Spaces:
Sleeping
Sleeping
Update data.py
Browse files
data.py
CHANGED
|
@@ -1,13 +1,10 @@
|
|
| 1 |
from __future__ import annotations
|
| 2 |
import os
|
| 3 |
import json
|
| 4 |
-
from typing import Optional, Dict, Any
|
| 5 |
from datetime import datetime
|
| 6 |
import pandas as pd
|
| 7 |
|
| 8 |
-
#
|
| 9 |
-
# - 既定: /tmp/adcopy_data(ephemeral)
|
| 10 |
-
# - 環境変数 DATA_DIR を設定すると、例: /data/adcopy_mab(HF Spaces の Persistent Storage)
|
| 11 |
DEFAULT_WRITABLE_DIR = "/tmp/adcopy_data"
|
| 12 |
DATA_DIR = os.environ.get("DATA_DIR", DEFAULT_WRITABLE_DIR)
|
| 13 |
|
|
@@ -15,20 +12,11 @@ LOG_PATH = os.path.join(DATA_DIR, "events.csv")
|
|
| 15 |
META_PATH = os.path.join(DATA_DIR, "meta.json")
|
| 16 |
|
| 17 |
SCHEMA = [
|
| 18 |
-
"ts",
|
| 19 |
-
"
|
| 20 |
-
"medium", # 媒体名 (例: FB, GDN)
|
| 21 |
-
"creative", # クリエイティブID/名前 (例: A1)
|
| 22 |
-
"is_control", # 0/1(コントロール群)
|
| 23 |
-
"impressions", # 表示数
|
| 24 |
-
"clicks", # クリック数(または目的コンバージョン)
|
| 25 |
-
"conversions", # 追加のCV(任意: 0 でもOK)
|
| 26 |
-
"cost", # コスト(任意)
|
| 27 |
-
"features_json" # クリエイティブ特徴量(dict をJSON文字列で)
|
| 28 |
]
|
| 29 |
|
| 30 |
def _ensure_storage():
|
| 31 |
-
"""初回起動時に保存先と空ファイルを準備。"""
|
| 32 |
os.makedirs(DATA_DIR, exist_ok=True)
|
| 33 |
if not os.path.exists(LOG_PATH):
|
| 34 |
pd.DataFrame(columns=SCHEMA).to_csv(LOG_PATH, index=False)
|
|
@@ -36,7 +24,7 @@ def _ensure_storage():
|
|
| 36 |
with open(META_PATH, "w", encoding="utf-8") as f:
|
| 37 |
json.dump({"created_at": datetime.utcnow().isoformat()}, f)
|
| 38 |
|
| 39 |
-
#
|
| 40 |
_ensure_storage()
|
| 41 |
|
| 42 |
def read_events() -> pd.DataFrame:
|
|
@@ -44,7 +32,6 @@ def read_events() -> pd.DataFrame:
|
|
| 44 |
df = pd.read_csv(LOG_PATH)
|
| 45 |
if df.empty:
|
| 46 |
return df
|
| 47 |
-
# 型整備
|
| 48 |
df["date"] = pd.to_datetime(df["date"]).dt.date.astype(str)
|
| 49 |
df["is_control"] = df["is_control"].fillna(0).astype(int)
|
| 50 |
for col in ["impressions", "clicks", "conversions"]:
|
|
@@ -55,7 +42,6 @@ def read_events() -> pd.DataFrame:
|
|
| 55 |
|
| 56 |
def append_events(rows: pd.DataFrame) -> None:
|
| 57 |
_ensure_storage()
|
| 58 |
-
# 必須列チェック & 補完
|
| 59 |
for c in SCHEMA:
|
| 60 |
if c not in rows.columns:
|
| 61 |
if c == "features_json":
|
|
|
|
| 1 |
from __future__ import annotations
|
| 2 |
import os
|
| 3 |
import json
|
|
|
|
| 4 |
from datetime import datetime
|
| 5 |
import pandas as pd
|
| 6 |
|
| 7 |
+
# 書き込み可能な場所をデフォルトにする
|
|
|
|
|
|
|
| 8 |
DEFAULT_WRITABLE_DIR = "/tmp/adcopy_data"
|
| 9 |
DATA_DIR = os.environ.get("DATA_DIR", DEFAULT_WRITABLE_DIR)
|
| 10 |
|
|
|
|
| 12 |
META_PATH = os.path.join(DATA_DIR, "meta.json")
|
| 13 |
|
| 14 |
SCHEMA = [
|
| 15 |
+
"ts", "date", "medium", "creative", "is_control",
|
| 16 |
+
"impressions", "clicks", "conversions", "cost", "features_json"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
]
|
| 18 |
|
| 19 |
def _ensure_storage():
|
|
|
|
| 20 |
os.makedirs(DATA_DIR, exist_ok=True)
|
| 21 |
if not os.path.exists(LOG_PATH):
|
| 22 |
pd.DataFrame(columns=SCHEMA).to_csv(LOG_PATH, index=False)
|
|
|
|
| 24 |
with open(META_PATH, "w", encoding="utf-8") as f:
|
| 25 |
json.dump({"created_at": datetime.utcnow().isoformat()}, f)
|
| 26 |
|
| 27 |
+
# インポート時に準備
|
| 28 |
_ensure_storage()
|
| 29 |
|
| 30 |
def read_events() -> pd.DataFrame:
|
|
|
|
| 32 |
df = pd.read_csv(LOG_PATH)
|
| 33 |
if df.empty:
|
| 34 |
return df
|
|
|
|
| 35 |
df["date"] = pd.to_datetime(df["date"]).dt.date.astype(str)
|
| 36 |
df["is_control"] = df["is_control"].fillna(0).astype(int)
|
| 37 |
for col in ["impressions", "clicks", "conversions"]:
|
|
|
|
| 42 |
|
| 43 |
def append_events(rows: pd.DataFrame) -> None:
|
| 44 |
_ensure_storage()
|
|
|
|
| 45 |
for c in SCHEMA:
|
| 46 |
if c not in rows.columns:
|
| 47 |
if c == "features_json":
|