Add Meta-Learners, Neural Causal, Causal Discovery, TMLE, Policy Learning, Robustness modules; update README

New modules:
- robustness: spec_curve (Simonsohn 2020), robustness_report, subgroup_analysis
- metalearners: S/T/X/R/DR-Learner + CATE diagnostics
- neural_causal: TARNet, CFRNet, DragonNet
- causal_discovery: NOTEARS, PC algorithm
- tmle: TMLE + Super Learner
- policy_learning: Policy Tree + policy value
- conformal_causal, bcf, bunching, matrix_completion (scaffolds)
- plots: interactive editor

README rewritten with full 120-API feature list, honest Stata/R comparison,
and robustness automation examples.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: brycewang-stanford

README.md

@@ -52,12 +52,23 @@
 from .postestimation import margins, marginsplot, test, lincom
 from .diagnostics import oster_bounds, mccrary_test, diagnose, het_test, reset_test, vif, sensemakr, rddensity, hausman_test, anderson_rubin_test
 from .inference import wild_cluster_bootstrap, aipw, ri_test
-from .plots import binscatter, set_theme
+from .plots import binscatter, set_theme, interactive, get_code
 from .utils import label_var, label_vars, get_label, get_labels, describe, pwcorr, winsor, read_data
 from .gmm import xtabond
+from .metalearners import metalearner, SLearner, TLearner, XLearner, RLearner, DRLearner
+from .metalearners import cate_summary, cate_by_group, cate_plot, cate_group_plot, predict_cate, compare_metalearners, gate_test, blp_test
 from .regression.heckman import heckman
 from .regression.quantile import qreg, sqreg
 from .regression.tobit import tobit
+from .neural_causal import tarnet, cfrnet, dragonnet, TARNet, CFRNet, DragonNet
+from .causal_discovery import notears, NOTEARS, pc_algorithm, PCAlgorithm
+from .tmle import tmle, TMLE, super_learner, SuperLearner
+from .policy_learning import policy_tree, PolicyTree, policy_value
+from .conformal_causal import conformal_cate, ConformalCATE
+from .bcf import bcf, BayesianCausalForest
+from .bunching import bunching, BunchingEstimator
+from .matrix_completion import mc_panel, MCPanel
+from .robustness import spec_curve, SpecCurveResult, robustness_report, RobustnessResult, subgroup_analysis, SubgroupResult
 
 __all__ = [
     # Core
@@ -120,6 +131,8 @@
     # Plots
     "binscatter",
     "set_theme",
+    "interactive",
+    "get_code",
     # Utils
     "label_var",
     "label_vars",
@@ -161,4 +174,48 @@
     "wild_cluster_bootstrap",
     "aipw",
     "ri_test",
+    # Meta-Learners (HTE)
+    "metalearner",
+    "SLearner",
+    "TLearner",
+    "XLearner",
+    "RLearner",
+    "DRLearner",
+    # CATE Diagnostics
+    "cate_summary",
+    "cate_by_group",
+    "cate_plot",
+    "cate_group_plot",
+    "predict_cate",
+    "compare_metalearners",
+    "gate_test",
+    "blp_test",
+    # Neural Causal Models
+    "tarnet",
+    "cfrnet",
+    "dragonnet",
+    "TARNet",
+    "CFRNet",
+    "DragonNet",
+    # Causal Discovery
+    "notears",
+    "NOTEARS",
+    "pc_algorithm",
+    "PCAlgorithm",
+    # TMLE
+    "tmle",
+    "TMLE",
+    "super_learner",
+    "SuperLearner",
+    # Policy Learning
+    "policy_tree",
+    "PolicyTree",
+    "policy_value",
+    # Robustness
+    "spec_curve",
+    "SpecCurveResult",
+    "robustness_report",
+    "RobustnessResult",
+    "subgroup_analysis",
+    "SubgroupResult",
 ]

src/statspai/__init__.py

@@ -0,0 +1,22 @@
+"""
+Bayesian Causal Forest (BCF) for heterogeneous treatment effects.
+
+Decomposes the outcome into a prognostic function mu(X) and a
+treatment effect function tau(X), each modeled by BART:
+
+    Y_i = mu(X_i) + tau(X_i) * D_i + epsilon_i
+
+This separation allows regularization-induced confounding (RIC) to be
+mitigated, producing better CATE estimates than standard BART.
+
+References
+----------
+Hahn, P. R., Murray, J. S., & Carvalho, C. M. (2020).
+Bayesian Regression Tree Models for Causal Inference: Regularization,
+Confounding, and Heterogeneous Effects.
+Bayesian Analysis, 15(3), 965-1056.
+"""
+
+from .bcf import bcf, BayesianCausalForest
+
+__all__ = ['bcf', 'BayesianCausalForest']