#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""Régénère paper/tables/regression.tex à partir de revision_extras.json
   avec SE clusterisées par municipalité (et stars associées)."""
import json

R = json.load(open("results/revision_extras.json"))

def star(p):
    return "^{***}" if p < .01 else "^{**}" if p < .05 else "^{*}" if p < .1 else ""

def cell(model, k):
    c = R[model]["coef"].get(k)
    if not c:
        return "", ""
    b, se, p = c["b"], c["se_cl"], c["p_cl"]
    s = "$-$" if b < 0 else ""
    st = star(p)
    coef = f"{s}{abs(b):.3f}" + (f"${st}$" if st else "")
    return coef, f"({se:.3f})"

disp = [("ln_sqft", r"$\ln(\text{Living area})$"), ("ln_lot", r"$\ln(\text{Lot area})$"),
        ("beds", "Bedrooms"), ("baths", "Bathrooms"), ("parking", "Parking"), ("age", "Age"),
        ("__A__", r"\textit{Amenities \& text (LLM)}"),
        ("central_air", "Central A/C"), ("transit", "Public transit"), ("waterfront", "Waterfront"),
        ("g_cond", "Condition (LLM, ord.)"), ("g_renovated_kitchen", r"Reno.\ kitchen (LLM)"),
        ("g_hardwood", "Hardwood floor (LLM)"), ("g_new_construction", "New construction (LLM)"),
        ("g_income_property", "Income property (LLM)"),
        ("__V__", r"\textit{Vision (façade)}"),
        ("v_curb", "Curb appeal (1--10)"), ("v_reno", "Renovation level (ord.)"),
        ("v_luxury", "Luxury (visual)"), ("v_garage_vis", "Garage visible"),
        ("v_mat_brique", "Brick façade"), ("v_mat_pierre", "Stone façade"),
        ("v_mat_vinyle", "Vinyl façade"), ("v_set_rural", "Rural setting")]

L = []
for k, lab in disp:
    if k in ("__A__", "__V__"):
        L.append(f"      \\multicolumn{{3}}{{@{{}}l}}{{{lab}}} \\\\[2pt]")
        continue
    a1, a2 = cell("modelA", k)
    b1, b2 = cell("modelB", k)
    L.append(f"      {lab} & {a1} & {b1} \\\\")
    L.append(f"       & {a2} & {b2} \\\\[3pt]")

nA, nB = R["modelA"]["n"], R["modelB"]["n"]
tex = r"""% Auto-generated (revision: clustered SEs)
\begin{table}[htbp]
  \centering
  \begin{threeparttable}
    \caption{Hedonic Regression Results (Text + Vision Features)}
    \label{tab:regression}
    \begin{tabular}{@{}lcc@{}}
      \toprule
      & (1) Model~A & (2) Model~B \\
      & \textit{rich} & \textit{broad} \\
      \midrule
""" + "\n".join(L) + f"""
      \\midrule
      City FE / Type FE  & Yes & Yes \\\\
      Living area        & Yes & No  \\\\
      Observations       & {nA:,} & {nB:,} \\\\
      $R^2$              & {R['modelA']['r2']:.3f} & {R['modelB']['r2']:.3f} \\\\
      Adjusted $R^2$     & {R['modelA']['adj']:.3f} & {R['modelB']['adj']:.3f} \\\\
      \\bottomrule
    \\end{{tabular}}
""".replace(f"{nA:,}", f"{nA:,}".replace(",", "{,}")).replace(f"{nB:,}", f"{nB:,}".replace(",", "{,}")) + r"""    \begin{tablenotes}[flushleft]
      \footnotesize
      \item \textit{Notes:} OLS estimates of Equation~\eqref{eq:hedonic} with the full
      feature set. Model~A includes $\ln(\text{Living area})$ (measured-area subsample);
      Model~B omits it (full sample). ``LLM'' variables are extracted from the
      description text; ``Vision'' variables from the façade photograph. Curb appeal and
      renovation level enter as scores; luxury, garage, and material/setting are
      indicators. All specifications include city and property-type fixed effects and the
      full amenity, room-detail, and listing-signal controls (only selected coefficients
      shown). Standard errors clustered by municipality (195 clusters) in parentheses.
      Heteroskedasticity-robust (HC3) standard errors are uniformly smaller and yield
      the same significance pattern for the text- and vision-derived attributes, except
      that curb appeal in Model~B is significant under HC3 but not under clustering.
      $^{***}$, $^{**}$, $^{*}$: significance at 1\%, 5\%, 10\%.
    \end{tablenotes}
  \end{threeparttable}
\end{table}
"""
open("paper/tables/regression.tex", "w").write(tex)
print("-> paper/tables/regression.tex (clustered SEs)")