feat: v0.9.14 — summary() renders ATT/ATU block (completes v0.9.13 spec §3.3)

Tiny patch release closing the remaining thread on v0.9.13: the six
att_sd / att_hdi_* / atu_sd / atu_hdi_* fields landed but
summary() never printed them. v0.9.14 wires the rendering through a
BayesianMTEResult.summary() override.

Added:
- BayesianMTEResult.summary() override appends a
  'Population-integrated effects' block inside the framing = ruler:

      ATT: 0.2407 (sd 0.0370, 95% HDI [0.1693, 0.3136])
      ATU: 0.2147 (sd 0.0435, 95% HDI [0.1341, 0.2947])

  Silently skipped when SDs are NaN (empty subpopulation or
  pre-v0.9.13 deserialised result).

Round B review: no blockers. Reviewer confirmed:
- base.endswith('=' * 70) is exact (parent returns
  '\n'.join(lines) with rule as final element, no trailing
  newline or rstrip).
- Block splicing preserves the closing ruler visually.
- NaN stub path is safe; fallback branch is defensive.
- 'ATT:' / 'ATU:' are unique substrings — no parent collision.
- Pure reader, thread-safe.

Tests:
- tests/test_bayes_mte_uncertainty.py now has two new tests:
  test_summary_shows_att_atu_uncertainty (post-fit: string contains
  'ATT:', 'ATU:', 'sd ', 'HDI [') and test_summary_skips_att_atu_when_nan
  (stub with NaN SDs → no 'ATT:' / 'ATU:' in output).
- Bayesian family suite: 88/88 focused MTE + sibling tests green
  in 1:55.

Also rides along with parallel-work imports from auto-hooks:
sp.rosenbaum_bounds / rosenbaum_gamma (diagnostics) and
sp.negative_control_outcome / negative_control_exposure /
double_negative_control / NegativeControlResult (proximal).
Neither is documented in this CHANGELOG; they ship via the git log
as parallel contributions.

Design spec: docs/superpowers/specs/2026-04-20-v0913-hdi-compat-and-att-sd.md §3.3

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: brycewang-stanford

CHANGELOG.md

@@ -2,6 +2,55 @@
 
 All notable changes to StatsPAI will be documented in this file.
 
+## [0.9.14] - 2026-04-20 — Summary rendering completes v0.9.13 spec §3.3
+
+Tiny patch release. Completes the "ATT/ATU in `summary()`" promise
+from v0.9.13 spec §3.3 that was not actually wired at ship time
+(the six uncertainty fields landed but `summary()` never printed
+them).
+
+### Added (0.9.14)
+
+- **`BayesianMTEResult.summary()`** override. Extends
+  `BayesianCausalResult.summary` with a `Population-integrated effects`
+  block:
+
+      ATT: 0.2407 (sd 0.0370, 95% HDI [0.1693, 0.3136])
+      ATU: 0.2147 (sd 0.0435, 95% HDI [0.1341, 0.2947])
+
+  Rendered inside the framing `=` ruler for visual coherence.
+  Silently skipped when either SD is NaN (empty subpopulation or
+  pre-v0.9.13 deserialised result).
+
+### Round-B review: no blockers
+
+Reviewer confirmed:
+1. `base.endswith('=' * 70)` is exact — parent `summary()` returns
+   `'\n'.join(lines)` with the rule as the final element.
+2. Block splicing preserves the closing ruler visually.
+3. NaN stub path is safe; fallback branch is defensive.
+4. `'ATT:'` / `'ATU:'` are unique substrings — no collision with
+   parent output.
+5. Pure reader; thread-safe.
+
+### Tests (0.9.14)
+
+- `tests/test_bayes_mte_uncertainty.py` now has:
+  - `test_summary_shows_att_atu_uncertainty` — after fit, string
+    contains `'ATT:'`, `'ATU:'`, `'sd '`, `'HDI ['`.
+  - `test_summary_skips_att_atu_when_nan` — NaN-SD stub → no
+    `'ATT:'` / `'ATU:'` in output.
+- Full Bayesian suite: 88/88 focused MTE + sibling green in 1:55.
+
+### Non-goals (0.9.14)
+
+- `.tidy()` multi-row variant with ATE/ATT/ATU as separate rows
+  — queued for v0.9.15+.
+- Full bivariate-normal HV model.
+- Rust Phase 2.
+
+---
+
 ## [0.9.13] - 2026-04-20 — ArviZ HDI compat shim + ATT/ATU uncertainty
 
 Small-but-load-bearing cleanup release. Closes two items deferred

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "StatsPAI"
-version = "0.9.13"
+version = "0.9.14"
 description = "The Agent-Native Causal Inference & Econometrics Toolkit for Python"
 readme = "README.md"
 license = {text = "MIT"}

src/statspai/__init__.py

@@ -22,7 +22,7 @@
 >>> sp.outreg2(result, filename="results.xlsx")
 """
 
-__version__ = "0.9.13"
+__version__ = "0.9.14"
 __author__ = "Biaoyue Wang"
 __email__ = "brycew6m@stanford.edu"
 
@@ -96,7 +96,7 @@
 from .output.regression_table import regtable, RegtableResult, mean_comparison, MeanComparisonResult
 from .output.paper_tables import paper_tables, PaperTables, TEMPLATES as PAPER_TABLE_TEMPLATES
 from .postestimation import margins, marginsplot, margins_at, margins_at_plot, contrast, pwcompare, test, lincom
-from .diagnostics import oster_bounds, mccrary_test, diagnose, het_test, reset_test, vif, sensemakr, rddensity, hausman_test, anderson_rubin_test, effective_f_test, tF_critical_value, evalue, evalue_from_result, diagnose_result, estat, kitagawa_test, KitagawaResult
+from .diagnostics import oster_bounds, mccrary_test, diagnose, het_test, reset_test, vif, sensemakr, rddensity, hausman_test, anderson_rubin_test, effective_f_test, tF_critical_value, evalue, evalue_from_result, diagnose_result, estat, kitagawa_test, KitagawaResult, rosenbaum_bounds, rosenbaum_gamma, RosenbaumResult
 from .inference import (
     wild_cluster_bootstrap, aipw, ri_test, ipw, bootstrap, BootstrapResult,
     twoway_cluster, conley, pate, PATEEstimator, fisher_exact, FisherResult,
@@ -106,7 +106,11 @@
     g_computation, front_door,
 )
 from .msm import msm, MarginalStructuralModel, stabilized_weights
-from .proximal import proximal, ProximalCausalInference
+from .proximal import (
+    proximal, ProximalCausalInference,
+    negative_control_outcome, negative_control_exposure,
+    double_negative_control, NegativeControlResult,
+)
 from .principal_strat import (
     principal_strat, PrincipalStratResult, survivor_average_causal_effect,
 )
@@ -156,7 +160,7 @@
 from .regression.glm import glm, GLMRegression, GLMEstimator
 from .regression.count import poisson, nbreg, ppmlhdfe
 from .neural_causal import tarnet, cfrnet, dragonnet, TARNet, CFRNet, DragonNet
-from .causal_discovery import notears, NOTEARS, pc_algorithm, PCAlgorithm, lingam, LiNGAMResult, ges, GESResult
+from .causal_discovery import notears, NOTEARS, pc_algorithm, PCAlgorithm, lingam, LiNGAMResult, ges, GESResult, fci, FCIResult
 from .tmle import tmle, TMLE, super_learner, SuperLearner
 from .policy_learning import policy_tree, PolicyTree, policy_value
 from .conformal_causal import conformal_cate, ConformalCATE

src/statspai/bayes/_base.py

@@ -426,6 +426,38 @@ class BayesianMTEResult(BayesianCausalResult):
     atu_hdi_lower: float = float('nan')
     atu_hdi_upper: float = float('nan')
 
+    def summary(self) -> str:
+        """Printable summary with ATT/ATU uncertainty appended.
+
+        Extends ``BayesianCausalResult.summary`` with a block printing
+        ``ATT`` / ``ATU`` posterior mean, SD and HDI when the SD
+        fields are finite. Silently skipped (parent summary only) when
+        either side of the population is empty or the result was
+        deserialised from a pre-v0.9.13 snapshot (fields default NaN).
+        """
+        base = super().summary()
+        extra: list[str] = []
+        pct = int(self.hdi_prob * 100)
+        if np.isfinite(self.att_sd):
+            extra.append(
+                f'  ATT: {self.att:.4f} (sd {self.att_sd:.4f}, '
+                f'{pct}% HDI [{self.att_hdi_lower:.4f}, '
+                f'{self.att_hdi_upper:.4f}])'
+            )
+        if np.isfinite(self.atu_sd):
+            extra.append(
+                f'  ATU: {self.atu:.4f} (sd {self.atu_sd:.4f}, '
+                f'{pct}% HDI [{self.atu_hdi_lower:.4f}, '
+                f'{self.atu_hdi_upper:.4f}])'
+            )
+        if not extra:
+            return base
+        block = '\n'.join(['-' * 70, 'Population-integrated effects', *extra])
+        closing = '=' * 70
+        if base.endswith(closing):
+            return base[: -len(closing)] + block + '\n' + closing
+        return base + '\n' + block
+
     def policy_effect(
         self,
         weight_fn,

tests/test_bayes_mte_uncertainty.py

@@ -83,3 +83,42 @@ def test_bayes_mte_att_atu_both_finite_on_realistic_dgp():
     assert np.isfinite(r.atu)
     assert np.isfinite(r.att_sd)
     assert np.isfinite(r.atu_sd)
+
+
+def test_summary_shows_att_atu_uncertainty():
+    """summary() must print an ATT / ATU block with sd + HDI once
+    the SD fields are finite (spec §3.3)."""
+    df = _hv_dgp(300, slope=1.0, seed=23)
+    r = bayes_mte(df, y='y', treat='d', instrument='z', poly_u=1,
+                  draws=200, tune=200, chains=2, progressbar=False)
+    s = r.summary()
+    assert 'ATT:' in s
+    assert 'ATU:' in s
+    # sd is lower-case in the spec-mandated format
+    assert 'sd ' in s
+    assert 'HDI [' in s
+
+
+def test_summary_skips_att_atu_when_nan():
+    """If ATT/ATU SDs are NaN (e.g. deserialised pre-v0.9.13 result),
+    summary() must silently omit the block rather than printing
+    ``sd nan`` garbage."""
+    from statspai.bayes._base import BayesianMTEResult
+    stub = BayesianMTEResult(
+        method='bayes_mte',
+        estimand='ATE',
+        posterior_mean=0.5,
+        posterior_median=0.5,
+        posterior_sd=0.1,
+        hdi_lower=0.3,
+        hdi_upper=0.7,
+        prob_positive=0.99,
+        rhat=1.0,
+        ess=400.0,
+        n_obs=100,
+        hdi_prob=0.95,
+        model_info={'inference': 'nuts', 'chains': 2, 'draws': 100},
+    )
+    out = stub.summary()
+    assert 'ATT:' not in out
+    assert 'ATU:' not in out