, defaults
+ * to `"pearson"`. Reserved key (prefix `_`) so it never collides with a
+ * column name.
+ *
  * # Output
  * `{ names, matrix (flattened n×n), n, high_pairs }`
  */
@@ -49,6 +69,24 @@ export function dbscan(data_json: any, config_json: any): any;
  */
 export function describe(data_json: any): any;
 
+/**
+ * Univariate outlier detection on a flat numeric vector.
+ *
+ * # Input
+ * ```json
+ * { "data": [1.0, 2.0, 3.0, 100.0], "method": "iqr" }
+ * ```
+ * `method` ∈ `{"iqr"|"tukey", "zscore"|"three_sigma", "modified_zscore"|"hampel"}`.
+ * Optional, defaults to `"iqr"`.
+ *
+ * # Output
+ * `{ method, indices, scores, count, pct, lower_fence, upper_fence, center, spread }`
+ *
+ * `method` aliases: `tukey` → IQR Tukey fences (k=1.5), `three_sigma` →
+ * mean ± 3·σ, `hampel` → robust median ± 3.5·(MAD/0.6745).
+ */
+export function detect_univariate_outliers(data_json: any): any;
+
 /**
  * Runs distribution analysis on a 1-D numeric array.
  *
@@ -182,3 +220,17 @@ export function pca(data_json: any, n_components: number): any;
  * `{ target_name, predictor_names, r_squared, adj_r_squared, coefficients, p_values, vif, f_p_value }`
  */
 export function regression(data_json: any): any;
+
+/**
+ * Variance Inflation Factor diagnostics for column-major numeric data.
+ *
+ * # Input
+ * ```json
+ * { "col1": [...], "col2": [...], "_threshold": 10.0 }
+ * ```
+ * `_threshold` is optional (default 10.0).
+ *
+ * # Output
+ * `{ vif_per_column, high_vif_columns, threshold, names }`
+ */
+export function vif_diagnostic(data_json: any): any;

package/u_insight.js

@@ -5,5 +5,5 @@ import { __wbg_set_wasm } from "./u_insight_bg.js";
 __wbg_set_wasm(wasm);
 wasm.__wbindgen_start();
 export {
-    correlation_matrix, dbscan, describe, distribution_analysis, feature_importance, hierarchical, isolation_forest, kmeans, lof, pca, regression
+    condition_number_diagnostic, correlation_matrix, dbscan, describe, detect_univariate_outliers, distribution_analysis, feature_importance, hierarchical, isolation_forest, kmeans, lof, pca, regression, vif_diagnostic
 } from "./u_insight_bg.js";

package/u_insight_bg.js

@@ -1,11 +1,39 @@
 /**
- * Computes a Pearson correlation matrix for a column-major dataset.
+ * 2-norm condition number of the sample covariance matrix.
  *
  * # Input
  * ```json
- * { "col1": [1.0, 2.0, 3.0], "col2": [4.0, 5.0, 6.0] }
+ * { "col1": [...], "col2": [...] }
+ * ```
+ *
+ * # Output
+ * `{ condition_number, names }`
+ *
+ * Standard threshold: `cond > 30` indicates multicollinearity (Belsley 1991).
+ * Returns `Infinity` for numerically singular input.
+ * @param {any} data_json
+ * @returns {any}
+ */
+export function condition_number_diagnostic(data_json) {
+    const ret = wasm.condition_number_diagnostic(data_json);
+    if (ret[2]) {
+        throw takeFromExternrefTable0(ret[1]);
+    }
+    return takeFromExternrefTable0(ret[0]);
+}
+
+/**
+ * Computes a correlation matrix for a column-major dataset.
+ *
+ * # Input
+ * ```json
+ * { "col1": [1.0, 2.0, 3.0], "col2": [4.0, 5.0, 6.0], "_method": "pearson" }
  * ```
  *
+ * `_method` ∈ `{"pearson", "spearman", "kendall"}` — optional, defaults
+ * to `"pearson"`. Reserved key (prefix `_`) so it never collides with a
+ * column name.
+ *
  * # Output
  * `{ names, matrix (flattened n×n), n, high_pairs }`
  * @param {any} data_json
@@ -71,6 +99,32 @@ export function describe(data_json) {
     return takeFromExternrefTable0(ret[0]);
 }
 
+/**
+ * Univariate outlier detection on a flat numeric vector.
+ *
+ * # Input
+ * ```json
+ * { "data": [1.0, 2.0, 3.0, 100.0], "method": "iqr" }
+ * ```
+ * `method` ∈ `{"iqr"|"tukey", "zscore"|"three_sigma", "modified_zscore"|"hampel"}`.
+ * Optional, defaults to `"iqr"`.
+ *
+ * # Output
+ * `{ method, indices, scores, count, pct, lower_fence, upper_fence, center, spread }`
+ *
+ * `method` aliases: `tukey` → IQR Tukey fences (k=1.5), `three_sigma` →
+ * mean ± 3·σ, `hampel` → robust median ± 3.5·(MAD/0.6745).
+ * @param {any} data_json
+ * @returns {any}
+ */
+export function detect_univariate_outliers(data_json) {
+    const ret = wasm.detect_univariate_outliers(data_json);
+    if (ret[2]) {
+        throw takeFromExternrefTable0(ret[1]);
+    }
+    return takeFromExternrefTable0(ret[0]);
+}
+
 /**
  * Runs distribution analysis on a 1-D numeric array.
  *
@@ -274,6 +328,28 @@ export function regression(data_json) {
     }
     return takeFromExternrefTable0(ret[0]);
 }
+
+/**
+ * Variance Inflation Factor diagnostics for column-major numeric data.
+ *
+ * # Input
+ * ```json
+ * { "col1": [...], "col2": [...], "_threshold": 10.0 }
+ * ```
+ * `_threshold` is optional (default 10.0).
+ *
+ * # Output
+ * `{ vif_per_column, high_vif_columns, threshold, names }`
+ * @param {any} data_json
+ * @returns {any}
+ */
+export function vif_diagnostic(data_json) {
+    const ret = wasm.vif_diagnostic(data_json);
+    if (ret[2]) {
+        throw takeFromExternrefTable0(ret[1]);
+    }
+    return takeFromExternrefTable0(ret[0]);
+}
 export function __wbg_Error_83742b46f01ce22d(arg0, arg1) {
     const ret = Error(getStringFromWasm0(arg0, arg1));
     return ret;