[test] expand diagnostics report coverage

python/tests/test_diagnostics/test_report.py

@@ -0,0 +1,141 @@
+from __future__ import annotations
+
+import json
+from types import SimpleNamespace
+
+import numpy as np
+
+from hyperion_diagnostics.report import DiagnosticsReport
+from hyperion_diagnostics.report import generate_report
+
+
+def test_diagnostics_report_json_serializes_core_fields() -> None:
+    report = DiagnosticsReport(
+        model_name="normal_model",
+        inference_method="hmc",
+        timestamp="2026-05-09T10:00:00",
+        config={"num_samples": 100},
+        summary_stats={"mu": {"mean": 0.1, "std": 1.2}},
+        convergence_metrics={"accept_rate": 0.8},
+    )
+
+    payload = json.loads(report.to_json())
+
+    assert payload["model_name"] == "normal_model"
+    assert payload["inference_method"] == "hmc"
+    assert payload["config"]["num_samples"] == 100
+    assert payload["summary_stats"]["mu"]["mean"] == 0.1
+
+
+def test_diagnostics_report_markdown_renders_configuration_and_summary_table() -> None:
+    report = DiagnosticsReport(
+        model_name="coin_model",
+        inference_method="nuts",
+        timestamp="2026-05-09T10:00:00",
+        config={"num_warmup": 50, "num_samples": 100},
+        summary_stats={
+            "mu": {
+                "mean": 2.0,
+                "std": 0.5,
+                "median": 2.1,
+                "ci_5.0%": 1.2,
+                "ci_95.0%": 2.8,
+                "ess": 88.0,
+            }
+        },
+        convergence_metrics={"mu/ess": 88.0},
+    )
+
+    markdown = report.to_markdown()
+
+    assert "# Diagnostics Report: coin_model" in markdown
+    assert "- **num_warmup:** 50" in markdown
+    assert "| Parameter | mean | std | median | ci_5.0% | ci_95.0% | ess |" in markdown
+    assert "| mu | 2.0000 | 0.5000 | 2.1000 | 1.2000 | 2.8000 | 88.0000 |" in markdown
+
+
+def test_generate_report_adds_clean_conclusion_when_diagnostics_pass() -> None:
+    rng = np.random.default_rng(5)
+    result = SimpleNamespace(
+        samples={"mu": rng.normal(size=500)},
+        diagnostics={"accept_rate": 0.75, "num_divergences": 0},
+    )
+
+    report = generate_report(
+        result,
+        model_name="normal_model",
+        inference_method="hmc",
+        config={"num_samples": 500},
+    )
+
+    assert report.model_name == "normal_model"
+    assert report.warnings == []
+    assert report.conclusions == ["No convergence issues detected. Results appear reliable."]
+    assert "mu" in report.summary_stats
+
+
+def test_generate_report_warns_on_low_effective_sample_size() -> None:
+    result = SimpleNamespace(
+        samples={"mu": np.ones(150)},
+        diagnostics={"accept_rate": 0.8, "num_divergences": 0},
+    )
+
+    report = generate_report(result, model_name="sticky_model", inference_method="hmc")
+
+    assert any("Low ESS for mu/ess" in warning for warning in report.warnings)
+    assert report.conclusions == [
+        "1 potential issue(s) detected. Review warnings before trusting results."
+    ]
+
+
+def test_generate_report_warns_on_sampler_diagnostics() -> None:
+    rng = np.random.default_rng(7)
+    result = SimpleNamespace(
+        samples={"mu": rng.normal(size=500)},
+        diagnostics={
+            "accept_rate": 0.4,
+            "num_divergences": 2,
+            "energy": np.cumsum(rng.normal(0.0, 0.01, size=500)),
+        },
+    )
+
+    report = generate_report(result, model_name="difficult_model", inference_method="nuts")
+
+    assert any("Low acceptance rate" in warning for warning in report.warnings)
+    assert any("divergent transitions" in warning for warning in report.warnings)
+    assert any("Low BFMI" in warning for warning in report.warnings)
+
+
+def test_generate_report_includes_multichain_vector_parameters() -> None:
+    rng = np.random.default_rng(11)
+    chains = rng.normal(size=(3, 120, 2))
+    result = SimpleNamespace(
+        samples={"beta": chains.reshape(-1, 2)},
+        diagnostics={"accept_rate": 0.9, "num_divergences": 0},
+        num_chains=3,
+        samples_by_chain={"beta": chains},
+    )
+
+    report = generate_report(result, model_name="linear_model", inference_method="hmc")
+
+    assert "beta[0]" in report.summary_stats
+    assert "beta[1]" in report.summary_stats
+    assert "beta[0]/r_hat" in report.convergence_metrics
+    assert "beta[1]/split_r_hat" in report.convergence_metrics
+
+
+def test_diagnostics_report_markdown_renders_warnings_and_conclusions() -> None:
+    report = DiagnosticsReport(
+        model_name="warning_model",
+        inference_method="vi",
+        timestamp="2026-05-09T10:00:00",
+        warnings=["Low ESS for z/ess: 30.0 (recommended > 100)"],
+        conclusions=["Review warnings before trusting results."],
+    )
+
+    markdown = report.to_markdown()
+
+    assert "## Warnings" in markdown
+    assert "Low ESS for z/ess" in markdown
+    assert "## Conclusions" in markdown
+    assert "- Review warnings before trusting results." in markdown