121 lines
3.7 KiB
Bash
Executable File
121 lines
3.7 KiB
Bash
Executable File
#!/bin/bash
|
||
# update.sh - 自动从 Google Sheet 拉取 EG Nagwa 当月数据并生成 data.json
|
||
# 数据源：公开导出的 Google Sheet CSV
|
||
# 每天由服务器自动执行一次（见 schedule.conf）
|
||
set -euo pipefail
|
||
|
||
SHEET_ID="11Npr4IpXT8jtgUnrkrjITJ8s1fHNYTwRRwSGIvcfeIs"
|
||
GID="699511110"
|
||
CSV_URL="https://docs.google.com/spreadsheets/d/${SHEET_ID}/export?format=csv&gid=${GID}"
|
||
|
||
OUT="data.json"
|
||
TMP_CSV="$(mktemp)"
|
||
trap 'rm -f "$TMP_CSV"' EXIT
|
||
|
||
# 当月，格式如 September-2026
|
||
MONTH_TAG="$(date +%B-%Y)"
|
||
|
||
echo "Fetching CSV for month: $MONTH_TAG"
|
||
curl -fsSL "$CSV_URL" -o "$TMP_CSV"
|
||
|
||
# 用 python 解析 CSV 生成 data.json（python3 处理带引号/逗号的字段更稳妥）
|
||
python3 - "$TMP_CSV" "$OUT" "$MONTH_TAG" <<'PY'
|
||
import csv, json, sys, re
|
||
from datetime import datetime
|
||
|
||
csv_path, out_path, month_tag = sys.argv[1], sys.argv[2], sys.argv[3]
|
||
|
||
def num(s):
|
||
if s is None: return None
|
||
s = s.strip().replace(',', '').replace('$', '').replace('%', '')
|
||
if s == '': return None
|
||
try:
|
||
f = float(s)
|
||
return int(f) if f.is_integer() else f
|
||
except ValueError:
|
||
return None
|
||
|
||
rows = list(csv.reader(open(csv_path, newline='', encoding='utf-8')))
|
||
|
||
# 列索引（0-based），依据表头布局：
|
||
# Month=7, DT=8, Geo=9, Install=10, Price=11, MidasSpend=12,
|
||
# PayableSpend=13, Enroll_Session=14, Enroll_Session%=15, Remark=16,
|
||
# meta-key=17, meta-value=18
|
||
COL_MONTH, COL_DT, COL_INSTALL = 7, 8, 10
|
||
COL_PAYABLE, COL_ENROLL, COL_ENROLLPCT = 13, 14, 15
|
||
COL_METAK, COL_METAV = 16, 17
|
||
|
||
daily = []
|
||
meta = {"geo": "EG", "product": "Nagwa"}
|
||
|
||
# 从 month_tag 解析目标年月，用于把 DT（如 2026/9/1 或 20260901）标准化成 YYYY-MM-DD
|
||
target_month = datetime.strptime(month_tag, "%B-%Y")
|
||
ty, tm = target_month.year, target_month.month
|
||
|
||
def parse_dt(dt):
|
||
dt = dt.strip()
|
||
if not dt: return None
|
||
# 形如 2026/9/1
|
||
m = re.match(r'^(\d{4})/(\d{1,2})/(\d{1,2})$', dt)
|
||
if m:
|
||
y, mo, d = map(int, m.groups())
|
||
return f"{y:04d}-{mo:02d}-{d:02d}"
|
||
# 形如 20260901
|
||
m = re.match(r'^(\d{4})(\d{2})(\d{2})$', dt)
|
||
if m:
|
||
y, mo, d = map(int, m.groups())
|
||
return f"{y:04d}-{mo:02d}-{d:02d}"
|
||
return None
|
||
|
||
price_seen = None
|
||
for r in rows:
|
||
if len(r) <= COL_ENROLLPCT:
|
||
continue
|
||
if r[COL_MONTH].strip() != month_tag:
|
||
continue
|
||
date = parse_dt(r[COL_DT])
|
||
if not date:
|
||
continue
|
||
install = num(r[COL_INSTALL]) or 0
|
||
spend = num(r[COL_PAYABLE]) or 0.0
|
||
enroll = num(r[COL_ENROLL]) or 0
|
||
pct = num(r[COL_ENROLLPCT]) or 0.0
|
||
daily.append({
|
||
"date": date,
|
||
"install": int(install),
|
||
"payable_spend": round(float(spend), 2),
|
||
"enroll_session": int(enroll),
|
||
"enroll_session_pct": round(float(pct), 2),
|
||
})
|
||
# 记录价格（取当月出现的最新价）
|
||
p = num(r[11])
|
||
if p is not None:
|
||
price_seen = p
|
||
# 抓取 meta（Remark 区块的 key/value）
|
||
k = r[COL_METAK].strip() if len(r) > COL_METAK else ""
|
||
v = r[COL_METAV].strip() if len(r) > COL_METAV else ""
|
||
if k and v:
|
||
kl = k.lower()
|
||
if "budget" in kl:
|
||
meta["sep_budget"] = num(v)
|
||
elif kl == "spend":
|
||
meta["spend"] = num(v)
|
||
elif kl == "left":
|
||
meta["left"] = num(v)
|
||
elif kl == "%":
|
||
meta["spend_pct"] = v if v.endswith('%') else (v + '%')
|
||
elif "daily install" in kl:
|
||
meta["rec_daily_install"] = num(v)
|
||
|
||
if price_seen is not None:
|
||
meta["price"] = price_seen
|
||
|
||
data = {"meta": meta, "daily": daily}
|
||
with open(out_path, "w", encoding="utf-8") as f:
|
||
json.dump(data, f, ensure_ascii=False, indent=2)
|
||
|
||
print(f"Wrote {len(daily)} days to {out_path}")
|
||
PY
|
||
|
||
echo "Done. Last run: $(date)"
|