crypto-mcp/services/__init__.py

"""Service layer."""

import time
from config import DEFAULT_OHLCV_LIMIT, MAX_OHLCV_LIMIT, TIMEFRAME_TO_BINANCE
from cache import get_cached_price, set_cached_price, get_cached_ohlcv, set_cached_ohlcv
import providers
import indicators as ind_module


async def get_price(symbol: str) -> dict:
    symbol = symbol.upper()
    cached = get_cached_price(symbol)
    if cached:
        return cached
    data = await providers.fetch_price(symbol)
    set_cached_price(symbol, data)
    return data


async def get_ohlcv(symbol: str, timeframe: str, limit: int = DEFAULT_OHLCV_LIMIT) -> dict:
    symbol = symbol.upper()
    limit = min(max(limit, 1), MAX_OHLCV_LIMIT)
    cached = get_cached_ohlcv(symbol, timeframe)
    if cached:
        result = dict(cached)
        result["candles"] = cached["candles"][-limit:]
        return result
    data = await providers.fetch_ohlcv(symbol, timeframe, limit=MAX_OHLCV_LIMIT)
    set_cached_ohlcv(symbol, timeframe, data)
    result = dict(data)
    result["candles"] = data["candles"][-limit:]
    return result


async def get_indicator(symbol: str, indicator: str, timeframe: str = "1h", params: dict = None, limit: int = 200) -> dict:
    params = params or {}
    symbol = symbol.upper()
    ohlcv_data = await get_ohlcv(symbol, timeframe, limit=limit)
    result = ind_module.compute_indicator(ohlcv_data["candles"], indicator, params)
    return {"symbol": symbol, "indicator": result["indicator"], "timeframe": timeframe, "value": result["value"], "timestamp": int(time.time())}


async def get_market_snapshot(symbol: str) -> dict:
    symbol = symbol.upper()
    price_data = await get_price(symbol)
    ohlcv_data = await get_ohlcv(symbol, "1h", limit=200)
    candles = ohlcv_data["candles"]
    price = price_data["price"]
    snapshot = {
        "symbol": symbol,
        "price": price,
        "rsi_1h": None,
        "ema_20_1h": None,
        "ema_50_1h": None,
        "ema_200_1h": None,
        "macd_histogram_1h": None,
        "atr_1h": None,
        "atr_percent_1h": None,
        "bollinger_1h": None,
        "vwap_1h": None,
        "trend_bias": None,
        "timestamp": price_data["timestamp"],
    }

    def _compute(name: str, params: dict):
        return ind_module.compute_indicator(candles, name, params)["value"]

    for key, ind, params in [
        ("rsi_1h", "rsi", {"period": 14}),
        ("ema_20_1h", "ema", {"period": 20}),
        ("ema_50_1h", "ema", {"period": 50}),
        ("ema_200_1h", "sma", {"period": 200}),
    ]:
        try:
            snapshot[key] = _compute(ind, params)
        except Exception:
            continue

    try:
        macd_vals = _compute("macd", {"fast_period": 12, "slow_period": 26, "signal_period": 9})
        if isinstance(macd_vals, dict):
            snapshot["macd_histogram_1h"] = macd_vals.get("histogram")
    except Exception:
        pass

    try:
        atr_val = _compute("atr", {"period": 14})
        snapshot["atr_1h"] = atr_val
        if atr_val is not None and price:
            snapshot["atr_percent_1h"] = round((atr_val / price) * 100, 4)
    except Exception:
        pass

    try:
        snapshot["bollinger_1h"] = _compute("bollinger", {"period": 20, "multiplier": 2.0})
    except Exception:
        pass

    try:
        snapshot["vwap_1h"] = _compute("vwap", {"period": 48})
    except Exception:
        pass

    ema_fast = snapshot.get("ema_20_1h")
    ema_slow = snapshot.get("ema_50_1h")
    if ema_fast is not None and ema_slow not in (None, 0):
        delta = (ema_fast - ema_slow) / ema_slow
        if delta > 0.002:
            snapshot["trend_bias"] = "bull"
        elif delta < -0.002:
            snapshot["trend_bias"] = "bear"
        else:
            snapshot["trend_bias"] = "range"
    return snapshot


async def get_top_movers(limit: int = 10) -> dict:
    return await providers.fetch_top_movers(min(max(limit, 1), 50))


async def get_capabilities() -> dict:
    timeframe_descriptions = {
        "1m": "1-minute candles — ultra-fast scalping / heartbeat data (highest volatility, shortest TTL)",
        "5m": "5-minute candles — intraday momentum and micro-structure",
        "15m": "15-minute candles — short-term swings, aligns with many bot cycles",
        "1h": "1-hour candles — default for balanced trend/momentum reads",
        "4h": "4-hour candles — swing/position traders",
        "1d": "1-day candles — macro trend / higher timeframe context",
    }
    timeframes = []
    for tf, interval in TIMEFRAME_TO_BINANCE.items():
        timeframes.append(
            {
                "timeframe": tf,
                "provider_interval": interval,
                "description": timeframe_descriptions.get(tf, ""),
            }
        )
    return {
        "description": "Supported technical indicators (with params/defaults) and the timeframes you can request via get_ohlcv / get_indicator / get_regime.",
        "indicators": ind_module.get_supported_indicators(),
        "timeframes": timeframes,
    }


async def get_regime(symbol: str, timeframe: str = "1h", limit: int = 200) -> dict:
    symbol = symbol.upper()
    ohlcv_data = await get_ohlcv(symbol, timeframe, limit=limit)
    candles = ohlcv_data["candles"]
    close_price = float(candles[-1][4]) if candles else None
    timestamp = int(time.time())

    def _compute(name: str, params: dict | None = None):
        try:
            return ind_module.compute_indicator(candles, name, params or {})["value"]
        except Exception:
            return None

    ema_fast = _compute("ema", {"period": 20})
    ema_slow = _compute("ema", {"period": 50})
    sma_long = _compute("sma", {"period": 200})
    atr_val = _compute("atr", {"period": 14})
    boll = _compute("bollinger", {"period": 20, "multiplier": 2.0})
    vwap_val = _compute("vwap", {"period": 48})
    rsi_val = _compute("rsi", {"period": 14})
    macd_vals = _compute("macd", {"fast_period": 12, "slow_period": 26, "signal_period": 9})

    regime_delta = None
    if ema_fast is not None and ema_slow is not None and ema_slow != 0:
        regime_delta = (ema_fast - ema_slow) / ema_slow
    if regime_delta is None:
        trend_state = "unknown"
    elif regime_delta > 0.002:
        trend_state = "bull"
    elif regime_delta < -0.002:
        trend_state = "bear"
    else:
        trend_state = "range"

    if rsi_val is None:
        momentum_state = "unknown"
    elif rsi_val >= 60:
        momentum_state = "bull"
    elif rsi_val <= 40:
        momentum_state = "bear"
    else:
        momentum_state = "neutral"

    atr_percent = None
    if atr_val is not None and close_price:
        atr_percent = round((atr_val / close_price) * 100, 4)

    macd_hist = macd_vals.get("histogram") if isinstance(macd_vals, dict) else None

    # --- Early reversal score (heuristic) ---
    # Goal: flag plausible “pivot attempts” before full trend confirmation.
    # Note: without previous-bar indicator deltas, this is a conservative, snapshot-based score.
    reversal = {"direction": "none", "score": 0, "triggers": []}
    if close_price is not None and vwap_val is not None and isinstance(boll, dict):
        upper = boll.get("upper")
        lower = boll.get("lower")
        if upper and lower and (upper > lower):
            dist_to_upper = (upper - close_price) / close_price if close_price else None
            dist_to_lower = (close_price - lower) / close_price if close_price else None
        else:
            dist_to_upper = None
            dist_to_lower = None

        # Bullish reversal attempt while trend is bearish
        if trend_state == "bear" and macd_hist is not None and rsi_val is not None and macd_hist > 0:
            score = 0
            if rsi_val >= 45:
                score += 25
                reversal["triggers"].append("RSI rising/near-neutral (>=45)")
            if vwap_val is not None and close_price > vwap_val:
                score += 25
                reversal["triggers"].append("Price reclaimed above VWAP")
            if dist_to_lower is not None and dist_to_lower <= 0.01:
                score += 20
                reversal["triggers"].append("Near lower Bollinger band")
            if atr_percent is not None and atr_percent >= 0.8:
                score += 10
                reversal["triggers"].append("Volatility elevated (ATR% >= 0.8)")

            if score >= 40:
                reversal.update({"direction": "bullish", "score": min(score, 95)})

        # Bearish reversal attempt while trend is bullish
        if trend_state == "bull" and macd_hist is not None and rsi_val is not None and macd_hist < 0:
            score = 0
            if rsi_val <= 55:
                score += 25
                reversal["triggers"].append("RSI falling/near-neutral (<=55)")
            if vwap_val is not None and close_price < vwap_val:
                score += 25
                reversal["triggers"].append("Price rejected below VWAP")
            if dist_to_upper is not None and dist_to_upper <= 0.01:
                score += 20
                reversal["triggers"].append("Near upper Bollinger band")
            if atr_percent is not None and atr_percent >= 0.8:
                score += 10
                reversal["triggers"].append("Volatility elevated (ATR% >= 0.8)")

            if score >= 40:
                reversal.update({"direction": "bearish", "score": min(score, 95)})

    return {
        "symbol": symbol,
        "timeframe": timeframe,
        "timestamp": timestamp,
        "price": close_price,
        "trend": {
            "ema_fast": ema_fast,
            "ema_slow": ema_slow,
            "sma_long": sma_long,
            "state": trend_state,
        },
        "momentum": {
            "rsi": rsi_val,
            "macd_histogram": macd_hist,
            "state": momentum_state,
        },
        "volatility": {
            "atr": atr_val,
            "atr_percent": atr_percent,
        },
        "bands": {
            "bollinger": boll,
        },
        "vwap": vwap_val,
        "reversal": reversal,
    }