348 lines
16 KiB
Plaintext
348 lines
16 KiB
Plaintext
# -*- coding: utf-8 -*-
|
||
"""手工示例数据灌入:按 schema_ODS_doc.sql 的表结构写入 ODS。"""
|
||
from __future__ import annotations
|
||
|
||
import json
|
||
import os
|
||
from datetime import datetime
|
||
from typing import Any, Iterable
|
||
|
||
from psycopg2.extras import Json
|
||
|
||
from .base_task import BaseTask
|
||
|
||
|
||
class ManualIngestTask(BaseTask):
|
||
"""本地示例 JSON 灌入 ODS,确保表名/主键/插入列与 schema_ODS_doc.sql 对齐。"""
|
||
|
||
FILE_MAPPING: list[tuple[tuple[str, ...], str]] = [
|
||
(("member_profiles",), "billiards_ods.member_profiles"),
|
||
(("member_balance_changes",), "billiards_ods.member_balance_changes"),
|
||
(("member_stored_value_cards",), "billiards_ods.member_stored_value_cards"),
|
||
(("recharge_settlements",), "billiards_ods.recharge_settlements"),
|
||
(("settlement_records",), "billiards_ods.settlement_records"),
|
||
(("assistant_cancellation_records",), "billiards_ods.assistant_cancellation_records"),
|
||
(("assistant_accounts_master",), "billiards_ods.assistant_accounts_master"),
|
||
(("assistant_service_records",), "billiards_ods.assistant_service_records"),
|
||
(("site_tables_master",), "billiards_ods.site_tables_master"),
|
||
(("table_fee_discount_records",), "billiards_ods.table_fee_discount_records"),
|
||
(("table_fee_transactions",), "billiards_ods.table_fee_transactions"),
|
||
(("goods_stock_movements",), "billiards_ods.goods_stock_movements"),
|
||
(("stock_goods_category_tree",), "billiards_ods.stock_goods_category_tree"),
|
||
(("goods_stock_summary",), "billiards_ods.goods_stock_summary"),
|
||
(("payment_transactions",), "billiards_ods.payment_transactions"),
|
||
(("refund_transactions",), "billiards_ods.refund_transactions"),
|
||
(("platform_coupon_redemption_records",), "billiards_ods.platform_coupon_redemption_records"),
|
||
(("group_buy_redemption_records",), "billiards_ods.group_buy_redemption_records"),
|
||
(("group_buy_packages",), "billiards_ods.group_buy_packages"),
|
||
(("settlement_ticket_details",), "billiards_ods.settlement_ticket_details"),
|
||
(("store_goods_master",), "billiards_ods.store_goods_master"),
|
||
(("tenant_goods_master",), "billiards_ods.tenant_goods_master"),
|
||
(("store_goods_sales_records",), "billiards_ods.store_goods_sales_records"),
|
||
]
|
||
|
||
TABLE_SPECS: dict[str, dict[str, Any]] = {
|
||
"billiards_ods.member_profiles": {"pk": "id"},
|
||
"billiards_ods.member_balance_changes": {"pk": "id"},
|
||
"billiards_ods.member_stored_value_cards": {"pk": "id"},
|
||
"billiards_ods.recharge_settlements": {"pk": "id"},
|
||
"billiards_ods.settlement_records": {"pk": "id"},
|
||
"billiards_ods.assistant_cancellation_records": {"pk": "id", "json_cols": ["siteProfile"]},
|
||
"billiards_ods.assistant_accounts_master": {"pk": "id"},
|
||
"billiards_ods.assistant_service_records": {"pk": "id", "json_cols": ["siteProfile"]},
|
||
"billiards_ods.site_tables_master": {"pk": "id"},
|
||
"billiards_ods.table_fee_discount_records": {"pk": "id", "json_cols": ["siteProfile", "tableProfile"]},
|
||
"billiards_ods.table_fee_transactions": {"pk": "id", "json_cols": ["siteProfile"]},
|
||
"billiards_ods.goods_stock_movements": {"pk": "siteGoodsStockId"},
|
||
"billiards_ods.stock_goods_category_tree": {"pk": "id", "json_cols": ["categoryBoxes"]},
|
||
"billiards_ods.goods_stock_summary": {"pk": "siteGoodsId"},
|
||
"billiards_ods.payment_transactions": {"pk": "id", "json_cols": ["siteProfile"]},
|
||
"billiards_ods.refund_transactions": {"pk": "id", "json_cols": ["siteProfile"]},
|
||
"billiards_ods.platform_coupon_redemption_records": {"pk": "id"},
|
||
"billiards_ods.tenant_goods_master": {"pk": "id"},
|
||
"billiards_ods.group_buy_packages": {"pk": "id"},
|
||
"billiards_ods.group_buy_redemption_records": {"pk": "id"},
|
||
"billiards_ods.settlement_ticket_details": {
|
||
"pk": "orderSettleId",
|
||
"json_cols": ["memberProfile", "orderItem", "tenantMemberCardLogs"],
|
||
},
|
||
"billiards_ods.store_goods_master": {"pk": "id"},
|
||
"billiards_ods.store_goods_sales_records": {"pk": "id"},
|
||
}
|
||
|
||
def get_task_code(self) -> str:
|
||
"""返回任务编码。"""
|
||
return "MANUAL_INGEST"
|
||
|
||
def execute(self, cursor_data: dict | None = None) -> dict:
|
||
"""从目录读取 JSON,按表定义批量入库。"""
|
||
data_dir = (
|
||
self.config.get("manual.data_dir")
|
||
or self.config.get("pipeline.ingest_source_dir")
|
||
or r"c:\dev\LLTQ\ETL\feiqiu-ETL\etl_billiards\tests\testdata_json"
|
||
)
|
||
if not os.path.exists(data_dir):
|
||
self.logger.error("Data directory not found: %s", data_dir)
|
||
return {"status": "error", "message": "Directory not found"}
|
||
|
||
counts = {"fetched": 0, "inserted": 0, "updated": 0, "skipped": 0, "errors": 0}
|
||
|
||
for filename in sorted(os.listdir(data_dir)):
|
||
if not filename.endswith(".json"):
|
||
continue
|
||
filepath = os.path.join(data_dir, filename)
|
||
try:
|
||
with open(filepath, "r", encoding="utf-8") as fh:
|
||
raw_entries = json.load(fh)
|
||
except Exception:
|
||
counts["errors"] += 1
|
||
self.logger.exception("Failed to read %s", filename)
|
||
continue
|
||
|
||
entries = raw_entries if isinstance(raw_entries, list) else [raw_entries]
|
||
records = self._extract_records(entries)
|
||
if not records:
|
||
counts["skipped"] += 1
|
||
continue
|
||
|
||
target_table = self._match_by_filename(filename)
|
||
if not target_table:
|
||
self.logger.warning("No mapping found for file: %s", filename)
|
||
counts["skipped"] += 1
|
||
continue
|
||
|
||
self.logger.info("Ingesting %s into %s", filename, target_table)
|
||
try:
|
||
inserted, updated = self._ingest_table(target_table, records, filename)
|
||
counts["inserted"] += inserted
|
||
counts["updated"] += updated
|
||
counts["fetched"] += len(records)
|
||
except Exception:
|
||
counts["errors"] += 1
|
||
self.logger.exception("Error processing %s", filename)
|
||
self.db.rollback()
|
||
continue
|
||
|
||
try:
|
||
self.db.commit()
|
||
except Exception:
|
||
self.db.rollback()
|
||
raise
|
||
|
||
return {"status": "SUCCESS", "counts": counts}
|
||
|
||
def _match_by_filename(self, filename: str) -> str | None:
|
||
"""根据文件名关键字匹配目标表。"""
|
||
for keywords, table in self.FILE_MAPPING:
|
||
if any(keyword and keyword in filename for keyword in keywords):
|
||
return table
|
||
return None
|
||
|
||
def _extract_records(self, raw_entries: Iterable[Any]) -> list[dict]:
|
||
"""兼容多层 data/list 包装,抽取记录列表。"""
|
||
records: list[dict] = []
|
||
for entry in raw_entries:
|
||
if isinstance(entry, dict):
|
||
preferred = entry
|
||
if "data" in entry and not any(k not in {"data", "code"} for k in entry.keys()):
|
||
preferred = entry["data"]
|
||
data = preferred
|
||
if isinstance(data, dict):
|
||
# 特殊处理 settleList(充值、结算记录):展开 data.settleList 下的 settleList,抛弃上层 siteProfile
|
||
if "settleList" in data:
|
||
settle_list_val = data.get("settleList")
|
||
if isinstance(settle_list_val, dict):
|
||
settle_list_iter = [settle_list_val]
|
||
elif isinstance(settle_list_val, list):
|
||
settle_list_iter = settle_list_val
|
||
else:
|
||
settle_list_iter = []
|
||
|
||
handled = False
|
||
for item in settle_list_iter or []:
|
||
if not isinstance(item, dict):
|
||
continue
|
||
inner = item.get("settleList")
|
||
merged = dict(inner) if isinstance(inner, dict) else dict(item)
|
||
# 保留 siteProfile 供后续字段补充,但不落库
|
||
site_profile = data.get("siteProfile")
|
||
if isinstance(site_profile, dict):
|
||
merged.setdefault("siteProfile", site_profile)
|
||
records.append(merged)
|
||
handled = True
|
||
if handled:
|
||
continue
|
||
|
||
list_used = False
|
||
for v in data.values():
|
||
if isinstance(v, list) and v and isinstance(v[0], dict):
|
||
records.extend(v)
|
||
list_used = True
|
||
break
|
||
if list_used:
|
||
continue
|
||
if isinstance(data, list) and data and isinstance(data[0], dict):
|
||
records.extend(data)
|
||
elif isinstance(data, dict):
|
||
records.append(data)
|
||
elif isinstance(entry, list):
|
||
records.extend([item for item in entry if isinstance(item, dict)])
|
||
return records
|
||
|
||
def _get_table_columns(self, table: str) -> list[tuple[str, str, str]]:
|
||
"""查询 information_schema,获取目标表列信息。"""
|
||
cache = getattr(self, "_table_columns_cache", {})
|
||
if table in cache:
|
||
return cache[table]
|
||
if "." in table:
|
||
schema, name = table.split(".", 1)
|
||
else:
|
||
schema, name = "public", table
|
||
sql = """
|
||
SELECT column_name, data_type, udt_name
|
||
FROM information_schema.columns
|
||
WHERE table_schema = %s AND table_name = %s
|
||
ORDER BY ordinal_position
|
||
"""
|
||
with self.db.conn.cursor() as cur:
|
||
cur.execute(sql, (schema, name))
|
||
cols = [(r[0], (r[1] or "").lower(), (r[2] or "").lower()) for r in cur.fetchall()]
|
||
cache[table] = cols
|
||
self._table_columns_cache = cache
|
||
return cols
|
||
|
||
def _ingest_table(self, table: str, records: list[dict], source_file: str) -> tuple[int, int]:
|
||
"""构建 INSERT/ON CONFLICT 语句并批量执行。"""
|
||
spec = self.TABLE_SPECS.get(table)
|
||
if not spec:
|
||
raise ValueError(f"No table spec for {table}")
|
||
|
||
pk_col = spec.get("pk")
|
||
json_cols = set(spec.get("json_cols", []))
|
||
json_cols_lower = {c.lower() for c in json_cols}
|
||
|
||
columns_info = self._get_table_columns(table)
|
||
columns = [c[0] for c in columns_info]
|
||
db_json_cols_lower = {
|
||
c[0].lower() for c in columns_info if c[1] in ("json", "jsonb") or c[2] in ("json", "jsonb")
|
||
}
|
||
pk_col_db = None
|
||
if pk_col:
|
||
pk_col_db = next((c for c in columns if c.lower() == pk_col.lower()), pk_col)
|
||
|
||
placeholders = ", ".join(["%s"] * len(columns))
|
||
col_list = ", ".join(f'"{c}"' for c in columns)
|
||
sql = f'INSERT INTO {table} ({col_list}) VALUES ({placeholders})'
|
||
if pk_col_db:
|
||
update_cols = [c for c in columns if c != pk_col_db]
|
||
set_clause = ", ".join(f'"{c}"=EXCLUDED."{c}"' for c in update_cols)
|
||
sql += f' ON CONFLICT ("{pk_col_db}") DO UPDATE SET {set_clause}'
|
||
sql += " RETURNING (xmax = 0) AS inserted"
|
||
|
||
params = []
|
||
now = datetime.now()
|
||
json_dump = lambda v: json.dumps(v, ensure_ascii=False) # noqa: E731
|
||
for rec in records:
|
||
merged_rec = rec if isinstance(rec, dict) else {}
|
||
data_part = merged_rec.get("data")
|
||
while isinstance(data_part, dict):
|
||
merged_rec = {**data_part, **merged_rec}
|
||
data_part = data_part.get("data")
|
||
|
||
# 针对充值/结算,补齐 siteProfile 中的店铺信息
|
||
if table in {
|
||
"billiards_ods.recharge_settlements",
|
||
"billiards_ods.settlement_records",
|
||
}:
|
||
site_profile = merged_rec.get("siteProfile") or merged_rec.get("site_profile")
|
||
if isinstance(site_profile, dict):
|
||
merged_rec.setdefault("tenantid", site_profile.get("tenant_id") or site_profile.get("tenantId"))
|
||
merged_rec.setdefault("siteid", site_profile.get("id") or site_profile.get("siteId"))
|
||
merged_rec.setdefault("sitename", site_profile.get("shop_name") or site_profile.get("siteName"))
|
||
|
||
pk_val = self._get_value_case_insensitive(merged_rec, pk_col) if pk_col else None
|
||
if pk_col and (pk_val is None or pk_val == ""):
|
||
continue
|
||
|
||
row_vals = []
|
||
for col_name, data_type, udt in columns_info:
|
||
col_lower = col_name.lower()
|
||
if col_lower == "payload":
|
||
row_vals.append(Json(rec, dumps=json_dump))
|
||
continue
|
||
if col_lower == "source_file":
|
||
row_vals.append(source_file)
|
||
continue
|
||
if col_lower == "fetched_at":
|
||
row_vals.append(merged_rec.get(col_name, now))
|
||
continue
|
||
|
||
value = self._normalize_scalar(self._get_value_case_insensitive(merged_rec, col_name))
|
||
|
||
if col_lower in json_cols_lower or col_lower in db_json_cols_lower:
|
||
row_vals.append(Json(value, dumps=json_dump) if value is not None else None)
|
||
continue
|
||
|
||
casted = self._cast_value(value, data_type)
|
||
row_vals.append(casted)
|
||
params.append(tuple(row_vals))
|
||
|
||
if not params:
|
||
return 0, 0
|
||
|
||
inserted = 0
|
||
updated = 0
|
||
with self.db.conn.cursor() as cur:
|
||
for row in params:
|
||
cur.execute(sql, row)
|
||
flag = cur.fetchone()[0]
|
||
if flag:
|
||
inserted += 1
|
||
else:
|
||
updated += 1
|
||
return inserted, updated
|
||
|
||
@staticmethod
|
||
def _get_value_case_insensitive(record: dict, col: str | None):
|
||
"""忽略大小写获取值,兼容 information_schema 与 JSON 原始字段。"""
|
||
if record is None or col is None:
|
||
return None
|
||
if col in record:
|
||
return record.get(col)
|
||
col_lower = col.lower()
|
||
for k, v in record.items():
|
||
if isinstance(k, str) and k.lower() == col_lower:
|
||
return v
|
||
return None
|
||
|
||
@staticmethod
|
||
def _normalize_scalar(value):
|
||
"""将空字符串/空 JSON 规范为 None,避免类型转换错误。"""
|
||
if value == "" or value == "{}" or value == "[]":
|
||
return None
|
||
return value
|
||
|
||
@staticmethod
|
||
def _cast_value(value, data_type: str):
|
||
"""根据列类型做简单转换,保证批量插入兼容。"""
|
||
if value is None:
|
||
return None
|
||
dt = (data_type or "").lower()
|
||
if dt in ("integer", "bigint", "smallint"):
|
||
if isinstance(value, bool):
|
||
return int(value)
|
||
try:
|
||
return int(value)
|
||
except Exception:
|
||
return None
|
||
if dt in ("numeric", "double precision", "real", "decimal"):
|
||
if isinstance(value, bool):
|
||
return int(value)
|
||
try:
|
||
return float(value)
|
||
except Exception:
|
||
return None
|
||
if dt.startswith("timestamp") or dt in ("date", "time", "interval"):
|
||
return value if isinstance(value, str) else None
|
||
return value
|