#!/usr/bin/env python3 """ 老狼履约明细 — 行课分析（用订单号匹配） 数据源：李丹发的5.23日老狼履约明细.xlsx（含订单号列） 规则： - 用订单号直接查 account_id，避免手机号脱敏碰撞 - 购买 L1+L2 联报 → 看 L1 行课 - 购买 L2 课包 → 看 L2 行课 - 含角色年龄、完课时长、加微状态 """ import os, sys import psycopg2 import pandas as pd from collections import defaultdict from datetime import date DB_HOST = "bj-postgres-16pob4sg.sql.tencentcdb.com" DB_PORT = 28591 DB_USER = "ai_member" DB_NAME = "vala_bi" def get_password(): pw = os.environ.get("PG_ONLINE_PASSWORD", "") if pw: return pw secrets_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "..", "secrets.env") if os.path.exists(secrets_path): with open(secrets_path) as f: for line in f: if line.startswith("PG_ONLINE_PASSWORD="): return line.strip().split("=", 1)[1].strip("'\"") raise RuntimeError("PG_ONLINE_PASSWORD not found") def get_conn(db=DB_NAME): return psycopg2.connect(host=DB_HOST, port=DB_PORT, user=DB_USER, password=get_password(), dbname=db, connect_timeout=120) def lesson_bucket(n): if n == 0: return "0课时" elif n <= 3: return "1-3课时" elif n <= 7: return "4-7课时" elif n <= 15: return "8-15课时" elif n <= 30: return "16-30课时" elif n <= 60: return "31-60课时" else: return "60课时以上" def time_bucket(m): if m == 0: return "0分钟" elif m <= 10: return "1-10分钟" elif m <= 30: return "11-30分钟" elif m <= 60: return "31-60分钟" elif m <= 120: return "61-120分钟" elif m <= 300: return "121-300分钟" else: return "300分钟以上" BUCKET_ORDER = ["0课时","1-3课时","4-7课时","8-15课时","16-30课时","31-60课时","60课时以上"] TIME_BUCKET_ORDER = ["0分钟","1-10分钟","11-30分钟","31-60分钟","61-120分钟","121-300分钟","300分钟以上"] L1_GOODS, L2_GOODS, L1L2_GOODS = {57,60,63}, {31,32,33,54}, {61} def classify_orders(goods_ids): gs = set(goods_ids) r = [] if gs & L1L2_GOODS: r.append(("L1+L2联报→看L1","L1")) if (gs & L2_GOODS) and not (gs & L1L2_GOODS): r.append(("仅L2→看L2","L2")) if (gs & L1L2_GOODS) and (gs & L2_GOODS): r.append(("联报+仅L2→看L2","L2")) return r def main(): input_file = sys.argv[1] if len(sys.argv) > 1 else "/root/.openclaw/media/inbound/5.23æ_è_ç_¼å_çº_æ_ç---c2361b75-b1c0-48c4-8302-d68ebb5e05d2.xlsx" df_raw = pd.read_excel(input_file, dtype=str) print(f"原始文件: {len(df_raw)} 行") # ── 用订单号直接查 account_id ── trade_nos = df_raw["订单号"].dropna().unique().tolist() print(f"订单号: {len(trade_nos)} 个") conn = get_conn() cur = conn.cursor() ph = ",".join(["%s"]*len(trade_nos)) cur.execute(f""" SELECT o.account_id, o.trade_no, o.goods_id, o.goods_name, o.pay_success_date, o.pay_amount_int, o.order_status, o.key_from FROM bi_vala_order o INNER JOIN bi_vala_app_account a ON o.account_id = a.id AND a.status = 1 WHERE o.trade_no IN ({ph}) AND o.deleted_at IS NULL """, trade_nos) orders = cur.fetchall() print(f" 订单: {len(orders)} 条") # 退款 refund_set = set() for i in range(0, len(trade_nos), 500): batch = trade_nos[i:i+500] p2 = ",".join(["%s"]*len(batch)) cur.execute(f"SELECT DISTINCT trade_no FROM bi_refund_order WHERE trade_no IN ({p2}) AND status=3 AND deleted_at IS NULL", batch) for (tn,) in cur.fetchall(): refund_set.add(tn) # 用户列表 user_ids = sorted(set(o[0] for o in orders)) print(f" 用户: {len(user_ids)} 人") # ── 课程结构 ── cur.execute("SELECT id, course_level FROM bi_level_unit_lesson WHERE course_level IN ('L1','L2')") l1_chs, l2_chs = set(), set() for cid, lv in cur.fetchall(): if lv == "L1": l1_chs.add(cid) else: l2_chs.add(cid) # ── 角色（含生日）── ph_u = ",".join(["%s"]*len(user_ids)) cur.execute(f""" SELECT id, account_id, nickname, birthday FROM bi_vala_app_character WHERE account_id IN ({ph_u}) AND nickname IS NOT NULL AND nickname!='' AND deleted_at IS NULL """, user_ids) chars = cur.fetchall() char_ids = [c[0] for c in chars] account_chars = defaultdict(list) char_info = {} today = date.today() for cid, aid, nick, bday in chars: account_chars[aid].append(cid) age = None if bday: try: bd = pd.Timestamp(str(bday)[:10]).date() age = today.year - bd.year - ((today.month, today.day) < (bd.month, bd.day)) except: pass char_info[cid] = (nick, str(bday)[:10] if bday else '', age) print(f" 角色: {len(chars)}") # ── 课时完成 ── char_chapter_cuid = defaultdict(dict) char_lessons_l1 = defaultdict(int) char_lessons_l2 = defaultdict(int) all_cuids = set() for tbl_idx in range(8): table = f"bi_user_chapter_play_record_{tbl_idx}" for i in range(0, len(char_ids), 2000): batch = char_ids[i:i+2000] p2 = ",".join(["%s"]*len(batch)) try: cur.execute(f""" SELECT user_id, chapter_id, chapter_unique_id FROM {table} WHERE user_id IN ({p2}) AND play_status=1 AND deleted_at IS NULL """, batch) for uid, ch_id, cuid in cur.fetchall(): if ch_id not in char_chapter_cuid[uid]: char_chapter_cuid[uid][ch_id] = cuid if ch_id in l1_chs: char_lessons_l1[uid] += 1 elif ch_id in l2_chs: char_lessons_l2[uid] += 1 all_cuids.add(cuid) except Exception as e: print(f" warn {table}: {e}") # ── 完课时长 ── cuid_duration = defaultdict(int) cuid_list = list(all_cuids) for tbl_idx in range(8): table = f"bi_user_component_play_record_{tbl_idx}" for i in range(0, len(cuid_list), 2000): batch = cuid_list[i:i+2000] p2 = ",".join(["%s"]*len(batch)) try: cur.execute(f""" SELECT chapter_unique_id, SUM(COALESCE(interval_time,0)) FROM {table} WHERE chapter_unique_id IN ({p2}) AND deleted_at IS NULL GROUP BY chapter_unique_id """, batch) for cuid, ms in cur.fetchall(): if ms: cuid_duration[cuid] += ms except: pass # ── 账户信息 ── cur.execute(f"SELECT id, created_at FROM bi_vala_app_account WHERE id IN ({ph_u}) AND status=1", user_ids) account_info = {aid: reg for aid, reg in cur.fetchall()} conn.close() # ── 加微（vala_class.student_info）── wechat_bound = set() try: conn_class = get_conn("vala_class") cur2 = conn_class.cursor() cur2.execute(f"SELECT DISTINCT vala_account_id FROM student_info WHERE vala_account_id IN ({ph_u})", user_ids) wechat_bound = {r[0] for r in cur2.fetchall()} cur2.close() conn_class.close() print(f" 加微: {len(wechat_bound)}/{len(user_ids)} 人") except Exception as e: print(f" 加微跳过: {e}") # ── 构建数据 ── user_orders_map = defaultdict(list) for o in orders: aid, tn, gid, gn, pd_, amt, os_, kf = o user_orders_map[aid].append({ "goods_id": gid, "goods_name": gn, "trade_no": tn, "pay_date": pd_, "amount": amt/100.0, "order_status": os_, "key_from": kf, "is_refunded": tn in refund_set, }) char_dur_l1 = defaultdict(int) char_dur_l2 = defaultdict(int) for cid, cmap in char_chapter_cuid.items(): for ch_id, cuid in cmap.items(): dur = cuid_duration.get(cuid, 0) if ch_id in l1_chs: char_dur_l1[cid] += dur elif ch_id in l2_chs: char_dur_l2[cid] += dur rows = [] for aid in user_ids: my_orders = user_orders_map.get(aid, []) my_chars = account_chars.get(aid, []) reg_time = account_info.get(aid) total_l1 = sum(char_lessons_l1.get(c,0) for c in my_chars) total_l2 = sum(char_lessons_l2.get(c,0) for c in my_chars) dur_l1 = sum(char_dur_l1.get(c,0) for c in my_chars) dur_l2 = sum(char_dur_l2.get(c,0) for c in my_chars) total_orders = len(my_orders) refunded_orders = sum(1 for o in my_orders if o["is_refunded"]) total_gmv = sum(o["amount"] for o in my_orders) total_refund = sum(o["amount"] for o in my_orders if o["is_refunded"]) all_refunded = (refunded_orders == total_orders and total_orders > 0) has_refund = refunded_orders > 0 goods_ids = [o["goods_id"] for o in my_orders] cats = classify_orders(goods_ids) char_details = [] for cid in my_chars: nick, bday_str, age = char_info.get(cid, ('?','',None)) age_str = f'{age}岁' if age is not None else '?' char_details.append(f'{nick}({age_str})') for cat_label, watch_level in cats: lessons = total_l1 if watch_level == "L1" else total_l2 dur_ms = dur_l1 if watch_level == "L1" else dur_l2 dur_min = round(dur_ms/60000.0, 1) if dur_ms else 0.0 avg_min = round(dur_min/lessons, 1) if lessons > 0 else 0.0 rows.append({ "用户ID": aid, "注册时间": reg_time, "购买分类": cat_label, "行课等级": watch_level, "角色数": len(my_chars), "角色信息(名称+年龄)": "; ".join(char_details), "完成课时数": lessons, "总完课时长(分钟)": dur_min, "平均每课时长(分钟)": avg_min, "订单数": total_orders, "退款订单数": refunded_orders, "GMV": round(total_gmv, 2), "GSV": round(total_gmv - total_refund, 2), "是否退款": "是" if has_refund else "否", "是否全部退款": "是" if all_refunded else "否", "是否加微": "是" if aid in wechat_bound else "否", "购买课包": ";".join(o["goods_name"] for o in my_orders), }) df = pd.DataFrame(rows) df["课时桶"] = df["完成课时数"].apply(lesson_bucket) df["时长桶"] = df["总完课时长(分钟)"].apply(time_bucket) # ── 输出 ── print(f"\n{'='*70}") print(f"老狼履约明细 — 行课分析（{len(user_ids)} 用户, {len(orders)} 订单）") print(f"{'='*70}") for cat_label in ["L1+L2联报→看L1", "仅L2→看L2"]: df_cat = df[df["购买分类"] == cat_label] if len(df_cat) == 0: continue print(f"\n{'='*60}") print(f"【{cat_label}】 {len(df_cat)} 个用户-分类记录") print(f"{'='*60}") print("\n ┌─ 按完成课时数分桶 ─────────────────────") bs = df_cat.groupby("课时桶").agg( 用户数=("用户ID","count"), 退款用户数=("是否退款", lambda x: (x=="是").sum()), 平均课时=("完成课时数","mean"), 平均完课时长=("总完课时长(分钟)","mean"), 中位完课时长=("总完课时长(分钟)","median"), ).reindex(BUCKET_ORDER).fillna(0) bs["退款率%"] = (bs["退款用户数"]/bs["用户数"]*100).round(1) bs["平均课时"] = bs["平均课时"].round(1) bs["平均完课时长"] = bs["平均完课时长"].round(1) bs["中位完课时长"] = bs["中位完课时长"].round(1) bs = bs[bs["用户数"]>0] print(bs.to_string()) print("\n ┌─ 按完课时长分桶 ─────────────────────") ts = df_cat.groupby("时长桶").agg( 用户数=("用户ID","count"), 退款用户数=("是否退款", lambda x: (x=="是").sum()), 平均课时=("完成课时数","mean"), 平均完课时长=("总完课时长(分钟)","mean"), ).reindex(TIME_BUCKET_ORDER).fillna(0) ts["退款率%"] = (ts["退款用户数"]/ts["用户数"]*100).round(1) ts["平均课时"] = ts["平均课时"].round(1) ts["平均完课时长"] = ts["平均完课时长"].round(1) ts = ts[ts["用户数"]>0] print(ts.to_string()) refund_users = df_cat[df_cat["是否退款"]=="是"] no_refund = df_cat[df_cat["是否退款"]=="否"] print(f"\n ┌─ 汇总 ──────────────────────────────") print(f" 退款用户: {len(refund_users)}, 平均课时: {refund_users['完成课时数'].mean():.1f}, 平均总时长: {refund_users['总完课时长(分钟)'].mean():.1f}min") print(f" 未退款用户: {len(no_refund)}, 平均课时: {no_refund['完成课时数'].mean():.1f}, 平均总时长: {no_refund['总完课时长(分钟)'].mean():.1f}min") print(f" 退费率: {len(refund_users)/len(df_cat)*100:.1f}%") print(f"\n ┌─ 明细 ─────────────────────────────") for _, r in df_cat.sort_values(["总完课时长(分钟)"]).iterrows(): print(f" UID={r['用户ID']} | {r['角色信息(名称+年龄)'][:30]} | {r['完成课时数']}课时 | {r['总完课时长(分钟)']}min | 加微={r['是否加微']} | {'退款' if r['是否退款']=='是' else '未退款'}") # ── Excel ── output_dir = os.path.join(os.path.dirname(os.path.abspath(__file__)), "..", "output") os.makedirs(output_dir, exist_ok=True) output_path = os.path.join(output_dir, "老狼履约_行课分析.xlsx") with pd.ExcelWriter(output_path, engine="openpyxl") as writer: df_out = df.drop(columns=["课时桶","时长桶"], errors="ignore").copy() for col in df_out.columns: if pd.api.types.is_datetime64_any_dtype(df_out[col]): df_out[col] = df_out[col].dt.tz_localize(None) df_out.to_excel(writer, sheet_name="明细", index=False) for cat_label in ["L1+L2联报→看L1", "仅L2→看L2"]: dc = df[df["购买分类"]==cat_label] if len(dc)==0: continue bs = dc.groupby("课时桶").agg(用户数=("用户ID","count"),退款用户数=("是否退款",lambda x:(x=="是").sum()),平均课时=("完成课时数","mean"),平均完课时长=("总完课时长(分钟)","mean"),中位完课时长=("总完课时长(分钟)","median")).reindex(BUCKET_ORDER).fillna(0) bs["退款率%"]=(bs["退款用户数"]/bs["用户数"]*100).round(1); bs["平均课时"]=bs["平均课时"].round(1); bs["平均完课时长"]=bs["平均完课时长"].round(1); bs["中位完课时长"]=bs["中位完课时长"].round(1) bs=bs[bs["用户数"]>0]; bs.to_excel(writer, sheet_name=f"{cat_label[:20]}_按课时") ts = dc.groupby("时长桶").agg(用户数=("用户ID","count"),退款用户数=("是否退款",lambda x:(x=="是").sum()),平均课时=("完成课时数","mean"),平均完课时长=("总完课时长(分钟)","mean")).reindex(TIME_BUCKET_ORDER).fillna(0) ts["退款率%"]=(ts["退款用户数"]/ts["用户数"]*100).round(1); ts["平均课时"]=ts["平均课时"].round(1); ts["平均完课时长"]=ts["平均完课时长"].round(1) ts=ts[ts["用户数"]>0]; ts.to_excel(writer, sheet_name=f"{cat_label[:20]}_按时长") print(f"\n✅ {output_path}") if __name__ == "__main__": main()