from utils.db import serenitas_pool, dbconn
import datetime
import pandas as pd
from pandas.tseries.offsets import BDay
from psycopg2.extras import execute_values
from pyisda.date import default_accrual, previous_twentieth
from analytics.index import CreditIndex
from copy import copy


def get_outstanding_positions(conn, trade_date, fcm, fund="SERCGMAST"):
    with conn.cursor() as c:
        c.execute(
            "SELECT security_id, notional, folder, nextredindexcode, currency, "
            "maturity, indexfactor "
            "FROM list_cds_positions_by_strat_fcm(%s, %s, %s) a "
            "JOIN index_version_markit "
            "ON a.security_id=index_version_markit.redindexcode "
            "WHERE nextredindexcode IS NOT NULL AND nextredindexcode !='2I65BRUV0'",
            (trade_date, fcm, fund),
        )
        yield from c


def default_adjustment(conn, company_id, seniority, end_date):
    with conn.cursor() as c:
        c.execute(
            "SELECT recovery, event_date, auction_date FROM defaulted WHERE id=%s "
            "AND seniority=%s",
            (company_id, seniority),
        )
        recovery, event_date, auction_date = next(c)
    fee = 1 - recovery
    start_date = previous_twentieth(event_date)
    accrual_days, _ = default_accrual(
        auction_date, event_date, start_date, end_date, 1.0, 1.0
    )
    return accrual_days, fee


PORTFOLIO = {
    "HYOPTDEL": "OPTIONS",
    "HEDGE_MBS": "MORTGAGES",
    "HYINX": "TRANCHE",
    "SER_IGCURVE": "CURVE",
    "HEDGE_CLO": "CLO",
}


def rebook(conn, trade_date, company_id, seniority, fcm, fund="SERCGMAST"):
    dawndb = dbconn("dawndb")
    upfront_settle_date = trade_date + 3 * BDay()
    effective_date = trade_date + datetime.timedelta(days=1)
    for r in get_outstanding_positions(dawndb, trade_date, fcm, fund):
        accrual_days, fee = default_adjustment(conn, company_id, seniority, r.maturity)
        index_new = CreditIndex(
            redcode=r.nextredindexcode,
            maturity=r.maturity,
            value_date=trade_date,
            notional=r.notional,
        )

        adj = (
            (fee - accrual_days * index_new.fixed_rate * 1e-4 / 360)
            * r.notional
            * (r.indexfactor - index_new.factor)
        )
        index_new.mark()
        trade_new = {
            "fund": fund,
            "action": "NEW",
            "portfolio": PORTFOLIO[r.folder],
            "folder": r.folder,
            "cp_code": "CONTRA",
            "custodian": "NONE",
            "trade_date": trade_date,
            "effective_date": effective_date,
            "maturity": r.maturity,
            "currency": r.currency,
            "payment_rolldate": "Following",
            "notional": abs(r.notional),
            "fixed_rate": index_new.fixed_rate / 100,
            "day_count": "ACT/360",
            "frequency": 4,
            "protection": index_new.direction,
            "security_id": r.nextredindexcode,
            "security_desc": f"CDX {index_new.index_type} CDSI S{index_new.series} 5Y",
            "upfront": index_new.pv,
            "upfront_settle_date": upfront_settle_date,
            "swap_type": "CD_INDEX",
            "account_code": fcm,
        }
        trade_prev = copy(trade_new)
        trade_prev["protection"] = (
            "Seller" if trade_new["protection"] == "Buyer" else "Buyer"
        )
        trade_prev["upfront"] = adj - index_new.pv
        trade_prev["security_id"] = r.security_id
        sql_str = f"INSERT INTO cds({','.join(trade_new.keys())}) VALUES %s"
        with dawndb.cursor() as c:
            execute_values(
                c, sql_str, [tuple(trade_prev.values()), tuple(trade_new.values())]
            )
    dawndb.commit()
    dawndb.close()


def insert_newids(dawndb, d: datetime.date, df: pd.DataFrame):
    with dawndb.cursor() as c:
        c.execute(
            "INSERT INTO id_mapping ( "
            "   SELECT %s, 'CDS', serenitas_id, new_id FROM ( "
            "       SELECT * FROM unnest(%s, %s) AS t(globeop_id, new_id) "
            "       LEFT JOIN id_mapping USING (globeop_id)) a"
            ")",
            (d, df.old_ids.tolist(), df.new_ids.to_list()),
        )
    dawndb.commit()


if __name__ == "__main__":
    conn = serenitas_pool.getconn()
    # PKD
    # rebook(datetime.date(2019, 1, 24), 101148)
    # WINDSSE
    # rebook(datetime.date(2019, 4, 8), 36806879)
    # WFT
    # rebook(datetime.date(2019, 7, 26), 103633, "WF")
    # rebook(datetime.date(2019, 7, 26), 103633, "BAML")
    # DF
    # rebook(datetime.date(2019, 12, 11), 154954, "Senior", "BAML")
    # MNI
    # rebook(datetime.date(2020, 3, 13), 100957, "Senior", "BAML")
    # rebook(datetime.date(2020, 3, 13), 100957, "Senior", "WF")
    # WLL
    # rebook(datetime.date(2020, 5, 7), 8240322, "Senior", "BAML")
    # rebook(datetime.date(2020, 5, 7), 8240322, "Senior", "WF")
    # rebook(conn, datetime.date(2020, 5, 7), 8240322, "Senior", "GS", "BOWDST")
    # FCA
    rebook(conn, datetime.date(2020, 5, 14), 100337, "Senior", "BAML")
    rebook(conn, datetime.date(2020, 5, 14), 100337, "Senior", "WF")
    rebook(conn, datetime.date(2020, 5, 14), 100337, "Senior", "GS", "BOWDST")
    serenitas_pool.putconn(conn)