"""LaTeX table generators for the Polymarket Arrow-Debreu paper."""

from __future__ import annotations

from pathlib import Path

import numpy as np
import pandas as pd

from src.config import TABLES


def _save_table(content: str, name: str) -> str:
    TABLES.mkdir(parents=True, exist_ok=True)
    path = TABLES / f"{name}.tex"
    path.write_text(content)
    return content


def _esc(s: str) -> str:
    """Escape LaTeX special characters in text."""
    return str(s).replace("$", r"\$").replace("_", r"\_").replace("&", r"\&").replace("%", r"\%").replace("#", r"\#")


def _fmt(x: float, decimals: int = 3) -> str:
    if pd.isna(x):
        return ""
    return f"{x:.{decimals}f}"


def _stars(p: float) -> str:
    if pd.isna(p):
        return ""
    if p < 0.01:
        return "^{***}"
    if p < 0.05:
        return "^{**}"
    if p < 0.10:
        return "^{*}"
    return ""


def table_descriptive_stats(metadata: pd.DataFrame) -> str:
    """Table 1: Descriptive statistics of Polymarket contracts."""
    rows = []
    for _, row in metadata.iterrows():
        name = _esc(str(row.get("name", row.get("question", "")))[:40])
        vol = row.get("volume", 0)
        liq = row.get("liquidity", 0)
        n_obs = int(row.get("n_obs", 0))
        mean_p = row.get("mean_price", np.nan)
        std_p = row.get("std_price", np.nan)
        rows.append(
            f"    {name} & {vol:,.0f} & {liq:,.0f} & {n_obs} & {_fmt(mean_p)} & {_fmt(std_p)} \\\\"
        )

    body = "\n".join(rows)
    tex = rf"""\begin{{table}}[htbp]
\centering
\begin{{threeparttable}}
\caption{{Descriptive Statistics of Polymarket Contracts}}
\label{{tab:descriptive_stats}}
\begin{{tabular}}{{l S S c S S}}
\toprule
Market & {{Volume (\$)}} & {{Liquidity (\$)}} & Obs & {{Mean Price}} & {{Std Dev}} \\
\midrule
{body}
\bottomrule
\end{{tabular}}
\begin{{tablenotes}}
\small
\item \textit{{Notes:}} This table reports summary statistics for the Polymarket contracts analyzed.
Volume and liquidity are reported in US dollars. Mean price represents the average implied probability.
\end{{tablenotes}}
\end{{threeparttable}}
\end{{table}}"""
    return _save_table(tex, "table1_descriptive_stats")


def table_summary_stats(prices: pd.DataFrame) -> str:
    """Table 2: Summary statistics of price series."""
    stats = []
    for col in prices.columns:
        s = prices[col].dropna()
        from scipy.stats import jarque_bera
        try:
            jb_stat, jb_p = jarque_bera(s)
        except Exception:
            jb_stat, jb_p = np.nan, np.nan
        stats.append({
            "Market": _esc(col[:30]),
            "Mean": s.mean(),
            "Std": s.std(),
            "Min": s.min(),
            "Max": s.max(),
            "Skew": s.skew(),
            "Kurt": s.kurtosis(),
            "JB": jb_stat,
        })

    rows = []
    for st in stats:
        rows.append(
            f"    {st['Market']} & {_fmt(st['Mean'])} & {_fmt(st['Std'])} & "
            f"{_fmt(st['Min'])} & {_fmt(st['Max'])} & {_fmt(st['Skew'],2)} & "
            f"{_fmt(st['Kurt'],2)} & {_fmt(st['JB'],1)} \\\\"
        )

    body = "\n".join(rows)
    tex = rf"""\begin{{table}}[htbp]
\centering
\begin{{threeparttable}}
\caption{{Summary Statistics of Prediction Market Prices}}
\label{{tab:summary_stats}}
\begin{{tabular}}{{l S S S S S S S}}
\toprule
Market & {{Mean}} & {{Std}} & {{Min}} & {{Max}} & {{Skew}} & {{Kurt}} & {{JB}} \\
\midrule
{body}
\bottomrule
\end{{tabular}}
\begin{{tablenotes}}
\small
\item \textit{{Notes:}} JB is the Jarque--Bera test statistic for normality.
$^{{***}}$, $^{{**}}$, and $^{{*}}$ denote significance at the 1\%, 5\%, and 10\% levels.
\end{{tablenotes}}
\end{{threeparttable}}
\end{{table}}"""
    return _save_table(tex, "table2_summary_stats")


def table_completeness_test(test_result) -> str:
    """Table 3: Arrow-Debreu completeness test results."""
    tex = rf"""\begin{{table}}[htbp]
\centering
\begin{{threeparttable}}
\caption{{Arrow--Debreu Market Completeness Tests}}
\label{{tab:completeness}}
\begin{{tabular}}{{l S}}
\toprule
Statistic & {{Value}} \\
\midrule
    Mean state price sum & {_fmt(test_result.mean_sum, 4)} \\
    Std of state price sum & {_fmt(test_result.std_sum, 4)} \\
    Mean deviation from unity & {_fmt(test_result.mean_deviation, 4)} \\
    Max absolute deviation & {_fmt(test_result.max_deviation, 4)} \\
    All state prices positive & {{{'Yes' if test_result.all_positive else 'No'}}} \\
    Implied discount factor & {_fmt(test_result.discount_factor, 4)} \\
    Number of observations & {len(test_result.daily_sums)} \\
\bottomrule
\end{{tabular}}
\begin{{tablenotes}}
\small
\item \textit{{Notes:}} This table reports results from testing Arrow--Debreu market completeness
conditions using the Fed rate decision markets on Polymarket. Under complete markets,
the sum of state prices equals $1/(1+r_f)$. Mean deviation measures the average
departure from this theoretical prediction.
\end{{tablenotes}}
\end{{threeparttable}}
\end{{table}}"""
    return _save_table(tex, "table3_completeness")


def table_event_study(results: dict[str, object]) -> str:
    """Table 4: Event study cumulative abnormal returns."""
    rows = []
    for event_type, result in results.items():
        if hasattr(result, "test_stats") and not result.test_stats.empty:
            for _, row in result.test_stats.iterrows():
                window = row.get("window", "")
                market = _esc(str(row.get("market", "")))
                car = row.get("mean_CAR", np.nan)
                t = row.get("t_stat", np.nan)
                p = row.get("p_value", np.nan)
                stars = _stars(p)
                rows.append(
                    f"    {_esc(event_type)}: {market} & {window} & ${_fmt(car)}{stars}$ & "
                    f"{_fmt(t, 2)} & {_fmt(p)} & {result.n_events} \\\\"
                )

    body = "\n".join(rows) if rows else "    \\multicolumn{6}{c}{No significant results} \\\\"
    tex = rf"""\begin{{table}}[htbp]
\centering
\begin{{threeparttable}}
\caption{{Event Study: Cumulative Abnormal Returns Around Macroeconomic Events}}
\label{{tab:event_study}}
\begin{{tabular}}{{l l c c c c}}
\toprule
Event Type & Window & {{CAR}} & {{$t$-stat}} & {{$p$-value}} & $N$ \\
\midrule
{body}
\bottomrule
\end{{tabular}}
\begin{{tablenotes}}
\small
\item \textit{{Notes:}} This table reports cumulative abnormal returns (CARs) for prediction market
prices around macroeconomic events. Abnormal returns are computed relative to a constant mean
model estimated over a 60-day estimation window. $t$-statistics are computed using the
cross-sectional standard deviation of CARs. $^{{***}}$, $^{{**}}$, $^{{*}}$ denote
significance at 1\%, 5\%, 10\%.
\end{{tablenotes}}
\end{{threeparttable}}
\end{{table}}"""
    return _save_table(tex, "table4_event_study")


def table_granger_causality(f_stats: pd.DataFrame, p_values: pd.DataFrame) -> str:
    """Table 5: Granger causality test results."""
    rows = []
    for caused in f_stats.index:
        for causing in f_stats.columns:
            if caused != causing:
                f_val = f_stats.loc[caused, causing]
                p_val = p_values.loc[caused, causing]
                stars = _stars(p_val)
                rows.append(
                    f"    {_esc(causing)} $\\to$ {_esc(caused)} & ${_fmt(f_val, 2)}{stars}$ & {_fmt(p_val)} \\\\"
                )

    body = "\n".join(rows)
    tex = rf"""\begin{{table}}[htbp]
\centering
\begin{{threeparttable}}
\caption{{Granger Causality Tests}}
\label{{tab:granger}}
\begin{{tabular}}{{l c c}}
\toprule
Direction & {{$F$-stat}} & {{$p$-value}} \\
\midrule
{body}
\bottomrule
\end{{tabular}}
\begin{{tablenotes}}
\small
\item \textit{{Notes:}} This table reports $F$-statistics and $p$-values for pairwise Granger
causality tests. The null hypothesis is that the causing variable does not Granger-cause
the caused variable. Lag length selected by AIC. $^{{***}}$, $^{{**}}$, $^{{*}}$ denote
significance at 1\%, 5\%, 10\%.
\end{{tablenotes}}
\end{{threeparttable}}
\end{{table}}"""
    return _save_table(tex, "table5_granger")


def table_var_results(var_results: dict) -> str:
    """Table 6: VAR estimation results."""
    tex = rf"""\begin{{table}}[htbp]
\centering
\begin{{threeparttable}}
\caption{{Vector Autoregression Estimates}}
\label{{tab:var}}
\begin{{tabular}}{{l {'S ' * min(len(var_results.get('coefficients', {}).keys()), 4)}}}
\toprule
"""
    coefs = var_results.get("coefficients", {})
    if coefs:
        cols = list(coefs.keys())[:4]
        tex += " & ".join(["Variable"] + [f"{{{_esc(c)}}}" for c in cols]) + " \\\\\n\\midrule\n"
        first_col = coefs[cols[0]]
        if isinstance(first_col, pd.Series):
            for var_name in first_col.index[:15]:
                row_vals = [f"{_fmt(coefs[c].get(var_name, np.nan), 3)}" for c in cols]
                tex += f"    {_esc(str(var_name))} & " + " & ".join(row_vals) + " \\\\\n"
    tex += rf"""\bottomrule
\end{{tabular}}
\begin{{tablenotes}}
\small
\item \textit{{Notes:}} VAR model estimates. Optimal lag length selected by AIC.
Standard errors in parentheses. $^{{***}}$, $^{{**}}$, $^{{*}}$ denote significance
at 1\%, 5\%, 10\%.
\end{{tablenotes}}
\end{{threeparttable}}
\end{{table}}"""
    return _save_table(tex, "table6_var")


def table_garch(garch_results: dict[str, object]) -> str:
    """Table 7: GARCH estimation results."""
    rows = []
    for name, result in garch_results.items():
        if hasattr(result, "params"):
            for param_name, value in result.params.items():
                se = result.std_errors.get(param_name, np.nan) if hasattr(result, "std_errors") else np.nan
                rows.append(f"    {_esc(name)}: {_esc(str(param_name))} & {_fmt(value, 4)} & ({_fmt(se, 4)}) \\\\")

    body = "\n".join(rows) if rows else "    \\multicolumn{3}{c}{No results} \\\\"
    tex = rf"""\begin{{table}}[htbp]
\centering
\begin{{threeparttable}}
\caption{{GARCH Model Estimates}}
\label{{tab:garch}}
\begin{{tabular}}{{l S l}}
\toprule
Parameter & {{Estimate}} & (Std Err) \\
\midrule
{body}
\bottomrule
\end{{tabular}}
\begin{{tablenotes}}
\small
\item \textit{{Notes:}} GARCH(1,1) model estimates for prediction market return volatility.
$^{{***}}$, $^{{**}}$, $^{{*}}$ denote significance at 1\%, 5\%, 10\%.
\end{{tablenotes}}
\end{{threeparttable}}
\end{{table}}"""
    return _save_table(tex, "table7_garch")


def table_har_rv(har_result) -> str:
    """Table 8: HAR-RV estimation results."""
    rows = []
    if hasattr(har_result, "coefficients"):
        for name, val in har_result.coefficients.items():
            se = har_result.std_errors.get(name, np.nan) if hasattr(har_result, "std_errors") else np.nan
            t = har_result.t_stats.get(name, np.nan) if hasattr(har_result, "t_stats") else np.nan
            rows.append(f"    {_esc(str(name))} & {_fmt(val, 4)} & ({_fmt(se, 4)}) & {_fmt(t, 2)} \\\\")

    body = "\n".join(rows) if rows else "    \\multicolumn{4}{c}{No results} \\\\"
    r2 = _fmt(har_result.r_squared, 3) if hasattr(har_result, "r_squared") else ""
    adj_r2 = _fmt(har_result.adj_r_squared, 3) if hasattr(har_result, "adj_r_squared") else ""

    tex = rf"""\begin{{table}}[htbp]
\centering
\begin{{threeparttable}}
\caption{{HAR-RV Model Estimates}}
\label{{tab:har_rv}}
\begin{{tabular}}{{l S l S}}
\toprule
Variable & {{Coefficient}} & (Std Err) & {{$t$-stat}} \\
\midrule
{body}
\midrule
    $R^2$ & \multicolumn{{3}}{{c}}{{{r2}}} \\
    Adj.~$R^2$ & \multicolumn{{3}}{{c}}{{{adj_r2}}} \\
\bottomrule
\end{{tabular}}
\begin{{tablenotes}}
\small
\item \textit{{Notes:}} HAR-RV model following \citet{{corsi2009simple}}.
RV$_{{t+1}}^{{(d)}} = c + \beta_d RV_t^{{(d)}} + \beta_w RV_t^{{(w)}} + \beta_m RV_t^{{(m)}} + \varepsilon_{{t+1}}$.
Newey--West standard errors with 5 lags.
\end{{tablenotes}}
\end{{threeparttable}}
\end{{table}}"""
    return _save_table(tex, "table8_har_rv")


def table_liquidity(liquidity_stats: pd.DataFrame) -> str:
    """Table 9: Liquidity regression results."""
    rows = []
    for _, row in liquidity_stats.iterrows():
        var_name = _esc(str(row.get("variable", "")))
        coef = row.get("coefficient", np.nan)
        se = row.get("std_error", np.nan)
        t = row.get("t_stat", np.nan)
        p = row.get("p_value", np.nan)
        stars = _stars(p)
        rows.append(
            f"    {var_name} & ${_fmt(coef, 4)}{stars}$ & ({_fmt(se, 4)}) & {_fmt(t, 2)} \\\\"
        )

    body = "\n".join(rows) if rows else "    \\multicolumn{4}{c}{No results} \\\\"
    tex = rf"""\begin{{table}}[htbp]
\centering
\begin{{threeparttable}}
\caption{{Liquidity Determinants}}
\label{{tab:liquidity}}
\begin{{tabular}}{{l S l S}}
\toprule
Variable & {{Coefficient}} & (Std Err) & {{$t$-stat}} \\
\midrule
{body}
\bottomrule
\end{{tabular}}
\begin{{tablenotes}}
\small
\item \textit{{Notes:}} OLS regression of bid--ask spread on volume, volatility, and event dummies.
Newey--West standard errors. $^{{***}}$, $^{{**}}$, $^{{*}}$ denote significance at 1\%, 5\%, 10\%.
\end{{tablenotes}}
\end{{threeparttable}}
\end{{table}}"""
    return _save_table(tex, "table9_liquidity")