paper-dynasty-database/app/routers_v2/battingcardratings.py

import os

from fastapi import APIRouter, Depends, HTTPException, Query, Response
from fastapi.responses import FileResponse
from scipy import stats
from typing import Literal, Optional, List
import logging
import pandas as pd
import pydantic
from pydantic import validator, root_validator

from ..db_engine import (
    db,
    BattingCardRatings,
    model_to_dict,
    chunked,
    BattingCard,
    Player,
    query_to_csv,
    Team,
    CardPosition,
)
from ..db_helpers import upsert_batting_card_ratings
from ..dependencies import oauth2_scheme, valid_token, LOG_DATA, PRIVATE_IN_SCHEMA

logging.basicConfig(
    filename=LOG_DATA["filename"],
    format=LOG_DATA["format"],
    level=LOG_DATA["log_level"],
)

router = APIRouter(prefix="/api/v2/battingcardratings", tags=["battingcardratings"])
RATINGS_FILE = "storage/batting-ratings.csv"
BASIC_FILE = "storage/batting-basic.csv"


class BattingCardRatingsModel(pydantic.BaseModel):
    battingcard_id: int
    vs_hand: Literal["R", "L", "vR", "vL"]
    homerun: float = 0.0
    bp_homerun: float = 0.0
    triple: float = 0.0
    double_three: float = 0.0
    double_two: float = 0.0
    double_pull: float = 0.0
    single_two: float = 0.0
    single_one: float = 0.0
    single_center: float = 0.0
    bp_single: float = 0.0
    hbp: float = 0.0
    walk: float = 0.0
    strikeout: float = 0.0
    lineout: float = 0.0
    popout: float = 0.0
    flyout_a: float = 0.0
    flyout_bq: float = 0.0
    flyout_lf_b: float = 0.0
    flyout_rf_b: float = 0.0
    groundout_a: float = 0.0
    groundout_b: float = 0.0
    groundout_c: float = 0.0
    avg: float = 0.0
    obp: float = 0.0
    slg: float = 0.0
    pull_rate: float = 0.0
    center_rate: float = 0.0
    slap_rate: float = 0.0

    @validator("avg", always=True)
    def avg_validator(cls, v, values, **kwargs):
        return (
            values["homerun"]
            + values["bp_homerun"] / 2
            + values["triple"]
            + values["double_three"]
            + values["double_two"]
            + values["double_pull"]
            + values["single_two"]
            + values["single_one"]
            + values["single_center"]
            + values["bp_single"] / 2
        ) / 108

    @validator("obp", always=True)
    def obp_validator(cls, v, values, **kwargs):
        return ((values["hbp"] + values["walk"]) / 108) + values["avg"]

    @validator("slg", always=True)
    def slg_validator(cls, v, values, **kwargs):
        return (
            values["homerun"] * 4
            + values["bp_homerun"] * 2
            + values["triple"] * 3
            + values["double_three"] * 2
            + values["double_two"] * 2
            + values["double_pull"] * 2
            + values["single_two"]
            + values["single_one"]
            + values["single_center"]
            + values["bp_single"] / 2
        ) / 108

    @root_validator(skip_on_failure=True)
    def validate_chance_total(cls, values):
        total_chances = (
            values["homerun"]
            + values["bp_homerun"]
            + values["triple"]
            + values["double_three"]
            + values["double_two"]
            + values["double_pull"]
            + values["single_two"]
            + values["single_one"]
            + values["single_center"]
            + values["bp_single"]
            + values["hbp"]
            + values["walk"]
            + values["strikeout"]
            + values["lineout"]
            + values["popout"]
            + values["flyout_a"]
            + values["flyout_bq"]
            + values["flyout_lf_b"]
            + values["flyout_rf_b"]
            + values["groundout_a"]
            + values["groundout_b"]
            + values["groundout_c"]
        )

        if round(total_chances) != 108:
            raise ValueError(
                f"BC {values['battingcard_id']} must have exactly 108 chances on the card "
                f"{values['vs_hand']}; {round(total_chances)} listed"
            )
        return values


class RatingsReturnList(pydantic.BaseModel):
    count: int
    awards: list[BattingCardRatingsModel]


class RatingsList(pydantic.BaseModel):
    ratings: List[BattingCardRatingsModel]


@router.get("")
async def get_card_ratings(
    team_id: int,
    ts: str,
    battingcard_id: list = Query(default=None),
    cardset_id: list = Query(default=None),
    vs_hand: Literal["R", "L", "vR", "vL"] = None,
    short_output: bool = False,
    csv: bool = False,
):
    this_team = Team.get_or_none(Team.id == team_id)
    logging.debug(f"Team: {this_team} / has_guide: {this_team.has_guide}")
    if this_team is None or ts != this_team.team_hash() or this_team.has_guide != 1:
        logging.warning(f"Team_id {team_id} attempted to pull ratings")
        db.close()
        raise HTTPException(
            status_code=401, detail="You are not authorized to pull card ratings."
        )
    # elif not valid_token(token):
    #     logging.warning(f'Bad Token: {token}')
    #     db.close()
    #     raise HTTPException(
    #         status_code=401,
    #         detail='You are not authorized to pull card ratings.'
    #     )

    all_ratings = BattingCardRatings.select()

    if battingcard_id is not None:
        all_ratings = all_ratings.where(
            BattingCardRatings.battingcard_id << battingcard_id
        )
    if vs_hand is not None:
        all_ratings = all_ratings.where(BattingCardRatings.vs_hand == vs_hand[-1])
    if cardset_id is not None:
        set_players = Player.select(Player.player_id).where(
            Player.cardset_id << cardset_id
        )
        set_cards = BattingCard.select(BattingCard.id).where(
            BattingCard.player << set_players
        )
        all_ratings = all_ratings.where(BattingCardRatings.battingcard << set_cards)

    if csv:
        # return_val = query_to_csv(all_ratings)
        return_vals = [model_to_dict(x) for x in all_ratings]
        for x in return_vals:
            x.update(x["battingcard"])
            x["player_id"] = x["battingcard"]["player"]["player_id"]
            del x["battingcard"], x["player"]

        db.close()
        return Response(
            content=pd.DataFrame(return_vals).to_csv(index=False), media_type="text/csv"
        )

    else:
        return_val = {
            "count": all_ratings.count(),
            "ratings": [
                model_to_dict(x, recurse=not short_output) for x in all_ratings
            ],
        }
        db.close()
        return return_val


def get_scouting_dfs(cardset_id: list = None):
    all_ratings = BattingCardRatings.select()
    if cardset_id is not None:
        set_players = Player.select(Player.player_id).where(
            Player.cardset_id << cardset_id
        )
        set_cards = BattingCard.select(BattingCard.id).where(
            BattingCard.player << set_players
        )
        all_ratings = all_ratings.where(BattingCardRatings.battingcard << set_cards)

    vl_query = all_ratings.where(BattingCardRatings.vs_hand == "L")
    vr_query = all_ratings.where(BattingCardRatings.vs_hand == "R")

    vl_vals = [model_to_dict(x) for x in vl_query]
    for x in vl_vals:
        x.update(x["battingcard"])
        x["player_id"] = x["battingcard"]["player"]["player_id"]
        x["player_name"] = x["battingcard"]["player"]["p_name"]
        x["rarity"] = x["battingcard"]["player"]["rarity"]["name"]
        x["cardset_id"] = x["battingcard"]["player"]["cardset"]["id"]
        x["cardset_name"] = x["battingcard"]["player"]["cardset"]["name"]
        del x["battingcard"]
        del x["player"]

    vr_vals = [model_to_dict(x) for x in vr_query]
    for x in vr_vals:
        x["player_id"] = x["battingcard"]["player"]["player_id"]
        del x["battingcard"]

    vl = pd.DataFrame(vl_vals)
    vr = pd.DataFrame(vr_vals)

    bat_df = pd.merge(vl, vr, on="player_id", suffixes=("_vl", "_vr")).set_index(
        "player_id", drop=False
    )

    logging.debug(f"bat_df: {bat_df}")

    positions = CardPosition.select()
    if cardset_id is not None:
        set_players = Player.select(Player.player_id).where(
            Player.cardset_id << cardset_id
        )
        positions = positions.where(CardPosition.player << set_players)

    series_list = []
    for pos_code in ["P", "C", "1B", "2B", "3B", "SS", "LF", "CF", "RF"]:
        series_list.append(
            pd.Series(
                dict(
                    [
                        (x.player.player_id, x.range)
                        for x in positions.where(CardPosition.position == pos_code)
                    ]
                ),
                name=f"Range {pos_code}",
            )
        )
        series_list.append(
            pd.Series(
                dict(
                    [
                        (x.player.player_id, x.error)
                        for x in positions.where(CardPosition.position == pos_code)
                    ]
                ),
                name=f"Error {pos_code}",
            )
        )

    series_list.append(
        pd.Series(
            dict(
                [
                    (x.player.player_id, x.arm)
                    for x in positions.where(
                        CardPosition.position << ["LF", "CF", "RF"]
                    )
                ]
            ),
            name=f"Arm OF",
        )
    )
    series_list.append(
        pd.Series(
            dict(
                [
                    (x.player.player_id, x.arm)
                    for x in positions.where(CardPosition.position == "C")
                ]
            ),
            name=f"Arm C",
        )
    )
    series_list.append(
        pd.Series(
            dict(
                [
                    (x.player.player_id, x.pb)
                    for x in positions.where(CardPosition.position == "C")
                ]
            ),
            name=f"PB C",
        )
    )
    series_list.append(
        pd.Series(
            dict(
                [
                    (x.player.player_id, x.overthrow)
                    for x in positions.where(CardPosition.position == "C")
                ]
            ),
            name=f"Throw C",
        )
    )
    db.close()
    logging.debug(f"series_list: {series_list}")

    return bat_df.join(series_list)


@router.get("/scouting")
async def get_card_scouting(team_id: int, ts: str):
    this_team = Team.get_or_none(Team.id == team_id)
    logging.debug(f"Team: {this_team} / has_guide: {this_team.has_guide}")
    if this_team is None or ts != this_team.team_hash() or this_team.has_guide != 1:
        logging.warning(f"Team_id {team_id} attempted to pull ratings")
        db.close()
        return (
            "Your team does not have the ratings guide enabled. If you have purchased a copy ping Cal to "
            "make sure it is enabled on your team. If you are interested you can pick it up here (thank you!): "
            "https://ko-fi.com/manticorum/shop"
        )

    if os.path.isfile(f"storage/batting-ratings.csv"):
        return FileResponse(
            path=f"storage/batting-ratings.csv",
            media_type="text/csv",
            # headers=headers
        )

    raise HTTPException(
        status_code=400, detail="Go pester Cal - the scouting file is missing"
    )


@router.post("/calculate/scouting", include_in_schema=PRIVATE_IN_SCHEMA)
async def post_calc_scouting(token: str = Depends(oauth2_scheme)):
    if not valid_token(token):
        logging.warning(f"Bad Token: {token}")
        db.close()
        raise HTTPException(
            status_code=401, detail="You are not authorized to calculate card ratings."
        )

    logging.warning(f"Re-calculating batting ratings\n\n")

    output = get_scouting_dfs()
    first = ["player_id", "player_name", "cardset_name", "rarity", "hand", "variant"]
    exclude = first + ["id_vl", "id_vr", "vs_hand_vl", "vs_hand_vr"]
    output = output[first + [col for col in output.columns if col not in exclude]]

    csv_file = pd.DataFrame(output).to_csv(index=False)
    with open(RATINGS_FILE, "w") as file:
        file.write(csv_file)

    return Response(content=csv_file, media_type="text/csv")


@router.get("/basic")
async def get_basic_scouting(cardset_id: list = Query(default=None)):
    if os.path.isfile(f"storage/batting-basic.csv"):
        return FileResponse(
            path=f"storage/batting-basic.csv",
            media_type="text/csv",
            # headers=headers
        )

    raise HTTPException(
        status_code=400, detail="Go pester Cal - the scouting file is missing"
    )


@router.post("/calculate/basic", include_in_schema=PRIVATE_IN_SCHEMA)
async def post_calc_basic(token: str = Depends(oauth2_scheme)):
    if not valid_token(token):
        logging.warning(f"Bad Token: {token}")
        db.close()
        raise HTTPException(
            status_code=401, detail="You are not authorized to calculate basic ratings."
        )

    logging.warning(f"Re-calculating basic batting ratings\n\n")

    raw_data = get_scouting_dfs()
    logging.debug(f"output: {raw_data}")

    def get_raw_speed(df_data):
        speed_raw = df_data["running"] / 20 + df_data["steal_jump"]
        if df_data["steal_auto"]:
            speed_raw += 0.5
        return speed_raw

    raw_series = raw_data.apply(get_raw_speed, axis=1)
    rank_series = raw_series.rank(pct=True)
    raw_data["Speed"] = round(rank_series * 100)

    def get_raw_steal(df_data):
        return ((df_data["steal_high"] / 20) + (df_data["steal_low"] / 20)) * df_data[
            "steal_jump"
        ]

    raw_series = raw_data.apply(get_raw_steal, axis=1)
    rank_series = raw_series.rank(pct=True)
    raw_data["Steal"] = round(rank_series * 100)

    def get_raw_reaction(df_data):
        raw_total = 0
        for pos_range in [
            df_data["Range C"],
            df_data["Range 1B"],
            df_data["Range 2B"],
            df_data["Range 3B"],
            df_data["Range SS"],
            df_data["Range LF"],
            df_data["Range CF"],
            df_data["Range RF"],
        ]:
            if pd.notna(pos_range):
                raw_total += 10 ** (5 - pos_range)
        return raw_total

    raw_series = raw_data.apply(get_raw_reaction, axis=1)
    rank_series = raw_series.rank(pct=True)
    raw_data["Reaction"] = round(rank_series * 100)

    def get_raw_arm(df_data):
        of_arm = None
        of_pos = None
        if pd.notna(df_data["Range RF"]):
            of_pos = "RF"
        elif pd.notna(df_data["Range CF"]):
            of_pos = "CF"
        elif pd.notna(df_data["Range LF"]):
            of_pos = "LF"

        if of_pos is not None:
            if df_data["Arm OF"] < 0:
                of_raw = df_data["Arm OF"] * -10
            else:
                of_raw = 5 - df_data["Arm OF"]

            if of_pos == "RF":
                of_raw = of_raw * 1.5
                of_raw += (6 - df_data["Range RF"]) * 4
            elif of_pos == "CF":
                of_raw += (6 - df_data["Range CF"]) * 3
            elif of_pos == "LF":
                of_raw = of_raw / 2
                of_raw += (6 - df_data["Range LF"]) * 2

            of_arm = of_raw

        if_arm = None
        if (
            pd.notna(df_data["Range 3B"])
            or pd.notna(df_data["Range 2B"])
            or pd.notna(df_data["Range 1B"])
            or pd.notna(df_data["Range SS"])
        ):
            range_totals = 0
            if pd.notna(df_data["Range 3B"]):
                range_totals += (6 - df_data["Range 3B"]) * 5
            if pd.notna(df_data["Range SS"]):
                range_totals += (6 - df_data["Range SS"]) * 4
            if pd.notna(df_data["Range 2B"]):
                range_totals += (6 - df_data["Range 2B"]) * 3
            if pd.notna(df_data["Range 1B"]):
                range_totals += 6 - df_data["Range 1B"]
            if_arm = 100 - (50 - range_totals)

        c_arm = None
        if pd.notna(df_data["Arm C"]):
            if df_data["Arm C"] == -5:
                c_arm = 100
            else:
                temp_arm = (
                    20
                    + ((10 - df_data["Arm C"]) * 3)
                    + (20 - df_data["PB C"])
                    + (20 - df_data["Throw C"])
                    - df_data["Error C"]
                )
                c_arm = min(100, temp_arm)

        if c_arm is not None:
            return c_arm
        elif of_arm is not None:
            return of_arm
        elif if_arm is not None:
            return if_arm
        else:
            return 1

    raw_series = raw_data.apply(get_raw_arm, axis=1)
    rank_series = raw_series.rank(pct=True)
    raw_data["Arm"] = round(rank_series * 100)

    def get_raw_fielding(df_data):
        if_error, of_error, c_error = 0, 0, 0
        denom = 0
        if (
            pd.notna(df_data["Error 3B"])
            or pd.notna(df_data["Error 2B"])
            or pd.notna(df_data["Error 1B"])
            or pd.notna(df_data["Error SS"])
        ):
            raw_if = 100
            if pd.notna(df_data["Error 3B"]):
                raw_if -= df_data["Error 3B"] * 2
            if pd.notna(df_data["Error SS"]):
                raw_if -= df_data["Error SS"] * 0.75
            if pd.notna(df_data["Error 2B"]):
                raw_if -= df_data["Error 2B"] * 1.25
            if pd.notna(df_data["Error 1B"]):
                raw_if -= df_data["Error 1B"] * 2

            if_error = max(1, raw_if)
            denom += 1

        if (
            pd.notna(df_data["Error LF"])
            or pd.notna(df_data["Error CF"])
            or pd.notna(df_data["Error RF"])
        ):
            raw_of = 100
            if pd.notna(df_data["Error LF"]):
                raw_of -= df_data["Error LF"] * 2
            if pd.notna(df_data["Error CF"]):
                raw_of -= df_data["Error CF"] * 0.75
            if pd.notna(df_data["Error RF"]):
                raw_of -= df_data["Error RF"] * 1.25

            of_error = max(1, raw_of)
            denom += 1

        if pd.notna(df_data["Error C"]):
            c_error = max(
                100 - (df_data["Error C"] * 5) - df_data["Throw C"] - df_data["PB C"], 1
            )
            denom += 1

        return sum([if_error, of_error, c_error]) / max(denom, 1)

    raw_series = raw_data.apply(get_raw_fielding, axis=1)
    rank_series = raw_series.rank(pct=True)
    raw_data["Fielding"] = round(rank_series * 100)

    rank_series = raw_data["avg_vl"].rank(pct=True)
    raw_data["Contact L"] = round(rank_series * 100)

    rank_series = raw_data["avg_vr"].rank(pct=True)
    raw_data["Contact R"] = round(rank_series * 100)

    rank_series = raw_data["slg_vl"].rank(pct=True)
    raw_data["Power L"] = round(rank_series * 100)

    rank_series = raw_data["slg_vr"].rank(pct=True)
    raw_data["Power R"] = round(rank_series * 100)

    def get_raw_vision(df_data):
        return (
            (
                ((df_data["obp_vr"] * 0.67) + (df_data["obp_vl"] * 0.33))
                - ((df_data["avg_vr"] * 0.67) + (df_data["avg_vl"] * 0.33))
            )
            * 5
        ) - (
            ((df_data["strikeout_vl"] * 0.33) + (df_data["strikeout_vr"] * 0.67)) / 208
        )

    raw_series = raw_data.apply(get_raw_vision, axis=1)
    rank_series = raw_series.rank(pct=True)
    raw_data["Vision"] = round(rank_series * 100)

    def get_raw_rating(df_data):
        return (
            ((df_data["Reaction"] + df_data["Arm"] + df_data["Fielding"]) * 2)
            + (df_data["Speed"] + df_data["Steal"])
            + (
                (
                    ((df_data["Contact R"] + df_data["Power R"]) * 0.67)
                    + ((df_data["Contact L"] + df_data["Power L"]) * 0.33)
                    + df_data["Vision"]
                )
                * 6
            )
        )

    raw_series = raw_data.apply(get_raw_rating, axis=1)
    rank_series = raw_series.rank(pct=True)
    raw_data["Rating"] = round(rank_series * 100)

    output = raw_data[
        [
            "player_id",
            "player_name",
            "Rating",
            "Contact R",
            "Contact L",
            "Power R",
            "Power L",
            "Vision",
            "Speed",
            "Steal",
            "Reaction",
            "Arm",
            "Fielding",
            "hand",
            "cardset_name",
        ]
    ]

    csv_file = pd.DataFrame(output).to_csv(index=False)
    with open(BASIC_FILE, "w") as file:
        file.write(csv_file)

    return Response(content=csv_file, media_type="text/csv")


@router.get("/{ratings_id}")
async def get_one_rating(ratings_id: int, token: str = Depends(oauth2_scheme)):
    if not valid_token(token):
        logging.warning(f"Bad Token: {token}")
        db.close()
        raise HTTPException(
            status_code=401, detail="You are not authorized to pull card ratings."
        )

    this_rating = BattingCardRatings.get_or_none(BattingCardRatings.id == ratings_id)
    if this_rating is None:
        db.close()
        raise HTTPException(
            status_code=404, detail=f"BattingCardRating id {ratings_id} not found"
        )

    r_data = model_to_dict(this_rating)
    db.close()
    return r_data


@router.get("/player/{player_id}")
async def get_player_ratings(
    player_id: int,
    variant: list = Query(default=None),
    short_output: bool = False,
    token: str = Depends(oauth2_scheme),
):
    if not valid_token(token):
        logging.warning(f"Bad Token: {token}")
        db.close()
        raise HTTPException(
            status_code=401, detail="You are not authorized to pull card ratings."
        )

    all_cards = (
        BattingCard.select()
        .where(BattingCard.player_id == player_id)
        .order_by(BattingCard.variant)
    )
    if variant is not None:
        all_cards = all_cards.where(BattingCard.variant << variant)

    all_ratings = BattingCardRatings.select().where(
        BattingCardRatings.battingcard << all_cards
    )

    return_val = {
        "count": all_ratings.count(),
        "ratings": [model_to_dict(x, recurse=not short_output) for x in all_ratings],
    }
    db.close()
    return return_val


@router.put("", include_in_schema=PRIVATE_IN_SCHEMA)
async def put_ratings(ratings: RatingsList, token: str = Depends(oauth2_scheme)):
    if not valid_token(token):
        logging.warning(f"Bad Token: {token}")
        db.close()
        raise HTTPException(
            status_code=401, detail="You are not authorized to post card ratings."
        )

    new_ratings = []
    updates = 0
    for x in ratings.ratings:
        try:
            BattingCardRatings.get(
                (BattingCardRatings.battingcard_id == x.battingcard_id)
                & (BattingCardRatings.vs_hand == x.vs_hand)
            )
            updates += (
                BattingCardRatings.update(x.dict())
                .where(
                    (BattingCardRatings.battingcard_id == x.battingcard_id)
                    & (BattingCardRatings.vs_hand == x.vs_hand)
                )
                .execute()
            )
        except BattingCardRatings.DoesNotExist:
            new_ratings.append(x.dict())

    with db.atomic():
        # Use PostgreSQL-compatible upsert helper
        upsert_batting_card_ratings(new_ratings, batch_size=30)

    db.close()
    return f"Updated ratings: {updates}; new ratings: {len(new_ratings)}"


@router.delete("/{ratings_id}", include_in_schema=PRIVATE_IN_SCHEMA)
async def delete_rating(ratings_id: int, token: str = Depends(oauth2_scheme)):
    if not valid_token(token):
        logging.warning(f"Bad Token: {token}")
        db.close()
        raise HTTPException(
            status_code=401, detail="You are not authorized to post card ratings."
        )

    this_rating = BattingCardRatings.get_or_none(BattingCardRatings.id == ratings_id)
    if this_rating is None:
        db.close()
        raise HTTPException(
            status_code=404, detail=f"BattingCardRating id {ratings_id} not found"
        )

    count = this_rating.delete_instance()
    db.close()

    if count == 1:
        return f"Rating {this_rating} has been deleted"
    else:
        raise HTTPException(
            status_code=500, detail=f"Rating {this_rating} could not be deleted"
        )