267 lines
9.2 KiB
Python
267 lines
9.2 KiB
Python
"""数据同步调度中心模块。
|
||
|
||
该模块作为数据同步的调度中心，统一管理各类型数据的同步流程。
|
||
|
||
【重要规范 - sync.py 职责范围】
|
||
本模块**仅包含每日更新的数据接口**，季度/低频数据不应放入此文件：
|
||
|
||
✅ 本模块包含的同步逻辑（每日更新）：
|
||
- api_daily.py: 日线数据同步 (DailySync 类)
|
||
- api_daily_basic.py: 每日指标数据同步 (DailyBasicSync 类)
|
||
- api_bak_basic.py: 历史股票列表同步 (BakBasicSync 类)
|
||
- api_pro_bar.py: Pro Bar 数据同步 (ProBarSync 类)
|
||
- api_stock_st.py: ST股票列表同步 (StockSTSync 类)
|
||
- api_stk_limit.py: 涨跌停价格同步 (StkLimitSync 类)
|
||
- api_cyq_perf.py: 筹码分布数据同步 (CyqPerfSync 类)
|
||
- api_stock_basic.py: 股票基本信息同步
|
||
- api_trade_cal.py: 交易日历同步
|
||
|
||
❌ 不应包含的同步逻辑（季度/低频更新）：
|
||
- financial_data/: 财务数据（利润表、资产负债表、现金流量表等）
|
||
使用方式：
|
||
from src.data.api_wrappers.financial_data.api_financial_sync import sync_financial
|
||
sync_financial()
|
||
|
||
- api_namechange.py: 股票名称变更（不频繁）
|
||
使用方式：
|
||
from src.data.api_wrappers import sync_namechange
|
||
sync_namechange(force=True)
|
||
|
||
【架构说明】
|
||
本模块使用 SyncRegistry 注册表模式管理同步任务，避免手动罗列各个接口。
|
||
同步任务在 api_wrappers/__init__.py 中自动注册，新增接口无需修改 sync.py。
|
||
|
||
使用方式：
|
||
# 预览同步（检查数据量，不写入）
|
||
from src.data.sync import preview_sync
|
||
preview = preview_sync()
|
||
|
||
# 同步所有每日更新数据（不包括财务数据、namechange）
|
||
from src.data.sync import sync_all_data
|
||
result = sync_all_data()
|
||
|
||
# 强制全量重载
|
||
result = sync_all_data(force_full=True)
|
||
|
||
# 查看已注册的所有同步任务
|
||
from src.data.sync_registry import sync_registry
|
||
tasks = sync_registry.list_tasks()
|
||
for task in tasks:
|
||
print(f"{task.name}: {task.display_name}")
|
||
"""
|
||
|
||
from typing import Optional, Dict, Union, Any
|
||
|
||
import pandas as pd
|
||
|
||
# 导入以触发自动注册
|
||
from src.data import api_wrappers # noqa: F401
|
||
from src.data.sync_registry import sync_registry
|
||
from src.data.api_wrappers import sync_all_stocks
|
||
from src.data.sync_logger import SyncLogManager
|
||
|
||
|
||
def sync_all_data(
|
||
force_full: bool = False,
|
||
max_workers: Optional[int] = None,
|
||
dry_run: bool = False,
|
||
selected: Optional[list[str]] = None,
|
||
) -> dict[str, Any]:
|
||
"""同步所有每日更新的数据类型。
|
||
|
||
【重要】本函数仅同步每日更新的数据，不包含季度/低频数据。
|
||
|
||
【自动注册机制】
|
||
同步任务在 api_wrappers/__init__.py 中自动注册到 SyncRegistry。
|
||
当前注册的同步任务（按执行顺序）：
|
||
1. trade_cal: 交易日历缓存
|
||
2. stock_basic: 股票基本信息
|
||
3. pro_bar: Pro Bar 数据（复权、换手率、量比）
|
||
4. daily_basic: 每日指标（PE、PB、换手率、市值）
|
||
5. bak_basic: 历史股票列表
|
||
6. stock_st: ST股票列表
|
||
7. stk_limit: 每日涨跌停价格
|
||
8. cyq_perf: 每日筹码及胜率
|
||
|
||
新增接口时，只需在 api_wrappers/__init__.py 中添加注册代码，
|
||
无需修改本函数。
|
||
|
||
【不包含的同步（需单独调用）】
|
||
- 财务数据: 利润表、资产负债表、现金流量表（季度更新）
|
||
使用: from src.data.api_wrappers.financial_data.api_financial_sync import sync_financial
|
||
调用: sync_financial()
|
||
|
||
- 名称变更 (namechange): 股票曾用名（低频更新）
|
||
使用: from src.data.api_wrappers import sync_namechange
|
||
调用: sync_namechange(force=True)
|
||
|
||
Args:
|
||
force_full: 若为 True，强制所有数据类型完整重载
|
||
max_workers: 日线数据同步的工作线程数（默认: 10）
|
||
dry_run: 若为 True，仅显示将要同步的内容，不写入数据
|
||
selected: 只同步指定的任务列表，None表示同步所有
|
||
例如: selected=["trade_cal", "stock_basic"] 只同步交易日历和股票基本信息
|
||
|
||
Returns:
|
||
映射数据类型到同步结果的字典
|
||
|
||
Example:
|
||
>>> result = sync_all_data()
|
||
>>>
|
||
>>> # 强制完整重载
|
||
>>> result = sync_all_data(force_full=True)
|
||
>>>
|
||
>>> # Dry run
|
||
>>> result = sync_all_data(dry_run=True)
|
||
>>>
|
||
>>> # 只同步特定任务
|
||
>>> result = sync_all_data(selected=['trade_cal', 'pro_bar'])
|
||
>>>
|
||
>>> # 查看所有可用任务
|
||
>>> from src.data.sync_registry import sync_registry
|
||
>>> tasks = sync_registry.list_tasks()
|
||
>>> for t in tasks:
|
||
... print(f"{t.name}: {t.display_name}")
|
||
"""
|
||
# 记录调度中心开始
|
||
log_manager = SyncLogManager()
|
||
sync_mode = "full" if force_full else "incremental"
|
||
selected_str = ",".join(selected) if selected else "all"
|
||
log_manager.start_sync(
|
||
table_name="daily_data_batch",
|
||
sync_type=sync_mode,
|
||
metadata={
|
||
"selected": selected_str,
|
||
"dry_run": dry_run,
|
||
"max_workers": max_workers,
|
||
},
|
||
)
|
||
|
||
try:
|
||
result = sync_registry.sync_all(
|
||
force_full=force_full,
|
||
max_workers=max_workers,
|
||
dry_run=dry_run,
|
||
selected=selected,
|
||
)
|
||
|
||
# 计算成功/失败数量
|
||
success_count = 0
|
||
failed_count = 0
|
||
total_records = 0
|
||
for task_name, task_result in result.items():
|
||
if isinstance(task_result, dict):
|
||
if task_result.get("status") == "error":
|
||
failed_count += 1
|
||
else:
|
||
success_count += 1
|
||
# 累加记录数（如果有）
|
||
if "rows" in task_result:
|
||
total_records += task_result.get("rows", 0)
|
||
elif isinstance(task_result, pd.DataFrame):
|
||
success_count += 1
|
||
total_records += len(task_result)
|
||
else:
|
||
success_count += 1
|
||
|
||
# 记录完成日志
|
||
status = "partial" if failed_count > 0 else "success"
|
||
error_msg = f"Failed: {failed_count} tasks" if failed_count > 0 else None
|
||
log_manager.complete_sync(
|
||
table_name="daily_data_batch",
|
||
sync_type=sync_mode,
|
||
status=status,
|
||
records_inserted=total_records,
|
||
error_message=error_msg,
|
||
metadata={
|
||
"selected": selected_str,
|
||
"dry_run": dry_run,
|
||
"max_workers": max_workers,
|
||
},
|
||
)
|
||
|
||
return result
|
||
|
||
except Exception as e:
|
||
# 记录失败日志
|
||
log_manager.complete_sync(
|
||
table_name="daily_data_batch",
|
||
sync_type=sync_mode,
|
||
status="failed",
|
||
error_message=str(e),
|
||
metadata={
|
||
"selected": selected_str,
|
||
"dry_run": dry_run,
|
||
"max_workers": max_workers,
|
||
},
|
||
)
|
||
raise
|
||
|
||
|
||
def list_sync_tasks() -> list[dict[str, Any]]:
|
||
"""列出所有已注册的同步任务。
|
||
|
||
Returns:
|
||
任务信息列表，每个任务包含 name, display_name, description, order, enabled
|
||
|
||
Example:
|
||
>>> tasks = list_sync_tasks()
|
||
>>> for task in tasks:
|
||
... print(f"{task['order']:2d}. {task['name']}: {task['display_name']}")
|
||
"""
|
||
tasks = sync_registry.list_tasks()
|
||
return [
|
||
{
|
||
"name": t.name,
|
||
"display_name": t.display_name,
|
||
"description": t.description,
|
||
"order": t.order,
|
||
"enabled": t.enabled,
|
||
}
|
||
for t in tasks
|
||
]
|
||
|
||
|
||
if __name__ == "__main__":
|
||
print("=" * 60)
|
||
print("Data Sync Module")
|
||
print("=" * 60)
|
||
print("\nRegistered sync tasks:")
|
||
print("-" * 60)
|
||
|
||
tasks = list_sync_tasks()
|
||
for task in tasks:
|
||
status = "[启用]" if task["enabled"] else "[禁用]"
|
||
print(f" {status} {task['order']:2d}. {task['name']}: {task['display_name']}")
|
||
|
||
print("-" * 60)
|
||
print(f"\nTotal: {len(tasks)} tasks")
|
||
print("\nUsage:")
|
||
print(" # Sync all data types at once (RECOMMENDED)")
|
||
print(" from src.data.sync import sync_all_data")
|
||
print(" result = sync_all_data() # Incremental sync all")
|
||
print(" result = sync_all_data(force_full=True) # Full reload")
|
||
print("")
|
||
print(" # Sync selected data types only")
|
||
print(" result = sync_all_data(selected=['trade_cal', 'pro_bar'])")
|
||
print("")
|
||
print(" # List all available sync tasks")
|
||
print(" tasks = list_sync_tasks()")
|
||
print("")
|
||
print(" # Preview before sync (recommended)")
|
||
print(" preview = preview_sync()")
|
||
print("")
|
||
print(" # Dry run (preview only)")
|
||
print(" result = sync_all(dry_run=True)")
|
||
print("")
|
||
print(" # Actual sync")
|
||
print(" result = sync_all() # Incremental sync")
|
||
print(" result = sync_all(force_full=True) # Full reload")
|
||
print("\n" + "=" * 60)
|
||
|
||
# Run sync_all_data by default
|
||
print("\n[Main] Running sync_all_data()...")
|
||
result = sync_all_data()
|
||
print("\n[Main] Sync completed!")
|
||
print(f"Total data types synced: {len(result)}")
|