fsci-conformance 0.3.0

#![forbid(unsafe_code)]
//! Live SciPy differential coverage for the Epps-Singleton
//! two-sample test `scipy.stats.epps_singleton_2samp(x, y)`.
//!
//! Resolves [frankenscipy-ll11q]. The test compares two
//! samples by matching characteristic-function moments at
//! the default `t = (0.4, 0.8)` evaluation points; the
//! resulting test statistic is asymptotically chi-squared
//! distributed with df=4.
//!
//! 4 (x, y) fixtures × 2 arms (statistic + pvalue) = 8 cases
//! via subprocess. Tol 1e-9 abs.

use std::collections::{BTreeMap, HashMap};
use std::fs;
use std::io::Write;
use std::path::PathBuf;
use std::process::Stdio;
use std::time::{Instant, SystemTime, UNIX_EPOCH};

use fsci_conformance::{ArmCounts, CompareLedger};
use fsci_stats::epps_singleton_2samp;
use serde::{Deserialize, Serialize};

const PACKET_ID: &str = "FSCI-P2C-007";
const ABS_TOL: f64 = 1.0e-9;
const REQUIRE_SCIPY_ENV: &str = "FSCI_REQUIRE_SCIPY_ORACLE";

#[derive(Debug, Clone, Serialize)]
struct PointCase {
    case_id: String,
    x: Vec<f64>,
    y: Vec<f64>,
}

#[derive(Debug, Clone, Serialize)]
struct OracleQuery {
    points: Vec<PointCase>,
}

#[derive(Debug, Clone, Deserialize)]
struct PointArm {
    case_id: String,
    statistic: Option<f64>,
    pvalue: Option<f64>,
}

#[derive(Debug, Clone, Deserialize)]
struct OracleResult {
    points: Vec<PointArm>,
}

#[derive(Debug, Clone, Serialize)]
struct CaseDiff {
    case_id: String,
    arm: String,
    abs_diff: f64,
    pass: bool,
}

#[derive(Debug, Clone, Serialize)]
struct DiffLog {
    test_id: String,
    category: String,
    case_count: usize,
    compared: BTreeMap<String, ArmCounts>,
    max_abs_diff: f64,
    pass: bool,
    timestamp_ms: u128,
    duration_ns: u128,
    cases: Vec<CaseDiff>,
}

fn output_dir() -> PathBuf {
    PathBuf::from(env!("CARGO_MANIFEST_DIR")).join(format!("fixtures/artifacts/{PACKET_ID}/diff"))
}

fn ensure_output_dir() {
    fs::create_dir_all(output_dir()).expect("create epps_singleton diff output dir");
}

fn timestamp_ms() -> u128 {
    SystemTime::now()
        .duration_since(UNIX_EPOCH)
        .map_or(0, |d| d.as_millis())
}

fn emit_log(log: &DiffLog) {
    ensure_output_dir();
    let path = output_dir().join(format!("{}.json", log.test_id));
    let json = serde_json::to_string_pretty(log).expect("serialize epps_singleton diff log");
    fs::write(path, json).expect("write epps_singleton diff log");
}

fn generate_query() -> OracleQuery {
    let fixtures: Vec<(&str, Vec<f64>, Vec<f64>)> = vec![
        // Mean-shifted
        (
            "mean_shift",
            (1..=20).map(|i| i as f64 / 5.0).collect(),
            (1..=20).map(|i| i as f64 / 5.0 + 1.5).collect(),
        ),
        // Scale-shifted
        (
            "scale_shift",
            (-10..=10).map(|i| i as f64 / 5.0).collect(),
            (-10..=10).map(|i| i as f64 / 2.0).collect(),
        ),
        // Bimodal vs unimodal
        (
            "shape_diff",
            (-10..=10).map(|i| i as f64 / 5.0).collect(),
            vec![
                -2.0, -1.8, -1.6, -1.4, -1.2, -1.0, 1.0, 1.2, 1.4, 1.6, 1.8, 2.0,
            ],
        ),
        // Larger sample
        (
            "larger",
            (0..30).map(|i| i as f64 / 10.0).collect(),
            (0..30).map(|i| 0.5 + i as f64 / 10.0).collect(),
        ),
    ];

    let points = fixtures
        .into_iter()
        .map(|(name, x, y)| PointCase {
            case_id: name.into(),
            x,
            y,
        })
        .collect();
    OracleQuery { points }
}

fn scipy_oracle_or_skip(query: &OracleQuery) -> Option<OracleResult> {
    let script = r#"
import json
import math
import sys
import numpy as np
from scipy import stats

def fnone(v):
    try:
        v = float(v)
    except Exception:
        return None
    return v if math.isfinite(v) else None

q = json.load(sys.stdin)
points = []
for case in q["points"]:
    cid = case["case_id"]
    x = np.array(case["x"], dtype=float)
    y = np.array(case["y"], dtype=float)
    try:
        res = stats.epps_singleton_2samp(x, y)
        points.append({
            "case_id": cid,
            "statistic": fnone(res.statistic),
            "pvalue": fnone(res.pvalue),
        })
    except Exception:
        points.append({"case_id": cid, "statistic": None, "pvalue": None})
print(json.dumps({"points": points}))
"#;
    let query_json = serde_json::to_string(query).expect("serialize epps_singleton query");
    let mut child = match fsci_conformance::scipy_oracle_command()
        .arg("-c")
        .arg(script)
        .stdin(Stdio::piped())
        .stdout(Stdio::piped())
        .stderr(Stdio::piped())
        .spawn()
    {
        Ok(c) => c,
        Err(e) => {
            assert!(
                std::env::var(REQUIRE_SCIPY_ENV).is_err(),
                "failed to spawn python3 for epps_singleton oracle: {e}"
            );
            eprintln!("skipping epps_singleton oracle: python3 not available ({e})");
            return None;
        }
    };
    {
        let stdin = child
            .stdin
            .as_mut()
            .expect("open epps_singleton oracle stdin");
        if let Err(err) = stdin.write_all(query_json.as_bytes()) {
            let output = child.wait_with_output().expect("wait for failed oracle");
            let stderr = String::from_utf8_lossy(&output.stderr);
            assert!(
                std::env::var(REQUIRE_SCIPY_ENV).is_err(),
                "epps_singleton oracle stdin write failed: {err}; stderr: {stderr}"
            );
            eprintln!("skipping epps_singleton oracle: stdin write failed ({err})\n{stderr}");
            return None;
        }
    }
    let output = child
        .wait_with_output()
        .expect("wait for epps_singleton oracle");
    if !output.status.success() {
        let stderr = String::from_utf8_lossy(&output.stderr);
        assert!(
            std::env::var(REQUIRE_SCIPY_ENV).is_err(),
            "epps_singleton oracle failed: {stderr}"
        );
        eprintln!("skipping epps_singleton oracle: scipy not available\n{stderr}");
        return None;
    }
    let stdout = String::from_utf8_lossy(&output.stdout);
    Some(serde_json::from_str(&stdout).expect("parse epps_singleton oracle JSON"))
}

#[test]
fn diff_stats_epps_singleton() {
    let query = generate_query();
    let Some(oracle) = scipy_oracle_or_skip(&query) else {
        return;
    };
    assert_eq!(oracle.points.len(), query.points.len());

    let pmap: HashMap<String, PointArm> = oracle
        .points
        .into_iter()
        .map(|r| (r.case_id.clone(), r))
        .collect();

    let start = Instant::now();
    let mut diffs = Vec::new();
    let mut max_overall = 0.0_f64;
    let mut ledger = CompareLedger::new("diff_stats_epps_singleton", &["statistic", "pvalue"]);

    for case in &query.points {
        let scipy_arm = pmap.get(&case.case_id).expect("validated oracle");
        let result = epps_singleton_2samp(&case.x, &case.y);

        let arms = [
            ("statistic", scipy_arm.statistic, result.statistic),
            ("pvalue", scipy_arm.pvalue, result.pvalue),
        ];
        for (arm, scipy, fsci) in arms {
            let Some((s, f)) = ledger.pair(arm, &case.case_id, scipy, Some(fsci)) else {
                continue;
            };
            let abs_diff = (f - s).abs();
            max_overall = max_overall.max(abs_diff);
            ledger.compared(arm, &case.case_id, abs_diff <= ABS_TOL);
            diffs.push(CaseDiff {
                case_id: case.case_id.clone(),
                arm: arm.into(),
                abs_diff,
                pass: abs_diff <= ABS_TOL,
            });
        }
    }

    let all_pass = diffs.iter().all(|d| d.pass);

    let log = DiffLog {
        test_id: "diff_stats_epps_singleton".into(),
        category: "scipy.stats.epps_singleton_2samp".into(),
        case_count: diffs.len(),
        compared: ledger.counts().clone(),
        max_abs_diff: max_overall,
        pass: all_pass,
        timestamp_ms: timestamp_ms(),
        duration_ns: start.elapsed().as_nanos(),
        cases: diffs.clone(),
    };

    emit_log(&log);

    for d in &diffs {
        if !d.pass {
            eprintln!(
                "epps_singleton mismatch: {} arm={} abs={}",
                d.case_id, d.arm, d.abs_diff
            );
        }
    }

    assert!(
        all_pass,
        "epps_singleton conformance failed: {} cases, max_abs={}",
        diffs.len(),
        max_overall
    );
    ledger.finish(query.points.len());
}