Add campaign experiment and control reporting

update.sh

@@ -16,7 +16,7 @@ RAW_CSV="$TMP_DIR/tableau.csv"
 if [ -n "${SOURCE_CSV:-}" ]; then
   cp "$SOURCE_CSV" "$RAW_CSV"
 else
-  TABLEAU_URL=${TABLEAU_CSV_URL:-'https://mars.ushareit.org/views/DSPSpend/DSPDailyData.csv?:showVizHome=no&%E7%BB%B4%E5%BA%A61=pkg'}
+  TABLEAU_URL=${TABLEAU_CSV_URL:-'https://mars.ushareit.org/views/DSPSpend/DSPDailyData.csv?:showVizHome=no&%E7%BB%B4%E5%BA%A61=pkg&%E7%BB%B4%E5%BA%A62=adping_id'}
   if [ -n "${TABLEAU_COOKIE:-}" ]; then
     curl --fail --silent --show-error --location \
       --cookie "$TABLEAU_COOKIE" "$TABLEAU_URL" --output "$RAW_CSV"
@@ -25,7 +25,7 @@ else
   fi
 fi
 
-python3 - "$RAW_CSV" "$ROOT_DIR/data.json" <<'PY'
+python3 - "$RAW_CSV" "$ROOT_DIR/campaigns.csv" "$ROOT_DIR/data.json" <<'PY'
 import csv
 import datetime as dt
 import io
@@ -34,7 +34,8 @@ import pathlib
 import sys
 
 source_path = pathlib.Path(sys.argv[1])
-output_path = pathlib.Path(sys.argv[2])
+campaign_path = pathlib.Path(sys.argv[2])
+output_path = pathlib.Path(sys.argv[3])
 raw = source_path.read_bytes()
 head = raw[:500].lower()
 if b"<html" in head or b"tableau server" in head:
@@ -48,26 +49,10 @@ else:
 sample = text[:4096]
 delimiter = "\t" if sample.count("\t") > sample.count(",") else ","
 reader = csv.DictReader(io.StringIO(text), delimiter=delimiter)
-
-active_pkgs = [
-    "com.alibaba.aliexpresshd",
-    "com.edadeal.android",
-    "com.fxpro.direct.application",
-    "com.king.candycrushsaga",
-    "com.krom.android",
-    "com.vitastudio.mahjong",
-    "com.yandex.bank",
-    "daily.number.match.free.puzzle",
-    "ru.aliexpress.buyer",
-    "ru.kinopoisk",
-    "ru.ozon.app.android",
-    "tech.im",
-    "com.esprit.rpg.cntrctfft",
-    "com.uma.musicvk",
-    "com.amazon.mShop.android.shopping",
-    "com.esprit.rpg.lttlefirykng",
-]
-active = set(active_pkgs)
+with campaign_path.open(encoding="utf-8-sig", newline="") as f:
+    campaigns = list(csv.DictReader(f))
+campaign_by_key = {(r["pkg"], r["adping_id"]): r for r in campaigns if r["status"] == "Active"}
+active_pkgs = sorted({r["pkg"] for r in campaigns if r["status"] == "Active"})
 
 def pick(row, *names):
     for name in names:
@@ -84,16 +69,23 @@ def number(value):
     except ValueError:
         return 0.0
 
-daily_map = {}
+campaign_daily_map = {}
 for row in reader:
     pkg = pick(row, "维度1", "pkg", "Pkg").strip()
-    if pkg not in active:
+    adping_id = pick(row, "维度2", "adping_id", "adpingid").strip()
+    campaign = campaign_by_key.get((pkg, adping_id))
+    if not campaign:
         continue
     date = pick(row, "dt1", "dt", "date").strip()
     if not date:
         continue
-    key = (date, pkg)
-    item = daily_map.setdefault(key, {"date": date, "pkg": pkg, "impression": 0, "click": 0, "conversion": 0, "spend": 0.0})
+    key = (date, pkg, adping_id)
+    item = campaign_daily_map.setdefault(key, {
+        "date": date, "pkg": pkg, "adping_id": adping_id,
+        "group": campaign["group"], "geo": campaign["geo"],
+        "platform": campaign["platform"], "owner": campaign["owner"],
+        "impression": 0, "click": 0, "conversion": 0, "spend": 0.0,
+    })
     item["impression"] += round(number(pick(row, "impression", "Impression")))
     item["click"] += round(number(pick(row, "click", "Click")))
     item["conversion"] += round(number(pick(row, "conversions", "conversion", "Conversion")))
@@ -112,6 +104,13 @@ def metrics(item):
     item["ecpi"] = spend / conv if conv else None
     return item
 
+campaign_daily = [metrics(v) for _, v in sorted(campaign_daily_map.items())]
+daily_map = {}
+for row in campaign_daily:
+    key = (row["date"], row["pkg"])
+    item = daily_map.setdefault(key, {"date": row["date"], "pkg": row["pkg"], "impression": 0, "click": 0, "conversion": 0, "spend": 0.0})
+    for field in ("impression", "click", "conversion", "spend"):
+        item[field] += row[field]
 daily = [metrics(v) for _, v in sorted(daily_map.items())]
 summary = []
 for pkg in active_pkgs:
@@ -133,6 +132,8 @@ payload = {
         "date_start": dates[0] if dates else None,
         "date_end": dates[-1] if dates else None,
         "active_pkg_count": len(active_pkgs),
+        "active_campaign_count": len(campaigns),
+        "campaign_with_data_count": len({(r["pkg"], r["adping_id"]) for r in campaign_daily}),
         "pkg_with_data_count": sum(1 for x in summary if x["has_data"]),
         "definitions": {
             "CTR": "click / impression",
@@ -144,10 +145,11 @@ payload = {
     },
     "summary": summary,
     "daily": daily,
+    "campaigns": campaigns,
+    "campaign_daily": campaign_daily,
 }
 
 tmp = output_path.with_suffix(".json.tmp")
 tmp.write_text(json.dumps(payload, ensure_ascii=False, separators=(",", ":")), encoding="utf-8")
 tmp.replace(output_path)
 PY
-