Aviasales Junior — рефакторинг исследовательского скрипта для регулярного запуска

import pandas as pd, numpy as np
from datetime import datetime, timedelta, date
from dateutil.relativedelta import relativedelta
import matplotlib.pyplot as plt
import warnings
warnings.filterwarnings("ignore")
 
today_dt = (datetime.today()).date()
today = today_dt.strftime('%Y-%m-%d')
yesterday_dt = today_dt - timedelta(days=1)
yesterday = yesterday_dt.strftime('%Y-%m-%d')
two_weeks_ago_start = (today_dt - timedelta(days=14)).strftime('%Y-%m-%d')
last_year_day = (yesterday_dt - relativedelta(years=1)).strftime('%Y-%m-%d')
 
total_aviasales_bookings = session.query(f"""
    select to_date(booked_at) as period, count(booking_id) as total_bookings_aviasales
    from bd.bookings
    where to_date(booked_at) >= '{two_weeks_ago_start}' and to_date(booked_at) <= '{yesterday}'
    group by 1 order by 1
""")
 
total_aviasales_gross_profit = session.query(f"""
    select to_date(booked_at) as period, sum(profit) as total_profit_aviasales
    from bd.bookings
    where to_date(booked_at) >= '{two_weeks_ago_start}' and to_date(booked_at) <= '{yesterday}'
    group by 1 order by 1
""")
 
# ... ещё две похожие выгрузки + 4 одинаковых по структуре графика

"""
aviasales_daily_dashboard.py
 
Регулярный скрипт сборки графиков:
- бронирования и прибыль за последние 2 недели
- доля заказов с sender_id ∈ (10,20,30) — 2 недели и год
"""
 
from __future__ import annotations
import logging
from dataclasses import dataclass
from datetime import date, timedelta
from pathlib import Path
 
import pandas as pd
import matplotlib.pyplot as plt
from dateutil.relativedelta import relativedelta
 
logging.basicConfig(level=logging.INFO, format="%(asctime)s [%(levelname)s] %(message)s")
log = logging.getLogger(__name__)
 
 
# --- 1. Конфиг -----------------------------------------------------------
@dataclass(frozen=True)
class Config:
    senders_of_interest: tuple = (10, 20, 30)
    short_window_days:    int  = 14
    long_window_years:    int  = 1
    output_dir:          Path = Path("artifacts")
    last_marker_color:    str  = "red"
 
 
# --- 2. Даты и окна ------------------------------------------------------
def windows(today: date, cfg: Config):
    yesterday = today - timedelta(days=1)
    short_start = today - timedelta(days=cfg.short_window_days)
    long_start = yesterday - relativedelta(years=cfg.long_window_years)
    return {
        "yesterday": yesterday,
        "short_start": short_start,
        "long_start": long_start,
    }
 
 
# --- 3. Параметризованные SQL-выгрузки ----------------------------------
SQL_BOOKINGS = """
    SELECT to_date(booked_at) AS period,
           COUNT(booking_id)  AS bookings,
           SUM(profit)        AS profit,
           COUNT(IF(sender_id IN :senders, booking_id, NULL)) AS bookings_sender
    FROM bd.bookings
    WHERE to_date(booked_at) BETWEEN :start AND :end
    GROUP BY 1 ORDER BY 1
"""
 
 
def fetch_period(session, start: date, end: date, senders: tuple) -> pd.DataFrame:
    df = session.query(SQL_BOOKINGS, params={
        "senders": senders, "start": start, "end": end
    })
    df["period"] = pd.to_datetime(df["period"])
    df["share_of_sender"] = df["bookings_sender"] / df["bookings"] * 100
    return df
 
 
# --- 4. Визуализация -----------------------------------------------------
def plot_panel(short: pd.DataFrame, long: pd.DataFrame, cfg: Config, out: Path):
    fig, axs = plt.subplots(2, 2, figsize=(12, 8))
    fig.suptitle("Daily metrics — Aviasales", fontsize=16)
 
    panels = [
        (axs[0, 0], short, "bookings",        "Bookings (2 weeks)"),
        (axs[0, 1], short, "profit",          "Profit (2 weeks)"),
        (axs[1, 0], short, "share_of_sender", "Sender share, % (2 weeks)"),
        (axs[1, 1], long,  "share_of_sender", "Sender share, % (year)"),
    ]
    for ax, df, col, title in panels:
        if df.empty:
            log.warning("No data for %s", title)
            ax.set_title(f"{title} — no data"); continue
        ax.plot(df["period"], df[col], marker="o")
        ax.scatter(df["period"].iloc[-1], df[col].iloc[-1],
                   color=cfg.last_marker_color, s=80, label="last day")
        ax.set_title(title); ax.set_xlabel("date"); ax.legend()
    fig.autofmt_xdate()
    out.mkdir(parents=True, exist_ok=True)
    path = out / f"daily_{date.today().isoformat()}.png"
    fig.savefig(path, dpi=120, bbox_inches="tight")
    log.info("Saved %s", path)
 
 
def main(session, today: date | None = None, cfg: Config = Config()):
    today = today or date.today()
    w = windows(today, cfg)
    log.info("Windows: %s", w)
    short = fetch_period(session, w["short_start"], w["yesterday"], cfg.senders_of_interest)
    long_ = fetch_period(session, w["long_start"],  w["yesterday"], cfg.senders_of_interest)
    plot_panel(short, long_, cfg, cfg.output_dir)
 
 
if __name__ == "__main__":
    from analytics_db import session  # placeholder: ваш коннектор
    main(session)