jiangcang_uo/follow_up.py

import os
import datetime
import time
import pandas as pd


def follow_up_handle():
    '''后续处理'''
    object_dir = "./predictions"
    output_dir = "./keep"

    # 创建输出目录
    os.makedirs(output_dir, exist_ok=True)   

    # 检查目录是否存在
    if not os.path.exists(object_dir):
        print(f"目录不存在: {object_dir}")
        return

    # 获取所有以 future_predictions_ 开头的 CSV 文件
    csv_files = []
    for file in os.listdir(object_dir):
        if file.startswith("future_predictions_") and file.endswith(".csv"):
            csv_files.append(file)
    
    if not csv_files:
        print(f"在 {object_dir} 中没有找到 future_predictions_ 开头的 CSV 文件")
        return
    
    csv_files.sort()

    # 调试分支
    # target_time = "202603271700"
    # matching_files = [f for f in csv_files if target_time in f]
    # if matching_files:
    #     last_csv_file = matching_files[0]
    #     print(f"指定时间的文件: {last_csv_file}")
    # else:
    #     print(f"未找到时间 {target_time} 的预测文件")
    #     return

    # 正式分支
    last_csv_file = csv_files[-1]   # 只看最新预测的文件
    print(f"最新预测文件: {last_csv_file}")
    if last_csv_file.startswith("future_predictions_") and last_csv_file.endswith(".csv"):
        target_time = last_csv_file.replace("future_predictions_", "").replace(".csv", "")
    else:
        target_time = datetime.datetime.now().strftime("%Y%m%d%H%M")
    
    # 读取最新预测文件
    last_csv_path = os.path.join(object_dir, last_csv_file)
    df_last_predict = pd.read_csv(last_csv_path)

    df_last_predict_will_drop = df_last_predict[df_last_predict["will_price_drop"] == 1].reset_index(drop=True)
    df_last_predict_not_drop = df_last_predict[df_last_predict["will_price_drop"] == 0].reset_index(drop=True)
    print(f"最新预测文件中，预测降价的航班有 {len(df_last_predict_will_drop)} 条，预测不降价的航班有 {len(df_last_predict_not_drop)} 条")

    # 建一张 维护表 keep_info.csv  附加一个维护表快照 keep_info_{target_time}.csv
    keep_info_path = os.path.join(output_dir, "keep_info.csv")
    keep_info_snapshot_path = os.path.join(output_dir, f"keep_info_{target_time}.csv")
    key_cols = ["citypair", "flight_numbers", "baggage_weight", "from_date"]

    # 去重操作
    df_last_predict_will_drop = df_last_predict_will_drop.drop_duplicates(
        subset=key_cols, keep="last"
    ).reset_index(drop=True)
    # df_last_predict_not_drop = df_last_predict_not_drop.drop_duplicates(
    #     subset=key_cols, keep="last"
    # ).reset_index(drop=True)

    # 读取维护表
    if os.path.exists(keep_info_path):
        try:
            df_keep_info = pd.read_csv(keep_info_path)
        except Exception as e:
            print(f"读取维护表失败: {keep_info_path}, error: {str(e)}")
            df_keep_info = pd.DataFrame()
    else:
        df_keep_info = pd.DataFrame()
    
    def _parse_dt(yyyymmddhhmm):
        try:
            return datetime.datetime.strptime(str(yyyymmddhhmm), "%Y%m%d%H%M")
        except Exception:
            return None
    
    current_dt = _parse_dt(target_time)
    prev_dt = None
    hud_decrement = 1

    # if not df_keep_info.empty and "last_predict_time" in df_keep_info.columns:
    #     prev_candidates = (
    #         df_keep_info["last_predict_time"].dropna().astype(str).tolist()
    #     )
    #     if prev_candidates:
    #         prev_dt = _parse_dt(max(prev_candidates))

    if prev_dt is None:
        snapshot_times = []
        for f in os.listdir(output_dir):
            if (
                f.startswith("keep_info_")
                and f.endswith(".csv")
                and f != f"keep_info_{target_time}.csv"
            ):
                ts = f.replace("keep_info_", "").replace(".csv", "")
                dt = _parse_dt(ts)
                if dt is not None:
                    snapshot_times.append(dt)
        if snapshot_times:
            prev_dt = max(snapshot_times)
    
    if current_dt is not None and prev_dt is not None:
        delta_seconds = (current_dt - prev_dt).total_seconds()
        if delta_seconds >= 0:
            hud_decrement = max(0, int(delta_seconds // 3600))
        else:
            hud_decrement = 0
    
    # 初始化维护表
    if df_keep_info.empty:
        df_keep_info = df_last_predict_will_drop.copy()
        df_keep_info["into_update_hour"] = df_keep_info['update_hour']
        # df_keep_info["into_price"] = df_keep_info['price_total']
        df_keep_info["keep_flag"] = 1
        # df_keep_info["last_predict_time"] = target_time

        df_keep_info.to_csv(keep_info_snapshot_path, index=False, encoding="utf-8-sig")
        print(f"维护表快照已保存: {keep_info_snapshot_path} (rows={len(df_keep_info)})")

        df_keep_info.to_csv(keep_info_path, index=False, encoding="utf-8-sig")
        print(f"维护表已初始化: {keep_info_path} (rows={len(df_keep_info)})")
    
    # 已存在维护表
    else:
        if "keep_flag" not in df_keep_info.columns:
            df_keep_info["keep_flag"] = 0

        df_keep_info["keep_flag"] = (
            pd.to_numeric(df_keep_info["keep_flag"], errors="coerce")
            .fillna(0)
            .astype(int)
        )

        missing_cols = [c for c in key_cols if c not in df_keep_info.columns]
        if missing_cols:
            print(f"维护表缺少字段: {missing_cols}, path={keep_info_path}")
            return
        
        for c in key_cols:
            df_last_predict_will_drop[c] = df_last_predict_will_drop[c].astype(str)
            # df_last_predict_not_drop[c] = df_last_predict_not_drop[c].astype(str)
            df_keep_info[c] = df_keep_info[c].astype(str)
        
        df_keep_info = df_keep_info.drop_duplicates(subset=key_cols, keep="last").reset_index(drop=True)

        # 提取两者的标志位
        df_last_keys = df_last_predict_will_drop[key_cols].drop_duplicates().reset_index(drop=True)
        df_keep_keys = df_keep_info[key_cols].drop_duplicates().reset_index(drop=True)

        df_last_with_merge = df_last_predict_will_drop.merge(
            df_keep_keys, on=key_cols, how="left", indicator=True
        )

        # 场景一: 如果某一行数据在 df_last_predict_will_drop 出现，没有在 df_keep_info 里
        df_to_add = (
            df_last_with_merge.loc[df_last_with_merge["_merge"] == "left_only"]
            .drop(columns=["_merge"])
            .copy()
        )

        # keep_flag 设为 1
        if not df_to_add.empty:
            df_to_add['into_update_hour'] = df_to_add['update_hour']
            # df_to_add['into_price'] = df_to_add['price_total']
            df_to_add["keep_flag"] = 1
        
        df_keep_with_merge = df_keep_info.reset_index().merge(
            df_last_keys, on=key_cols, how="left", indicator=True
        )

        # 场景二: 如果某一行数据在 df_last_predict_will_drop 和 df_keep_info 里都出现
        matched_idx = df_keep_with_merge.loc[df_keep_with_merge["_merge"] == "both", "index"].tolist()
        # 场景三: 如果某一行数据在 df_last_predict_will_drop 没有出现，却在 df_keep_info 里都出现
        keep_only_idx = df_keep_with_merge.loc[df_keep_with_merge["_merge"] == "left_only", "index"].tolist()

        # 符合场景二的索引 (在 df_keep_with_merge 中)
        if matched_idx:
            df_matched_keys = df_keep_info.loc[matched_idx, key_cols]
            df_latest_matched = df_matched_keys.merge(
                df_last_predict_will_drop, on=key_cols, how="left"
            )
            # 将 df_keep_info 的 df_matched_keys 的内容更新为 df_last_predict_will_drop 里对应的内容
            update_cols = [c for c in df_last_predict_will_drop.columns if c not in key_cols]
            for c in update_cols:
                if c == "keep_flag":
                    continue
                if c not in df_keep_info.columns:
                    df_keep_info[c] = pd.NA
                df_keep_info.loc[matched_idx, c] = df_latest_matched[c].values
            
            # 重新标记 原来是1 -> 0  原来是0 -> 0  原来是2 -> 0, 原来是-1 -> 1
            old_flags = df_keep_info.loc[matched_idx, "keep_flag"]
            df_keep_info.loc[matched_idx, "keep_flag"] = old_flags.apply(
                lambda x: 0 if x in (0, 1, 2) else (1 if x == -1 else 1)
            )
            
        # 符合场景三的索引 (在 df_keep_with_merge 中)
        if keep_only_idx:
            mask_keep_only = df_keep_info.index.isin(keep_only_idx)  # 布尔索引序列
            
            # 如果 df_keep_info 的 keep_flag 大于等于0
            mask_need_observe = mask_keep_only & (df_keep_info["keep_flag"] >= 0)  # 布尔索引序列
            if mask_need_observe.any():
                if "hours_until_departure" not in df_keep_info.columns:
                    df_keep_info.loc[mask_need_observe, "keep_flag"] = -1
                else:
                    hud = pd.to_numeric(
                        df_keep_info.loc[mask_need_observe, "hours_until_departure"],
                        errors="coerce",
                    )
                    # hours_until_departure自动减1
                    # new_hud = hud - 1
                    new_hud = hud - hud_decrement
                    df_keep_info.loc[mask_need_observe, "hours_until_departure"] = new_hud
                    df_keep_info.loc[mask_need_observe, "keep_flag"] = -1       # 删除标志

        # 将 df_to_add 添加到 df_keep_info 之后
        add_rows = len(df_to_add) if "df_to_add" in locals() else 0
        if add_rows:
            df_keep_info = pd.concat([df_keep_info, df_to_add], ignore_index=True)
        
        df_keep_info_snapshot = df_keep_info.copy()
        df_keep_info_snapshot.to_csv(keep_info_snapshot_path, index=False, encoding="utf-8-sig")
        print(
            f"维护表快照已保存: {keep_info_snapshot_path} (rows={len(df_keep_info_snapshot)})"
        )

        # 移除 keep_flag 为 -1 的行
        before_rm = len(df_keep_info)
        df_keep_info = df_keep_info.loc[df_keep_info["keep_flag"] != -1].reset_index(drop=True)
        rm_rows = before_rm - len(df_keep_info)

        # 保存更新后的 df_keep_info 到维护表csv文件
        df_keep_info.to_csv(keep_info_path, index=False, encoding="utf-8-sig")
        print(
            f"维护表已更新: {keep_info_path} (rows={len(df_keep_info)} add={add_rows} rm={rm_rows})"
        )

    pass

if __name__ == "__main__":
    time.sleep(5)
    follow_up_handle()
    time.sleep(5)