logistics/物流t投递审核.py

"""
fetch_order_data函数只是获取源数据,是一个sql语句，可以更改
cal_min_fee 函数是分别以一票一件和一票多件计算出最小的费用和渠道
analyze_orders 订单层面的业务逻辑判断，防止出现混合渠道投递，卡派订单包含多个不同快递追踪单号，多渠道订单总重量小于1000KG(因为1000KG以内一个卡派可以搞定，不应该出现多渠道)
analyze_logistics 真正的物流投递层面去分析，先判断投递渠道和最优渠道是否一致，再判断偶发估算费用和最优渠道费用是否一致
"""
import pandas as pd
from utils.gtools import MySQLconnect
from utils.logisticsBill import BillFactory, Billing
from utils.countryOperator import OperateCountry
from utils.Package import Package, Package_group
from utils.logistics_name_config import logistics_name
from datetime import date
# 货币转换，其他转RMB
def convert_currency(amount, current_currency):
    """
    货币转换
    """
    if amount is None or amount ==0:
        return "金额为空"
    if amount >=9999:
        return "无可用渠道"
    if current_currency == "USD":
        amount=amount*7
    elif current_currency == "GBP":
        amount =amount*9
    elif current_currency == "EUR":
        amount = amount*8
    elif current_currency == "AUD":
        amount = amount*5
    elif current_currency == "CAD":
        amount = amount*5
    elif current_currency == "JPY":
        amount =amount*0.05
    return amount
        
# 获取数据
def fetch_order_data():
    """从数据库获取原始订单数据"""
    with MySQLconnect('ods') as db:
        sql = """
        SELECT
        DATE_FORMAT(ol.order_date, '%%Y-%%m-%%d') AS order_date,
        DATE_FORMAT(oe.投递时间, '%%Y-%%m-%%d') AS 投递时间,
        ol.fund_status,
        oe.`包裹状态`,
            oe.包裹号 AS package,
            oe.单号 AS order_id,
                        oe.运输方式,
                        oe.`目的国`,
            ol.postcode AS postcode,
                        oe.`快递分区`,
            oe.快递跟踪号,
            ecm.类型 AS 渠道类型, -- 包裹类型
            pvi.length AS 长,
            pvi.width AS 宽,
            pvi.hight AS 高,
            pvi.weight AS 重量,
            pfi.express_fee AS 基础估算,
            pfi.express_additional_fee AS 偶发估算,
            pfi.express_fee + pfi.express_additional_fee AS 包裹总估算,
            oe.快递公司 AS 投递渠道
        FROM
            ods.order_express oe
            LEFT JOIN ods.express_company ecm ON oe.快递公司 = ecm.快递公司
            LEFT JOIN ods.package_vol_info pvi ON oe.包裹号 = pvi.package
            LEFT JOIN ods.package_fee_info pfi ON oe.包裹号 = pfi.package
            LEFT JOIN ods.order_list ol ON oe.单号 = ol.order_id
        WHERE
            oe.包裹状态 not REGEXP '已作废|--|客户签收'
            # AND oe.`快递公司` NOT REGEXP "--"
            AND `卡板发货时间` REGEXP "--"
            AND ol.fund_status NOT REGEXP '等待|全额退款'
            AND ol.site_name REGEXP 'litfad|kwoking|lakiq'
            AND oe.投递时间 >= DATE_SUB(NOW(), INTERVAL 3 DAY) 
			AND pvi.length>0 AND pvi.width >0 AND pvi.hight>0 AND pvi.weight>0
            and oe.目的国 regexp 'United States|Australia|United Kingdom|Germany|France|Spain|Italy|Netherlands|Belgium'
        order by ol.order_id,ol.order_date
        
        """

        return pd.read_sql(sql, db.engine())
    

def cal_min_fee(raw_data: pd.DataFrame):
    """
    处理物流费用数据并实现业务逻辑判断
    1.用
    """    
    df = raw_data.copy()
    # 包裹层面审核
    for order_id, group in df.groupby('order_id'):
        package_group = Package_group()
        opCountry = OperateCountry(group['目的国'].iloc[0])
        express_fee = 0
        express_type=''
        for index, row in group.iterrows():
            # 计算一票一件
            packages=Package_group()
            package = Package(row['package'], row['长'], row['宽'], row['高'], row['重量'])
            packages.add_package(package)
            bill_express = Billing("1",opCountry,packages,row['postcode'],company_name=None,head_type=1,beizhu="") 
            if bill_express.tail_amount[0] == 0 or bill_express.tail_amount[0] >=9999:
                df.loc[index,"单票最小费用"] = ""
                df.loc[index,"单票渠道"] = ""
                express_fee = 999999
                express_type = '不可派'
            else:
                df.loc[index,"单票最小费用"] = bill_express.tail_amount[0]
                df.loc[index,"单票渠道"] = bill_express.company_name
                express_fee += bill_express.tail_amount[0]
                express_type = bill_express.logistic_type
                if bill_express.logistic_type == '卡派':
                    express_type = '卡派单包裹'
                
            # 计算一票多件
            package_group.add_package(package)
        # 计算一票多件
        if len(package_group) > 1:
            bill_ltl = Billing("1",opCountry,package_group,row['postcode'],company_name=None,head_type=1,beizhu="") 
            if bill_ltl.tail_amount[0] == 0 or bill_ltl.tail_amount[0] >=9999:
                df.loc[df['order_id']==order_id,'多票最小费用'] = ""
                df.loc[df['order_id']==order_id,'多票渠道'] = "不可派"
            df.loc[df['order_id']==order_id,'多票最小费用'] = bill_ltl.tail_amount[0]/len(package_group)
            df.loc[df['order_id']==order_id,'多票渠道'] = bill_ltl.company_name
            min_fee = min(bill_ltl.tail_amount[0],express_fee)
            df.loc[df['order_id']==order_id,'最优总费用'] = min_fee
            df.loc[df['order_id']==order_id,'最优渠道类型'] =  bill_ltl.logistic_type if min_fee == bill_ltl.tail_amount[0] else express_type
        else:
            min_fee = express_fee
            df.loc[df['order_id']==order_id,'最优总费用'] = min_fee
            df.loc[df['order_id']==order_id,'最优渠道类型'] = express_type
            
        df.loc[df['order_id']==order_id,'尾端货币'] = bill_express.tail_amount[1]
    return df

# 订单层面审核，防止出现混合渠道投递，卡派订单包含多个不同快递单号，多渠道订单总重量小于1000KG
def analyze_orders(raw_data: pd.DataFrame):
    """
    处理订单数据并实现业务逻辑判断
    返回聚合后的订单数据和分析结果，包裹信息按指定字典格式输出
    
    """
    data = raw_data.copy()

    # 1. 预处理 - 处理空值
    data.fillna({
        '渠道类型': '未知类型',
        '基础估算': 0,
        '偶发估算': 0,
        '包裹总估算': 0,
        '重量': 0,
        '长': 0,
        '宽': 0,
        '高': 0,
        'postcode': '未知'
    }, inplace=True)

    # 2. 按订单聚合数据
    def create_package_details(group):
        """创建包裹详情字典，严格按照要求的格式"""
        details = {}
        for i, (_, row) in enumerate(group.iterrows(), 1):
            details[f"包裹{i}"] = {
                "宽": f"{float(row['宽']):.2f}",
                "长": f"{float(row['长']):.2f}",
                "高": f"{float(row['高']):.2f}",
                "重量": f"{float(row['重量']):.2f}"
            }
        return details

    grouped = data.groupby('order_id')

    aggregated = pd.DataFrame({
        '订单时间': grouped['order_date'].first(),
        '最晚投递时间': grouped['投递时间'].max(),
        '包裹数量': grouped.size(),
        '总重量': grouped['重量'].sum(),
        '订单总估算': grouped['包裹总估算'].sum(),
        '包裹数据': grouped.apply(create_package_details),  # 使用新函数
        '投递渠道列表': grouped['投递渠道'].unique(),
        '渠道类型列表': grouped['渠道类型'].unique(),
        '快递跟踪号': grouped['快递跟踪号'].unique(),
        '最优渠道推荐':grouped['最优渠道'].first(),
        '最优渠道类型':grouped['最优渠道类型'].first(),
        '最优总费用':grouped['最优总费用'].first(),
        '费用差(RMB)':grouped['费用差(RMB)'].first(),
    }).reset_index()

    # 3. 实现业务逻辑判断（保持不变）
    def determine_order_type(row):
        if len(row['渠道类型列表']) > 1:
            return '混合'
        elif len(row['渠道类型列表']) == 1:
            return row['渠道类型列表'][0]
        else:
            return '未知类型'

    def determine_channel_type(row):
        if len(row['投递渠道列表']) > 1:
            return '多渠道'
        else:
            return '单渠道'

    aggregated['订单类型'] = aggregated.apply(determine_order_type, axis=1)
    aggregated['渠道种类'] = aggregated.apply(determine_channel_type, axis=1)

    # 4. 实现业务规则检查（保持不变）
    def apply_business_rules(row):
        actions = []
        status = '正常'
        comments = []

        if row['订单类型'] == '卡派' and len(row['快递跟踪号']) > 1:
            # tracking_nos = [list(p.values())[0] for p in row['包裹数据'].values()]
            # if len(set(tracking_nos)) > 1:
            #     status = '异常'
            status = '异常'
            comments.append('卡派订单包含多个不同快递单号')
        elif row['订单类型'] == '混合':
            status = '异常'
            comments.append('出现混合渠道类型订单，需要核查')

        if row['渠道种类'] == '多渠道':
            if row['总重量'] < 1000:
                comments.append(f'多渠道订单总重量{row["总重量"]:.2f}KG < 1000KG')

        return pd.Series({
            '状态': status,
            '建议操作': '; '.join(actions) if actions else '下一步',
            '备注': ' | '.join(comments) if comments else ''
        })

    rule_results = aggregated.apply(apply_business_rules, axis=1)
    aggregated = pd.concat([aggregated, rule_results], axis=1)
    aggregated['测算日期'] = date.today().strftime("%Y-%m-%d")
    # 5. 整理最终输出列
    final_columns = [
        'order_id','订单时间','最晚投递时间', '订单类型', '渠道种类','快递跟踪号',
        '包裹数量', '总重量',
        '订单总估算',
        '投递渠道列表', 
        '包裹数据' ,'状态', '备注','最优渠道推荐','最优总费用','费用差(RMB)','测算日期'# 使用新列名
    ]

    return aggregated[final_columns]

# 物流费用层面审核
def analyze_logistics(df: pd.DataFrame):
    """
    1.判断实际投递物流渠道和cal_min_fee计算的最优物流渠道是否一致
    2.物流渠道一致的情况下，判断费用是否一样
    """
    # 1. 计算最优渠道和费用
    df= cal_min_fee(df)
    # 判断渠道是否一致
    df['测算日期'] = date.today().strftime("%Y-%m-%d")
    df['最优渠道'] = df.apply(lambda row: row['单票渠道'] if row['最优渠道类型'] == "快递" or row['最优渠道类型'] == "卡派单包裹" else row['多票渠道'], axis=1)
    df['渠道一致'] = df.apply(lambda row: row['最优渠道'] == logistics_name.get(row['投递渠道']), axis=1)
    # 2. 计算费用是否一致
    
    def all_estimate(row):
        if row['最优总费用'] >=9999:
            return "费用有误"
        all_estimate = convert_currency(row['最优总费用'], row['尾端货币'])
        return all_estimate
    
    df['订单总估算']= df.groupby('order_id')['包裹总估算'].transform('sum')
    df['费用一致'] = df.apply(lambda row: False if isinstance(all_estimate(row), str) else abs(all_estimate(row) - row['订单总估算']) < 1,axis=1)

    df['费用差(RMB)'] = df.apply(lambda row: "费用有误" if isinstance(all_estimate(row), str) else round( all_estimate(row)-row['订单总估算'],2),axis=1)
    df['是否改投'] = df.apply(lambda row: "不改投" if row['渠道一致'] == True else 0,axis=1) # 渠道一致只检查费用问题，无需改投，0不确定，需要人工确认
    df['异常情况'] = None
    # 调整输出列
    final_columns = ['order_date','投递时间','fund_status','包裹状态','运输方式','快递跟踪号','目的国','postcode','快递分区','order_id','package','长','宽','高','重量',
                     '基础估算','偶发估算','包裹总估算','订单总估算','本地估算RMB','渠道类型','投递渠道','单票最小费用','单票渠道','多票最小费用','多票渠道','最优总费用',
                     '最优渠道','最优渠道类型','尾端货币','渠道一致','费用一致','费用差(RMB)','测算日期','是否改投','异常情况']
    return df[final_columns]

# 系统渠道下的本地计算费用
def local_fee_cal(df: pd.DataFrame):
    df_grouped= df.groupby('快递跟踪号')
    for order_num, group in df_grouped:
        postcode = group['postcode'].iloc[0]
        if pd.isna(postcode) or str(postcode).lower() == "nan":
            continue

        packages= Package_group() # Metro-SAIR
        company_name = logistics_name.get(group['投递渠道'].iloc[0])
        opCountry = OperateCountry(group['目的国'].iloc[0])

        total_weight=0 # 按体积重分费用
        for index,row in group.iterrows():
            if row['长'] == 0 or row['宽'] == 0 or row['高'] == 0 or row['重量'] == 0:
                continue
            total_weight = row['长']*row['宽']*row['高']/6000
            package = Package(row['package'],row['长'],row['宽'],row['高'],row['重量'])
            packages.add_package(package)
        try:
            bill = Billing(str(index),opCountry,packages,postcode,company_name=company_name,head_type=1,beizhu='1')
            for index,row in group.iterrows():
                propertion = bill.bill_dict()["体积重"]/total_weight
                tail_fee = bill.tail_amount[0]*propertion
                # 转rmb
                tail_fee = convert_currency(tail_fee, bill.tail_amount[1])
                df.loc[df['package']==row['package'],'本地估算RMB'] =round(tail_fee,2) if tail_fee <9999 else "暂无配置"
        except:
            df.loc[df['快递跟踪号'] == order_num, '本地估算RMB']= "暂无配置"
            continue
        print(bill)
    return df


# 合并新旧df并写入
def append_result(new_data, excel_path, only_columns):
    try:
        df_existing = pd.read_excel(excel_path,dtype={'order_id': str})
    except FileNotFoundError:
        # 文件不存在就直接存
        new_data.to_excel(excel_path, index=False)
        return
    # 识别老表里的特殊列
    special_cols = [col for col in ['是否改投', '异常情况','是否处理'] if col in df_existing.columns]
    # 新老合并（先全部concat起来以便后面筛选）
    df_all = pd.concat([df_existing, new_data], ignore_index=True)
    # 找出：重复的（即同时在新旧里都有的 only_columns 值）
    duplicated_keys = set(df_existing[only_columns]) & set(new_data[only_columns])
    # 1️⃣ 对有重复的 key → 保留 旧表的特殊列 + 新表的其他列
    if duplicated_keys:
        duplicated_keys = list(duplicated_keys)
        # 老表保留特殊列
        old_part = df_existing[df_existing[only_columns].isin(duplicated_keys)][[only_columns] + special_cols]
        # 新表保留除特殊列外的所有列
        new_part = new_data[new_data[only_columns].isin(duplicated_keys)]
        new_part_no_special = new_part.drop(columns=special_cols, errors='ignore')
        # 合并
        merged_part = new_part_no_special.merge(old_part, on=only_columns, how='left')
    else:
        merged_part = pd.DataFrame(columns=df_all.columns)  # 空
    # 2️⃣ 对没有重复的 → 直接保留新表的完整行
    unique_new_part = new_data[~new_data[only_columns].isin(duplicated_keys)]
    # 3️⃣ 把 老数据的全部 + 处理好的新数据拼起来
    final_result = pd.concat([df_existing, merged_part, unique_new_part], ignore_index=True)
    # 去重（以 only_columns 为唯一键，保留最后一次出现的）
    final_result = final_result.drop_duplicates(subset=[only_columns], keep='last')
    # 写回
    final_result.to_excel(excel_path, index=False)


def main():
    # 将前一天改投的数据保存到excel
    # 1.先读取logistics_analysis，并筛选是否改投列为1的数据
    # 2.将筛选结果追加到另一个excel
    df_new  = pd.read_excel(r'D:\test\logistics\拦截数据\logistics_analysis.xlsx')
    df_new  = df_new [df_new ['是否改投'] == "是"]
    df_new = df_new[['目的国','运输方式','order_id','package','基础估算','偶发估算','包裹总估算',
                    '渠道类型','最优渠道类型','投递渠道','最优渠道','尾端货币','订单总估算','最优总费用','费用差(RMB)','测算日期','是否改投','异常情况']]
    target_file1 = r'D:\test\logistics\拦截数据\改投记录表.xlsx'
    append_result(df_new,target_file1,'package')
    print("前一天的数据已保存")

    # 获取数据
    raw_data = fetch_order_data()
    print('已获取数据')
    # 本地计算投递渠道的费用
    order_result =local_fee_cal(raw_data)
    # 计算最优渠道和费用
    raw_data = analyze_logistics(raw_data)
    target_file2 = r'D:\test\logistics\拦截数据\logistics_analysis.xlsx'
    append_result(raw_data,target_file2,'package')
    print('已完成物流费用层面审核')
    # 订单层面审核
    order_result = analyze_orders(raw_data)
    target_file3 = r'D:\test\logistics\拦截数据\order_analysis.xlsx'
    append_result(order_result,target_file3,'order_id')
    print('已完成订单层面审核')

    
if __name__ == '__main__':
    main()
    # 取数