File size: 10,810 Bytes
3291500
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
90719f1
 
3291500
 
 
 
 
 
 
 
 
 
 
 
 
 
 
90719f1
3291500
 
 
 
 
 
 
 
 
 
 
 
90719f1
3291500
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
90719f1
 
3291500
 
 
 
 
 
 
90719f1
 
3291500
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ec533d1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3291500
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
"""
预算追踪:从 Google Sheets 读取 Budget 工作表,
结合实际花费数据计算 pacing 和 KPI 对比。

Budget 工作表表头(宽格式,每行 = Month × Country):
  Month, Country, SPF_Revenue_Target, DM_Revenue_Target,
  Conversion_Budget, Consideration_Budget, Awareness_Budget,
  DM_Total_Budget, Target_ROAS,
  Conversion_%, Consideration_%, Awareness_%
"""
import calendar
import logging
from datetime import date, datetime
from pathlib import Path
from typing import Any

import pandas as pd

from src.loader import load_config, load_daily_data, _get_spreadsheet

log = logging.getLogger("ad-agent")

FUNNEL_COLS = {
    "Conversion": "Conversion_Budget",
    "Consideration": "Consideration_Budget",
    "Awareness": "Awareness_Budget",
}

FUNNEL_PCT_COLS = {
    "Conversion": "Conversion_%",
    "Consideration": "Consideration_%",
    "Awareness": "Awareness_%",
}

NUMERIC_COLS = [
    "SPF_Revenue_Target", "DM_Revenue_Target",
    "Conversion_Budget", "Consideration_Budget", "Awareness_Budget",
    "DM_Total_Budget", "Target_ROAS",
]

PCT_COLS = ["Conversion_%", "Consideration_%", "Awareness_%"]


def _read_sheet(config: dict, name: str) -> pd.DataFrame | None:
    try:
        sh = _get_spreadsheet(config)
        ws = sh.worksheet(name)
        rows = ws.get_all_values()
        if len(rows) < 2:
            return None
        return pd.DataFrame(rows[1:], columns=[c.strip() for c in rows[0]])
    except Exception as e:
        log.warning(f"Sheet [{name}] 读取失败: {e}")
        return None


def load_budget_config(config: dict) -> pd.DataFrame | None:
    df = _read_sheet(config, "Budget")
    if df is None:
        return None
    for col in NUMERIC_COLS:
        if col in df.columns:
            cleaned = df[col].astype(str).str.replace(r"[$,\s]", "", regex=True)
            df[col] = pd.to_numeric(cleaned, errors="coerce").fillna(0)
    for col in PCT_COLS:
        if col in df.columns:
            cleaned = df[col].astype(str).str.replace(r"[%\s]", "", regex=True)
            df[col] = pd.to_numeric(cleaned, errors="coerce").fillna(0)
    return df


def _pacing_status(pacing: float | None) -> str:
    if pacing is None:
        return "no_data"
    if pacing > 1.15:
        return "overspending"
    if pacing < 0.85:
        return "underspending"
    return "on_track"


def _safe_cost(df: pd.DataFrame) -> float:
    if df.empty or "Cost" not in df.columns:
        return 0.0
    return float(df["Cost"].sum())


def _safe_revenue(df: pd.DataFrame) -> float:
    if df.empty or "Ads.Sales" not in df.columns:
        return 0.0
    return float(df["Ads.Sales"].sum())


def _node(
    budget: float,
    spent: float,
    pct_elapsed: float,
    days_elapsed: int,
    days_in_month: int,
) -> dict:
    pct = spent / budget if budget > 0 else 0
    pacing = pct / pct_elapsed if pct_elapsed > 0 else 0
    projected = spent / days_elapsed * days_in_month if days_elapsed > 0 else 0
    return {
        "budget": round(budget, 2),
        "spent": round(spent, 2),
        "pct_used": round(pct * 100, 1),
        "pacing": round(pacing, 2),
        "pacing_status": _pacing_status(pacing if budget > 0 else None),
        "projected": round(projected, 2),
    }


def _compute_roas(actual_df: pd.DataFrame) -> float | None:
    if actual_df.empty:
        return None
    cost = actual_df["Cost"].sum() if "Cost" in actual_df.columns else 0
    sales = actual_df["Ads.Sales"].sum() if "Ads.Sales" in actual_df.columns else 0
    if cost > 0:
        return round(sales / cost, 2)
    return None


def compute_budget_tracking(
    config_path: Path | None = None,
    target_month: str | None = None,
) -> dict[str, Any]:
    config = load_config(config_path)
    bcfg = config.get("budget", {})
    if not bcfg.get("enabled"):
        return {"error": "budget_disabled"}

    budget_df = load_budget_config(config)
    if budget_df is None or budget_df.empty:
        return {"error": "no_budget_data"}

    today = date.today()
    if target_month:
        try:
            ym = datetime.strptime(target_month, "%Y-%m")
            ty, tm = ym.year, ym.month
        except ValueError:
            return {"error": "bad_month"}
    else:
        ty, tm = today.year, today.month

    ms = f"{ty}-{tm:02d}"
    _, dim = calendar.monthrange(ty, tm)

    if ty == today.year and tm == today.month:
        de = max(today.day - 1, 1)
    elif date(ty, tm, 1) > today:
        de = 1
    else:
        de = dim

    pe = de / dim

    mb = budget_df[budget_df["Month"].astype(str).str.strip() == ms]
    avail = sorted(budget_df["Month"].astype(str).str.strip().unique().tolist())

    if mb.empty:
        return {"error": "no_month_data", "available_months": avail, "month": ms}

    try:
        adf = load_daily_data(config_path)
    except Exception as e:
        return {"error": f"data_load_fail: {e}"}

    m_start = pd.Timestamp(ty, tm, 1)
    m_end = pd.Timestamp(ty, tm, dim)
    am = adf[(adf["date"] >= m_start) & (adf["date"] <= m_end)].copy()

    # ── Total level ──
    tb = float(mb["DM_Total_Budget"].sum())
    ts = _safe_cost(am)
    total_rev = _safe_revenue(am)
    am_conv = am[am["campaign_type"].astype(str).str.strip() == "Conversion"] if "campaign_type" in am.columns else am
    total_roas_actual = _compute_roas(am_conv)

    total_spf = float(mb["SPF_Revenue_Target"].sum())
    total_dm_rev = float(mb["DM_Revenue_Target"].sum())

    weights = mb[["Target_ROAS", "DM_Total_Budget"]].copy()
    weights = weights[weights["Target_ROAS"] > 0]
    if not weights.empty and weights["DM_Total_Budget"].sum() > 0:
        total_target_roas = round(
            (weights["Target_ROAS"] * weights["DM_Total_Budget"]).sum()
            / weights["DM_Total_Budget"].sum(), 2
        )
    else:
        total_target_roas = None

    nd = _node(tb, ts, pe, de, dim)
    total_conv_rev = _safe_revenue(am_conv)
    nd.update({
        "month": ms,
        "remaining": round(tb - ts, 2),
        "days_elapsed": de,
        "days_total": dim,
        "days_remaining": dim - de,
        "pct_elapsed": round(pe * 100, 1),
        "available_months": avail,
        "revenue_target": {
            "spf": round(total_spf, 2),
            "dm": round(total_dm_rev, 2),
        },
        "actual_revenue": round(total_conv_rev, 2),
        "target_roas": total_target_roas,
        "kpi": {},
    })

    if total_roas_actual is not None:
        kpi_roas: dict[str, Any] = {"actual": total_roas_actual}
        if total_target_roas and total_target_roas > 0:
            kpi_roas["target"] = total_target_roas
            kpi_roas["diff_pct"] = round(
                (total_roas_actual - total_target_roas) / total_target_roas * 100, 1
            )
        nd["kpi"]["ROAS"] = kpi_roas

    # ── Country → Funnel → Platform ──
    hier: dict[str, Any] = {}
    for country in sorted(mb["Country"].astype(str).str.strip().unique()):
        crow = mb[mb["Country"].astype(str).str.strip() == country]
        ca = (
            am[am["country"].astype(str).str.strip() == country]
            if "country" in am.columns
            else pd.DataFrame()
        )

        c_budget = float(crow["DM_Total_Budget"].sum())
        c_spent = _safe_cost(ca)
        c_rev = _safe_revenue(ca)
        ca_conv = ca[ca["campaign_type"].astype(str).str.strip() == "Conversion"] if not ca.empty and "campaign_type" in ca.columns else ca
        c_roas_actual = _compute_roas(ca_conv)
        c_target_roas = float(crow["Target_ROAS"].iloc[0]) if len(crow) > 0 else None

        cn = _node(c_budget, c_spent, pe, de, dim)
        cn["revenue_target"] = {
            "spf": round(float(crow["SPF_Revenue_Target"].sum()), 2),
            "dm": round(float(crow["DM_Revenue_Target"].sum()), 2),
        }
        c_conv_rev = _safe_revenue(ca_conv)
        cn["actual_revenue"] = round(c_conv_rev, 2)
        cn["target_roas"] = round(c_target_roas, 2) if c_target_roas else None

        kpi_c: dict[str, Any] = {}
        if c_roas_actual is not None:
            kpi_entry: dict[str, Any] = {"actual": c_roas_actual}
            if c_target_roas and c_target_roas > 0:
                kpi_entry["target"] = round(c_target_roas, 2)
                kpi_entry["diff_pct"] = round(
                    (c_roas_actual - c_target_roas) / c_target_roas * 100, 1
                )
            kpi_c["ROAS"] = kpi_entry
        cn["kpi"] = kpi_c

        # ── Funnel level ──
        cn["by_funnel"] = {}
        for funnel_name, budget_col in FUNNEL_COLS.items():
            if budget_col not in crow.columns:
                continue
            f_budget = float(crow[budget_col].sum())
            pct_col = FUNNEL_PCT_COLS.get(funnel_name, "")
            f_budget_pct = float(crow[pct_col].iloc[0]) if pct_col in crow.columns and len(crow) > 0 else 0

            fa = (
                ca[ca["campaign_type"].astype(str).str.strip() == funnel_name]
                if not ca.empty and "campaign_type" in ca.columns
                else pd.DataFrame()
            )
            f_spent = _safe_cost(fa)

            fn = _node(f_budget, f_spent, pe, de, dim)
            fn["budget_pct"] = round(f_budget_pct, 1)

            # ── Platform level (from actual data only) ──
            fn["by_platform"] = {}
            if not fa.empty and "platform" in fa.columns:
                for plat in sorted(fa["platform"].astype(str).str.strip().unique()):
                    pa = fa[fa["platform"].astype(str).str.strip() == plat]
                    p_spent = _safe_cost(pa)
                    p_rev = _safe_revenue(pa)
                    p_roas = _compute_roas(pa)
                    fn["by_platform"][plat] = {
                        "spent": round(p_spent, 2),
                        "revenue": round(p_rev, 2),
                        "roas": p_roas,
                    }

            cn["by_funnel"][funnel_name] = fn

        hier[country] = cn

    nd["by_country"] = hier

    # ── 日度 pacing 曲线数据 ──
    if not am.empty and "date" in am.columns:
        daily_spend = am.groupby(am["date"].dt.day)["Cost"].sum().sort_index()
        cum_actual = []
        running = 0.0
        pacing_dates = []
        daily_ideal = tb / dim if dim > 0 else 0
        cum_ideal = []
        for day_num in range(1, dim + 1):
            pacing_dates.append(f"{ms}-{day_num:02d}")
            running += float(daily_spend.get(day_num, 0))
            cum_actual.append(round(running, 2))
            cum_ideal.append(round(daily_ideal * day_num, 2))
        nd["pacing_chart"] = {
            "dates": pacing_dates,
            "cumulative_actual": cum_actual,
            "cumulative_ideal": cum_ideal,
            "budget_total": round(tb, 2),
        }

    return nd