"""
Market sentiment tracker — CNN Fear & Greed + NAAIM + AAII.
All free/public sources. No API key required.
"""
from __future__ import annotations

import logging
import re
from datetime import datetime
from typing import Any, Dict, List, Optional, Tuple

import requests

logger = logging.getLogger(__name__)

CNN_FG_URL = "https://production.dataviz.cnn.io/index/fearandgreed/graphdata/"

NAAIM_URLS = [
    "https://www.naaim.org/wp-content/uploads/NAAIM-Exposure-Index-Weekly-Data.csv",
    "https://www.naaim.org/programs/naaim-exposure-index/",
]

AAII_URL = "https://www.aaii.com/sentimentsurvey/sent_results.js"

_HEADERS = {"User-Agent": "Mozilla/5.0 (compatible; OpenFin/1.0)"}
_TIMEOUT = 15


def _fetch_cnn_fear_greed() -> Optional[Dict[str, Any]]:
    try:
        resp = requests.get(CNN_FG_URL, timeout=_TIMEOUT, headers=_HEADERS)
        resp.raise_for_status()
        data = resp.json()
    except Exception as e:
        logger.warning(f"[SENTIMENT] CNN F&G fetch failed: {e}")
        return None

    fg = data.get("fear_and_greed", {})
    try:
        score         = float(fg.get("score", 0))
        rating        = str(fg.get("rating", "unknown")).lower()
        prev_close    = float(fg.get("previous_close", score))
        prev_1_week   = float(fg.get("previous_1_week", score))
        prev_1_month  = float(fg.get("previous_1_month", score))
        prev_1_year   = float(fg.get("previous_1_year", score))
    except (TypeError, ValueError) as e:
        logger.warning(f"[SENTIMENT] CNN F&G parse error: {e}")
        return None

    logger.info(f"[SENTIMENT] CNN F&G: {score:.1f} / 100 ({rating})")
    return {
        "score":           score,
        "rating":          rating,
        "previous_close":  prev_close,
        "previous_1_week": prev_1_week,
        "previous_1_month": prev_1_month,
        "previous_1_year":  prev_1_year,
    }


_NAAIM_NUMBER_RE = re.compile(r"(\d+\.?\d*)\s*(?:Average|Mean|Exposure)", re.IGNORECASE)


def _parse_naaim_csv(text: str) -> Optional[float]:
    lines = [l.strip() for l in text.splitlines() if l.strip()]
    if len(lines) < 2:
        return None
    data_line = lines[1]
    parts = re.split(r"[,\t]", data_line)
    for part in reversed(parts):
        try:
            val = float(part.strip())
            if 0.0 <= val <= 200.0:
                return val
        except ValueError:
            continue
    return None


def _fetch_naaim() -> Optional[float]:
    try:
        resp = requests.get(NAAIM_URLS[0], timeout=_TIMEOUT, headers=_HEADERS)
        if resp.status_code == 200 and resp.text.strip():
            val = _parse_naaim_csv(resp.text)
            if val is not None:
                logger.info(f"[SENTIMENT] NAAIM exposure (CSV): {val:.1f}")
                return val
    except Exception as e:
        logger.debug(f"[SENTIMENT] NAAIM CSV failed: {e}")

    try:
        resp = requests.get(NAAIM_URLS[1], timeout=_TIMEOUT, headers=_HEADERS)
        if resp.status_code == 200:
            m = _NAAIM_NUMBER_RE.search(resp.text)
            if m:
                val = float(m.group(1))
                logger.info(f"[SENTIMENT] NAAIM exposure (HTML): {val:.1f}")
                return val
    except Exception as e:
        logger.debug(f"[SENTIMENT] NAAIM HTML failed: {e}")

    logger.info("[SENTIMENT] NAAIM not available — skipping")
    return None


def _fetch_aaii() -> Optional[Tuple[float, float, float]]:
    try:
        resp = requests.get(AAII_URL, timeout=_TIMEOUT, headers=_HEADERS)
        if resp.status_code != 200:
            return None
        text = resp.text
    except Exception as e:
        logger.debug(f"[SENTIMENT] AAII fetch failed: {e}")
        return None

    try:
        bull_m = re.search(r'"bullish"\s*:\s*([\d.]+)', text, re.IGNORECASE)
        bear_m = re.search(r'"bearish"\s*:\s*([\d.]+)', text, re.IGNORECASE)
        neut_m = re.search(r'"neutral"\s*:\s*([\d.]+)', text, re.IGNORECASE)

        if bull_m and bear_m:
            bull = float(bull_m.group(1))
            bear = float(bear_m.group(1))
            neut = float(neut_m.group(1)) if neut_m else max(0.0, 100.0 - bull - bear)

            if 0 <= bull <= 100 and 0 <= bear <= 100:
                logger.info(f"[SENTIMENT] AAII Bulls={bull:.1f}% Bears={bear:.1f}%")
                return bull, bear, neut
    except (TypeError, ValueError) as e:
        logger.debug(f"[SENTIMENT] AAII parse error: {e}")

    logger.info("[SENTIMENT] AAII not available — skipping")
    return None


def _signal_indices_from_fg(score: float) -> str:
    if score < 25:
        return "bullish"
    if score > 75:
        return "bearish"
    if score < 40:
        return "bullish"
    if score > 60:
        return "bearish"
    return "neutral"


def _build_key_points(
    fg: Dict[str, Any],
    naaim: Optional[float],
    aaii: Optional[Tuple[float, float, float]],
) -> List[str]:
    points: List[str] = []

    score      = fg["score"]
    rating     = fg["rating"]
    prev_week  = fg["previous_1_week"]
    wow_delta  = score - prev_week

    points.append(
        f"CNN Fear & Greed: {score:.0f}/100 ({rating.upper()}) — WoW {wow_delta:+.0f}pt"
    )

    prev_close = fg["previous_close"]
    dod_delta  = score - prev_close
    points.append(f"CNN F&G DoD change: {dod_delta:+.0f}pt (prev close {prev_close:.0f})")

    if naaim is not None:
        naaim_level = "HIGH" if naaim > 80 else ("LOW" if naaim < 30 else "NORMAL")
        points.append(f"NAAIM Exposure Index: {naaim:.0f}/200 ({naaim_level})")
        if naaim > 80:
            points.append("NAAIM: Active managers at high equity exposure — limited incremental buying power")
        elif naaim < 30:
            points.append("NAAIM: Active managers heavily derisked — potential fuel for snapback rally")

    if aaii is not None:
        bull, bear, neut = aaii
        points.append(f"AAII Bulls: {bull:.0f}% | Bears: {bear:.0f}% | Neutral: {neut:.0f}%")
        if bull > 40:
            points.append(f"AAII: Bullish sentiment elevated ({bull:.0f}%) — mild contrarian bearish signal")
        elif bull < 20:
            points.append(f"AAII: Extreme pessimism ({bull:.0f}% bulls) — contrarian bullish signal")
        if bear > 40:
            points.append(f"AAII: High bear reading ({bear:.0f}%) — contrarian bullish signal")

    if score < 20:
        points.append("EXTREME FEAR: historically a strong mean-reversion buy signal (>80% 1M forward returns positive)")
    elif score > 80:
        points.append("EXTREME GREED: market complacency elevated — tail-risk hedging warranted")

    return points


def _build_implications(
    fg: Dict[str, Any],
    naaim: Optional[float],
    aaii: Optional[Tuple[float, float, float]],
) -> List[str]:
    score  = fg["score"]
    rating = fg["rating"]
    implications: List[str] = []

    if rating == "extreme_fear" or score < 20:
        implications.append(
            "Extreme Fear reading — historically strong contrarian buy signal for long delta"
        )
    elif rating == "extreme_greed" or score > 80:
        implications.append(
            "Extreme Greed — market may be overbought, consider protective puts or vol selling"
        )
    elif score < 40:
        implications.append(
            "Fear sentiment — elevated risk premium may support long vol or mean-reversion longs"
        )
    elif score > 60:
        implications.append(
            "Greed sentiment — consider trimming high-beta risk, watch for vol compression unwind"
        )
    else:
        implications.append(
            "Neutral sentiment — no contrarian edge; follow underlying trend and macro catalysts"
        )

    if naaim is not None:
        if naaim > 80:
            implications.append(
                "Active managers fully exposed — limited buying power remaining, correction risk elevated"
            )
        elif naaim < 30:
            implications.append(
                "Active managers heavily derisked — potential fuel for snapback rally on positive catalyst"
            )

    if aaii is not None:
        bull, bear, _ = aaii
        if bull > 40 and score > 60:
            implications.append(
                "AAII + CNN both elevated — dual sentiment warning, consider vol hedges"
            )
        elif bear > 40 and score < 40:
            implications.append(
                "AAII bears elevated + CNN fear — strong contrarian setup, watch for reversal catalyst"
            )

    return implications


def fetch_sentiment_report() -> Optional[Dict[str, Any]]:
    """
    Fetch CNN Fear & Greed (primary) + NAAIM + AAII (optional).
    Returns a single structured report dict compatible with institutional_reports table.
    Returns None only if CNN F&G fetch fails entirely.
    """
    fg = _fetch_cnn_fear_greed()
    if fg is None:
        logger.error("[SENTIMENT] CNN Fear & Greed unavailable — aborting sentiment report")
        return None

    naaim = _fetch_naaim()
    aaii  = _fetch_aaii()

    score  = fg["score"]
    rating = fg["rating"]

    key_points   = _build_key_points(fg, naaim, aaii)
    implications = _build_implications(fg, naaim, aaii)
    signal_indices = _signal_indices_from_fg(score)

    importance = 3 if (score < 20 or score > 80) else 2

    report_date = datetime.utcnow().strftime("%Y-%m-%d")

    sources = ["CNN Fear & Greed"]
    if naaim is not None:
        sources.append("NAAIM")
    if aaii is not None:
        sources.append("AAII")
    source_label = " + ".join(sources)

    raw: Dict[str, Any] = {"cnn_fear_greed": fg}
    if naaim is not None:
        raw["naaim"] = {
            "exposure_index": naaim,
            "level": "HIGH" if naaim > 80 else ("LOW" if naaim < 30 else "NORMAL"),
        }
    if aaii is not None:
        bull, bear, neut = aaii
        raw["aaii"] = {"bullish_pct": bull, "bearish_pct": bear, "neutral_pct": neut}

    ai_summary = (
        f"Market Sentiment ({report_date}). "
        f"CNN F&G: {score:.0f}/100 ({rating.upper()}), "
        f"WoW {score - fg['previous_1_week']:+.0f}pt. "
        + (f"NAAIM: {naaim:.0f}/200. " if naaim is not None else "")
        + (f"AAII Bulls: {aaii[0]:.0f}%. " if aaii is not None else "")
        + f"Signal: {signal_indices.upper()}. "
        + implications[0]
    )

    return {
        "report_type": "sentiment",
        "report_date": report_date,
        "title": f"Market Sentiment — CNN F&G {score:.0f}/100 ({rating.upper()}) — {report_date}",
        "source": source_label,
        "importance": importance,
        "category": "sentiment",
        "raw_data": raw,
        "key_points": key_points,
        "trading_implications": " | ".join(implications),
        "signal_energy": "neutral",
        "signal_metals": "neutral",
        "signal_indices": signal_indices,
        "signal_forex": "neutral",
        "ai_summary": ai_summary,
    }