ai_member_xiaoxi/scripts/query_chapter_play.sh

#!/bin/bash
# 课时完成记录查询脚本
# 查询指定课时(chapter_id)的首次完成记录，包含角色信息、最近登录、课程耗时
# 数据库：线上 PostgreSQL (vala_bi)
# 输出：xlsx 格式
# 用法：bash scripts/query_chapter_play.sh

set -euo pipefail

SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
source "${SCRIPT_DIR}/../secrets.env"

DB_HOST="bj-postgres-16pob4sg.sql.tencentcdb.com"
DB_PORT="28591"
DB_USER="ai_member"
DB_NAME="vala_bi"

OUTPUT_DIR="${SCRIPT_DIR}/../output"
mkdir -p "$OUTPUT_DIR"

TIMESTAMP=$(date +%Y%m%d_%H%M%S)
CSV_TMP="${OUTPUT_DIR}/.chapter_play_${TIMESTAMP}.csv"
OUTPUT_FILE="${OUTPUT_DIR}/chapter_play_${TIMESTAMP}.xlsx"

SQL=$(cat <<'EOF'
WITH chapter_play AS (
    SELECT user_id, chapter_id, chapter_unique_id, date(updated_at) AS finish_date
    FROM bi_user_chapter_play_record_0
    WHERE chapter_id IN (55,56,57,58,59,343,344,345,346,348) AND play_status = 1
    UNION ALL
    SELECT user_id, chapter_id, chapter_unique_id, date(updated_at)
    FROM bi_user_chapter_play_record_1
    WHERE chapter_id IN (55,56,57,58,59,343,344,345,346,348) AND play_status = 1
    UNION ALL
    SELECT user_id, chapter_id, chapter_unique_id, date(updated_at)
    FROM bi_user_chapter_play_record_2
    WHERE chapter_id IN (55,56,57,58,59,343,344,345,346,348) AND play_status = 1
    UNION ALL
    SELECT user_id, chapter_id, chapter_unique_id, date(updated_at)
    FROM bi_user_chapter_play_record_3
    WHERE chapter_id IN (55,56,57,58,59,343,344,345,346,348) AND play_status = 1
    UNION ALL
    SELECT user_id, chapter_id, chapter_unique_id, date(updated_at)
    FROM bi_user_chapter_play_record_4
    WHERE chapter_id IN (55,56,57,58,59,343,344,345,346,348) AND play_status = 1
    UNION ALL
    SELECT user_id, chapter_id, chapter_unique_id, date(updated_at)
    FROM bi_user_chapter_play_record_5
    WHERE chapter_id IN (55,56,57,58,59,343,344,345,346,348) AND play_status = 1
    UNION ALL
    SELECT user_id, chapter_id, chapter_unique_id, date(updated_at)
    FROM bi_user_chapter_play_record_6
    WHERE chapter_id IN (55,56,57,58,59,343,344,345,346,348) AND play_status = 1
    UNION ALL
    SELECT user_id, chapter_id, chapter_unique_id, date(updated_at)
    FROM bi_user_chapter_play_record_7
    WHERE chapter_id IN (55,56,57,58,59,343,344,345,346,348) AND play_status = 1
),
comp_time AS (
    SELECT chapter_unique_id, SUM(interval_time) AS total_interval
    FROM (
        SELECT chapter_unique_id, interval_time FROM bi_user_component_play_record_0
        UNION ALL
        SELECT chapter_unique_id, interval_time FROM bi_user_component_play_record_1
        UNION ALL
        SELECT chapter_unique_id, interval_time FROM bi_user_component_play_record_2
        UNION ALL
        SELECT chapter_unique_id, interval_time FROM bi_user_component_play_record_3
        UNION ALL
        SELECT chapter_unique_id, interval_time FROM bi_user_component_play_record_4
        UNION ALL
        SELECT chapter_unique_id, interval_time FROM bi_user_component_play_record_5
        UNION ALL
        SELECT chapter_unique_id, interval_time FROM bi_user_component_play_record_6
        UNION ALL
        SELECT chapter_unique_id, interval_time FROM bi_user_component_play_record_7
    ) t
    GROUP BY chapter_unique_id
),
course_detail AS (
    SELECT
        cp.user_id,
        cp.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,
        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 (
        SELECT user_id, chapter_id, chapter_unique_id, finish_date,
            ROW_NUMBER() OVER (PARTITION BY user_id, chapter_id ORDER BY finish_date) AS rn
        FROM chapter_play
    ) 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
    a.account_id              AS "账号ID",
    a.id                      AS "角色ID",
    DATE(a.created_at)        AS "注册日期",
    DATE(TO_TIMESTAMP(b.latest_active_time)) AS "最近登录时间",
    c.course_id               AS "课程名称",
    c.finish_date             AS "课程完成时间",
    c.finish_time             AS "课程耗时"
FROM bi_vala_app_character a
LEFT JOIN user_detail_info b ON a.id = b.user_id
LEFT JOIN course_detail c ON a.id = c.user_id
WHERE a.status = 1
ORDER BY a.id, c.course_id;
EOF
)

echo "正在查询..."
PGPASSWORD="$PG_ONLINE_PASSWORD" psql \
    -h "$DB_HOST" \
    -p "$DB_PORT" \
    -U "$DB_USER" \
    -d "$DB_NAME" \
    --csv \
    -c "$SQL" \
    > "$CSV_TMP"

ROW_COUNT=$(tail -n +2 "$CSV_TMP" | wc -l)
echo "查询完成，共 ${ROW_COUNT} 条记录，正在转换为 xlsx..."

python3 - "$CSV_TMP" "$OUTPUT_FILE" <<'PYEOF'
import csv, sys
from openpyxl import Workbook
from openpyxl.styles import Font, Alignment, PatternFill

csv_path, xlsx_path = sys.argv[1], sys.argv[2]

wb = Workbook()
ws = wb.active
ws.title = "课时完成记录"

# 读取 CSV
with open(csv_path, 'r', encoding='utf-8') as f:
    reader = csv.reader(f)
    for row in reader:
        ws.append(row)

# 表头样式
header_font = Font(bold=True, color="FFFFFF", size=11)
header_fill = PatternFill(start_color="4472C4", end_color="4472C4", fill_type="solid")
header_align = Alignment(horizontal="center", vertical="center")

for cell in ws[1]:
    cell.font = header_font
    cell.fill = header_fill
    cell.alignment = header_align

# 自动列宽
for col in ws.columns:
    max_len = 0
    col_letter = col[0].column_letter
    for cell in col:
        val = str(cell.value) if cell.value else ""
        # 中文字符算 2 宽度
        width = sum(2 if ord(c) > 127 else 1 for c in val)
        if width > max_len:
            max_len = width
    ws.column_dimensions[col_letter].width = min(max_len + 3, 30)

# 冻结首行
ws.freeze_panes = "A2"

wb.save(xlsx_path)
print(f"xlsx 已保存: {xlsx_path}")
PYEOF

# 清理临时 CSV
rm -f "$CSV_TMP"

echo "完成！结果已保存至: ${OUTPUT_FILE}"