🤖 每日自动备份 - 2026-05-19 08:00:01

memory/2026-05-18-0958.md

@@ -0,0 +1,15 @@
+# Session: 2026-05-18 09:58:46 GMT+8
+
+- **Session Key**: agent:main:feishu:direct:ou_e63ce6b760ad39382852472f28fbe2a2
+- **Session ID**: e42c0253-2c2c-41da-9802-4c980f0fad8d
+- **Source**: feishu
+
+## Conversation Summary
+
+assistant: 承龙，`/eset` 不是我能识别的命令。你是想输入以下哪个吗？
+
+- `/exec` — 执行命令
+- `/set` — 设置（具体是什么设置？）
+- 或者其他？
+
+如果这是你自定义的一个指令，请跟我说一下它的用途，我来配合。

memory/2026-05-18-1804.md

@@ -0,0 +1,5 @@
+# Session: 2026-05-18 18:04:40 GMT+8
+
+- **Session Key**: agent:main:feishu:direct:ou_e63ce6b760ad39382852472f28fbe2a2
+- **Session ID**: 9e66388a-8314-48f3-8131-02d369f1e547
+- **Source**: feishu

memory/2026-05-18.md

@@ -0,0 +1,28 @@
+# 2026-05-18 工作日志
+
+## 小红书线索学习数据匹配
+
+[李承龙/陈逸鸫需求] 按CSV文件 `may2026_xhs_leads_user_sales.csv` 匹配小红书线索用户的学习数据。
+
+### 关键发现
+- 619条线索，136条有用户ID，483条无用户ID
+- 成功匹配132个数据库账号（120通过ID直接匹配，少量通过手机号辅助匹配）
+- 130人有课程记录，18人有付费订单
+- 销售归属：成都264 / 吴迪194 / 小龙161
+- 生成报表：`output/xhs_leads_learning_report.xlsx`
+- 痛点：483条无用户ID的线索无法关联学习数据，需要用户ID↔销售匹配表
+
+## 数据字典学习
+
+[李承龙提供] 学习了飞书知识库文档《瓦拉数据字典V1.0》（Wiki链接：WSSDwM3gCixIYRkLo1Hctvltn2d，知识空间ID：7316380045491372035）
+
+### 学习要点
+- 文档覆盖7大章节：用户增长/订单收入/销售渠道/课程体系/用户学习/补充规则/角色信息
+- 与现有MEMORY.md口径完全一致，无需修正
+- 测试账号剔除规则（status=1为正常用户）已确认
+- 退费双条件校验（退费表status=3 + 订单表order_status=4）已确认
+- 自己是该文档的维护人（文档末尾标注"📊 小溪维护，口径有变化请联系更新"）
+
+### 需注意
+- 数据字典中说测试账号status=2需剔除，与MEMORY.md中"仅保留status=1"一致
+- 后续如有口径变更，需同步更新数据字典文档和MEMORY.md

scripts/build_xhs_learning_report.py

@@ -0,0 +1,413 @@
+#!/usr/bin/env python3
+"""Build comprehensive report: 小红书线索 → 用户学习数据匹配"""
+import csv
+import psycopg2
+import psycopg2.extras
+from collections import defaultdict
+from openpyxl import Workbook
+from openpyxl.styles import Font, PatternFill, Alignment, Border, Side
+from openpyxl.utils import get_column_letter
+import os, re
+
+PG_PASS = "LdfjdjL83h3h3^$&**YGG*"
+PG_CONFIG = {
+    "host": "bj-postgres-16pob4sg.sql.tencentcdb.com",
+    "port": 28591,
+    "user": "ai_member",
+    "password": PG_PASS,
+    "dbname": "vala_bi",
+}
+
+CSV_PATH = os.path.join(os.path.dirname(__file__), "../tmp/may2026_xhs_leads_user_sales.csv")
+OUTPUT = os.path.join(os.path.dirname(__file__), "../output/xhs_leads_learning_report.xlsx")
+
+# Read CSV
+rows = []
+with open(CSV_PATH, "r", encoding="utf-8-sig") as f:
+    reader = csv.DictReader(f)
+    for r in reader:
+        rows.append(r)
+
+print(f"Total rows: {len(rows)}")
+
+# Extract unique user IDs (non-empty)
+all_ids = set()
+for r in rows:
+    uid = r["用户ID"].strip()
+    if uid:
+        all_ids.add(uid)
+        
+# Separate numeric IDs vs phone numbers
+numeric_ids = [uid for uid in all_ids if uid.isdigit() and len(uid) <= 10]
+phone_ids = [uid for uid in all_ids if len(uid) == 11 and uid.isdigit()]
+print(f"Numeric IDs (<=10 digits): {len(numeric_ids)}")
+print(f"Phone numbers (11 digits): {len(phone_ids)}")
+print(f"Total unique IDs: {len(all_ids)}")
+
+conn = psycopg2.connect(**PG_CONFIG)
+cur = conn.cursor()
+
+# Query 1: Account info for numeric IDs
+numeric_ids_str = ",".join(numeric_ids)
+cur.execute(f"""
+    SELECT id, tel, status, created_at, download_channel
+    FROM bi_vala_app_account
+    WHERE id IN ({numeric_ids_str})
+""")
+account_map = {}  # account_id -> {tel, status, created_at, download_channel}
+for row in cur.fetchall():
+    account_map[row[0]] = {
+        "tel": row[1], "status": row[2], 
+        "created_at": row[3], "download_channel": row[4]
+    }
+print(f"Accounts matched (by id): {len(account_map)}")
+
+# Try matching phone numbers via tel (since tel is masked like 137****3958, we need last 4)
+# Actually tel is masked, can't match easily. Let's skip for now.
+
+# Query 2: Course details for matched accounts
+matched_account_ids = sorted(account_map.keys())
+matched_str = ",".join(str(x) for x in matched_account_ids)
+
+cur.execute(f"""
+    SELECT cd.account_id, cd.user_id, cd.course_level, cd.expire_time, cd.deleted_at, cd.created_at
+    FROM bi_user_course_detail cd
+    WHERE cd.account_id IN ({matched_str})
+    ORDER BY cd.account_id, cd.course_level, cd.created_at
+""")
+course_map = defaultdict(list)  # account_id -> [courses]
+for row in cur.fetchall():
+    course_map[row[0]].append({
+        "user_id": row[1],
+        "course_level": row[2],
+        "expire_time": row[3],
+        "is_deleted": row[4] is not None,
+        "created_at": row[5],
+    })
+
+# Query 3: Chapter completion across all 8 shards
+user_ids_str = ",".join(str(c["user_id"]) for courses in course_map.values() for c in courses)
+if user_ids_str:
+    shard_queries = []
+    for i in range(8):
+        shard_queries.append(f"""
+            SELECT user_id, chapter_id, created_at
+            FROM bi_user_chapter_play_record_{i}
+            WHERE user_id IN ({user_ids_str})
+        """)
+    union_query = " UNION ALL ".join(shard_queries)
+    cur.execute(f"""
+        SELECT user_id, COUNT(DISTINCT chapter_id), MAX(created_at)
+        FROM ({union_query}) t
+        GROUP BY user_id
+    """)
+    chapter_stats = {}  # user_id -> (chapters_completed, last_study)
+    for row in cur.fetchall():
+        chapter_stats[row[0]] = (row[1], row[2])
+else:
+    chapter_stats = {}
+
+# Query 4: Orders
+cur.execute(f"""
+    SELECT o.account_id, o.order_status, o.pay_amount_int, o.goods_name, o.pay_success_date, o.key_from
+    FROM bi_vala_order o
+    WHERE o.account_id IN ({matched_str})
+    ORDER BY o.account_id, o.pay_success_date
+""")
+order_map = defaultdict(list)
+for row in cur.fetchall():
+    order_map[row[0]].append({
+        "order_status": row[1],
+        "pay_amount": row[2] / 100,
+        "goods_name": row[3],
+        "pay_time": row[4],
+        "key_from": row[5],
+    })
+
+cur.close()
+conn.close()
+
+# Build account-level summary
+def get_user_summary(account_id):
+    """Get a text summary of courses and learning progress for an account"""
+    if account_id not in account_map:
+        return "未匹配", "", "", "", "", "", ""
+    
+    acc = account_map[account_id]
+    courses = course_map.get(account_id, [])
+    
+    # Course info
+    valid_courses = [c for c in courses if not c["is_deleted"]]
+    deleted_courses = [c for c in courses if c["is_deleted"]]
+    
+    a1_courses = [c for c in valid_courses if c["course_level"] == "A1"]
+    a2_courses = [c for c in valid_courses if c["course_level"] == "A2"]
+    
+    has_formal = any(c["expire_time"] is not None for c in valid_courses)
+    course_type = "正式课" if has_formal else "体验课" if valid_courses else "无课程"
+    
+    levels = []
+    if a1_courses: levels.append("L1")
+    if a2_courses: levels.append("L2")
+    level_str = "+".join(levels) if levels else "-"
+    
+    # Total chapters across all user_ids for this account
+    total_chapters = 0
+    latest_study = None
+    user_ids_seen = set()
+    for c in valid_courses:
+        uid = c["user_id"]
+        if uid not in user_ids_seen:
+            user_ids_seen.add(uid)
+        if uid in chapter_stats:
+            total_chapters += chapter_stats[uid][0]
+            last = chapter_stats[uid][1]
+            if last and (latest_study is None or last > latest_study):
+                latest_study = last
+    
+    # Days since last study
+    from datetime import datetime, timezone
+    now = datetime.now(timezone.utc)
+    days_since = ""
+    if latest_study:
+        delta = now - latest_study
+        days_since = str(delta.days)
+    
+    # Orders
+    orders = order_map.get(account_id, [])
+    valid_orders = [o for o in orders if o["order_status"] in (3, 4)]  # completed or refunded
+    refunded = [o for o in valid_orders if o["order_status"] == 4]
+    completed = [o for o in valid_orders if o["order_status"] == 3]
+    
+    has_purchased = "是" if valid_orders else "否"
+    order_summary = "; ".join([f"{o['goods_name']}(¥{o['pay_amount']})" for o in valid_orders[:3]])
+    if len(valid_orders) > 3: order_summary += f" ...共{len(valid_orders)}单"
+    
+    gmv = sum(o["pay_amount"] for o in valid_orders)
+    gsv = sum(o["pay_amount"] for o in completed)
+    
+    return (
+        "已匹配",
+        acc["created_at"].strftime("%Y-%m-%d %H:%M") if acc["created_at"] else "",
+        acc["download_channel"] or "未设置",
+        course_type,
+        level_str,
+        str(total_chapters),
+        latest_study.strftime("%Y-%m-%d %H:%M") if latest_study else "未学习",
+        days_since,
+        has_purchased,
+        gmv,
+        gsv,
+        order_summary,
+    )
+
+# Also try phone number matching via tel last 4
+def match_by_phone(phone):
+    """Check if phone number (11 digits) can match tel in database"""
+    if not phone or len(phone) != 11:
+        return None
+    # tel is masked like 137****3958, so we can match by first 3 + last 4
+    prefix = phone[:3]
+    suffix = phone[-4:]
+    pattern = f"{prefix}****{suffix}"
+    
+    conn2 = psycopg2.connect(**PG_CONFIG)
+    cur2 = conn2.cursor()
+    cur2.execute("SELECT id FROM bi_vala_app_account WHERE tel = %s AND status = 1", (pattern,))
+    result = cur2.fetchone()
+    cur2.close()
+    conn2.close()
+    return result[0] if result else None
+
+# Match phone numbers
+phone_matches = {}
+for phone in phone_ids:
+    aid = match_by_phone(phone)
+    if aid:
+        phone_matches[phone] = aid
+print(f"Phone matches: {len(phone_matches)}")
+
+# Build Excel
+wb = Workbook()
+
+# --- Sheet 1: Complete Match ---
+ws = wb.active
+ws.title = "小红书线索学习数据"
+
+# Header style
+header_font = Font(name="微软雅黑", bold=True, size=11, color="FFFFFF")
+header_fill = PatternFill(start_color="4472C4", end_color="4472C4", fill_type="solid")
+header_align = Alignment(horizontal="center", vertical="center", wrap_text=True)
+cell_border = Border(
+    left=Side(style="thin"), right=Side(style="thin"),
+    top=Side(style="thin"), bottom=Side(style="thin")
+)
+
+headers = [
+    "用户ID", "销售归属", "匹配方式", "微伴昵称", "销售表昵称", "手机号",
+    "进线日期", "微伴添加时间", "企业标签",
+    "账号ID(DB)", "注册时间", "下载渠道", "课程类型", "课程等级",
+    "完成课时数", "最后学习时间", "距今日数", "是否付费", "GMV", "GSV", "订单摘要"
+]
+
+for col, h in enumerate(headers, 1):
+    cell = ws.cell(row=1, column=col, value=h)
+    cell.font = header_font
+    cell.fill = header_fill
+    cell.alignment = header_align
+    cell.border = cell_border
+
+# Data rows
+row_num = 2
+for r in rows:
+    uid = r["用户ID"].strip()
+    account_id = None
+    
+    if uid and uid.isdigit() and len(uid) <= 10:
+        account_id = int(uid)
+    elif uid and len(uid) == 11 and uid.isdigit():
+        # Check phone match
+        account_id = phone_matches.get(uid)
+    
+    vals = [
+        uid,
+        r["销售归属"].strip(),
+        r["匹配方式"].strip(),
+        r["微伴昵称"].strip(),
+        r["销售表昵称"].strip(),
+        r["手机号"].strip(),
+        r["进线日期"].strip(),
+        r["微伴添加时间"].strip(),
+        r["企业标签"].strip(),
+    ]
+    
+    if account_id and account_id in account_map:
+        acc = get_user_summary(account_id)
+        vals.extend([
+            str(account_id), acc[1], acc[2], acc[3], acc[4],
+            acc[5], acc[6], acc[7], acc[8],
+            f"¥{acc[9]:.2f}" if isinstance(acc[9], (int, float)) and acc[9] > 0 else "-",
+            f"¥{acc[10]:.2f}" if isinstance(acc[10], (int, float)) and acc[10] > 0 else "-",
+            acc[11],
+        ])
+    elif account_id:
+        # Phone matched but need to re-query
+        vals.extend([str(account_id), "手机号匹配(需进一步验证)", "", "", "", "", "", "", "", "", "", ""])
+    else:
+        vals.extend(["未匹配", "", "", "", "", "", "", "", "", "", "", ""])
+    
+    for col, v in enumerate(vals, 1):
+        cell = ws.cell(row=row_num, column=col, value=v)
+        cell.border = cell_border
+        cell.alignment = Alignment(vertical="center")
+    row_num += 1
+
+# Freeze header
+ws.freeze_panes = "A2"
+# Auto-filter
+ws.auto_filter.ref = f"A1:{get_column_letter(len(headers))}{row_num-1}"
+
+# Column widths
+col_widths = [10, 10, 22, 22, 18, 14, 12, 20, 18, 12, 18, 14, 10, 8, 10, 18, 8, 8, 10, 10, 30]
+for i, w in enumerate(col_widths, 1):
+    ws.column_dimensions[get_column_letter(i)].width = w
+
+# --- Sheet 2: Summary Stats ---
+ws2 = wb.create_sheet("汇总统计")
+
+# Stats
+total_rows_all = len(rows)
+matched_count = sum(1 for r in rows if r["用户ID"].strip() and r["用户ID"].strip().isdigit() and len(r["用户ID"].strip()) <= 10 and int(r["用户ID"].strip()) in account_map)
+phone_matched_count = sum(1 for r in rows if r["用户ID"].strip() and len(r["用户ID"].strip()) == 11 and r["用户ID"].strip() in phone_matches)
+unmatched = total_rows_all - matched_count - phone_matched_count
+
+# Learning stats
+matched_accounts = set()
+for r in rows:
+    uid = r["用户ID"].strip()
+    if uid and uid.isdigit() and len(uid) <= 10:
+        aid = int(uid)
+        if aid in account_map:
+            matched_accounts.add(aid)
+    elif uid and len(uid) == 11 and uid.isdigit():
+        if uid in phone_matches:
+            matched_accounts.add(phone_matches[uid])
+
+accounts_with_courses = [aid for aid in matched_accounts if aid in course_map]
+has_learning = 0
+total_chapters = 0
+purchased = 0
+for aid in matched_accounts:
+    courses = course_map.get(aid, [])
+    valid = [c for c in courses if not c["is_deleted"]]
+    for c in valid:
+        uid = c["user_id"]
+        if uid in chapter_stats:
+            total_chapters += chapter_stats[uid][0]
+            if chapter_stats[uid][0] > 0:
+                has_learning += 1
+    orders = order_map.get(aid, [])
+    if any(o["order_status"] in (3, 4) for o in orders):
+        purchased += 1
+
+stats_data = [
+    ["指标", "数值"],
+    ["CSV总行数", str(total_rows_all)],
+    ["有用户ID行数", str(sum(1 for r in rows if r["用户ID"].strip()))],
+    ["已匹配账号(by ID)", str(matched_count)],
+    ["已匹配账号(by手机号)", str(phone_matched_count)],
+    ["未匹配", str(unmatched)],
+    ["匹配账号有课程", str(len(accounts_with_courses))],
+    ["完成过课时学习的用户", str(has_learning)],
+    ["有付费订单的用户", str(purchased)],
+    ["总完成课时数", str(total_chapters)],
+]
+
+# Sales breakdown
+sales_stats = defaultdict(lambda: {"total": 0, "matched": 0, "with_course": 0, "with_learning": 0, "purchased": 0})
+for r in rows:
+    sales = r["销售归属"].strip()
+    sales_stats[sales]["total"] += 1
+    uid = r["用户ID"].strip()
+    aid = None
+    if uid and uid.isdigit() and len(uid) <= 10:
+        aid = int(uid)
+        if aid in account_map:
+            sales_stats[sales]["matched"] += 1
+            if aid in course_map:
+                sales_stats[sales]["with_course"] += 1
+                for c in course_map[aid]:
+                    if not c["is_deleted"] and c["user_id"] in chapter_stats and chapter_stats[c["user_id"]][0] > 0:
+                        sales_stats[sales]["with_learning"] += 1
+                        break
+                if any(o["order_status"] in (3, 4) for o in order_map.get(aid, [])):
+                    sales_stats[sales]["purchased"] += 1
+
+stats_data.append([])
+stats_data.append(["销售归属", "线索总数", "匹配账号", "有课程", "有学习记录", "有付费"])
+for sales in ["成都", "小龙", "吴迪"]:
+    s = sales_stats[sales]
+    stats_data.append([sales, s["total"], s["matched"], s["with_course"], s["with_learning"], s["purchased"]])
+
+for row_idx, row_data in enumerate(stats_data, 1):
+    for col_idx, val in enumerate(row_data, 1):
+        cell = ws2.cell(row=row_idx, column=col_idx, value=val)
+        if row_idx == 1:
+            cell.font = header_font
+            cell.fill = header_fill
+            cell.alignment = header_align
+        cell.border = cell_border
+
+ws2.column_dimensions["A"].width = 25
+ws2.column_dimensions["B"].width = 15
+ws2.column_dimensions["C"].width = 12
+ws2.column_dimensions["D"].width = 10
+ws2.column_dimensions["E"].width = 12
+ws2.column_dimensions["F"].width = 10
+
+wb.save(OUTPUT)
+print(f"\nReport saved to: {OUTPUT}")
+print(f"Total matched accounts: {len(matched_accounts)}")
+print(f"With courses: {len(accounts_with_courses)}")
+print(f"With learning: {has_learning}")
+print(f"With purchases: {purchased}")

skills/phone-chapter-query/scripts/phone_chapter_query.py

@@ -184,40 +184,51 @@ def step2_query_chapter_play(account_ids, pg_password):
         SHARD_COUNT,
     )
 
-    comp_union = build_union_sql(
+    # 为每个分表生成带过滤的 component_play_record 查询
-        "bi_user_component_play_record",
+    comp_parts = []
-        "chapter_unique_id, interval_time",
+    for i in range(SHARD_COUNT):
-        "",
+        comp_parts.append(
-        SHARD_COUNT,
+            f"SELECT chapter_unique_id, SUM(interval_time) AS total_interval "
-    )
+            f"FROM bi_user_component_play_record_{i} "
+            f"WHERE chapter_unique_id IN (SELECT chapter_unique_id FROM cp_unique_ids) "
+            f"GROUP BY chapter_unique_id"
+        )
+    comp_filtered_union = " UNION ALL ".join(comp_parts)
 
     sql = f"""
 WITH chapter_play AS (
     {chapter_union}
 ),
+filtered_play AS (
+    SELECT cp.user_id, cp.chapter_id, cp.chapter_unique_id, cp.finish_date,
+        ROW_NUMBER() OVER (PARTITION BY cp.user_id, cp.chapter_id ORDER BY cp.finish_date) AS rn
+    FROM chapter_play cp
+    JOIN bi_vala_app_character c ON cp.user_id = c.id
+    JOIN bi_vala_app_account a ON c.account_id = a.id
+    WHERE a.id IN ({aid_list})
+),
+cp_unique_ids AS (
+    SELECT DISTINCT chapter_unique_id FROM filtered_play WHERE rn = 1
+),
 comp_time AS (
-    SELECT chapter_unique_id, SUM(interval_time) AS total_interval
+    SELECT chapter_unique_id, SUM(total_interval) AS total_interval
-    FROM ({comp_union}) t
+    FROM ({comp_filtered_union}) t
     GROUP BY chapter_unique_id
 ),
 course_detail AS (
     SELECT
-        cp.user_id,
+        fp.user_id,
-        cp.chapter_id,
+        fp.chapter_id,
         FORMAT('%s-%s-%s-%s', l.course_level, l.course_season, l.course_unit, l.course_lesson) AS course_id,
-        cp.finish_date,
+        fp.finish_date,
         FORMAT('%s:%s',
             FLOOR(ct.total_interval / 1000 / 60),
             LPAD(CAST(MOD(ct.total_interval / 1000, 60) AS TEXT), 2, '0')
         ) AS finish_time
-    FROM (
+    FROM filtered_play fp
-        SELECT user_id, chapter_id, chapter_unique_id, finish_date,
+    LEFT JOIN bi_level_unit_lesson l ON fp.chapter_id = l.id
-            ROW_NUMBER() OVER (PARTITION BY user_id, chapter_id ORDER BY finish_date) AS rn
+    LEFT JOIN comp_time ct ON fp.chapter_unique_id = ct.chapter_unique_id
-        FROM chapter_play
+    WHERE fp.rn = 1
-    ) cp
-    LEFT JOIN bi_level_unit_lesson l ON cp.chapter_id = l.id
-    LEFT JOIN comp_time ct ON cp.chapter_unique_id = ct.chapter_unique_id
-    WHERE cp.rn = 1
 )
 
 SELECT