feat: add TDengine connector + schema modules with tests

This commit is contained in:
张宗平
2026-06-11 15:38:59 +08:00
parent 0788daf2af
commit 7d1c03b98d
3 changed files with 276 additions and 0 deletions
+99
View File
@@ -0,0 +1,99 @@
"""connector 模块测试(mock taosws.connect)。"""
from unittest.mock import patch, MagicMock
import pytest
from ts_anomaly_td.connector import TDConnection, ConnectionError, _MAX_RETRY_DURATION
@pytest.fixture
def mock_connect():
"""mock taosws.connect 返回假连接。"""
with patch("ts_anomaly_td.connector.taosws.connect") as m:
mock_conn = MagicMock()
mock_conn.execute.return_value = [[("3.3.6.0",)]]
m.return_value = mock_conn
yield m
def test_connect_success(mock_connect):
"""首次连接成功。"""
conn = TDConnection()
assert conn.is_healthy()
mock_connect.assert_called_once()
def test_execute_query(mock_connect):
"""execute 返回行列表。"""
mock_conn = mock_connect.return_value
# taosws 返回可迭代的行对象;每行 row 是 (1, 'a') / (2, 'b')
mock_conn.execute.return_value = [(1, "a"), (2, "b")]
conn = TDConnection()
rows = conn.execute("SELECT 1")
assert rows == [(1, "a"), (2, "b")]
def test_execute_no_result(mock_connect):
"""execute_no_result 不返回值。"""
conn = TDConnection()
result = conn.execute_no_result("CREATE DATABASE test")
assert result is None
def test_retry_logic(mock_connect):
"""前 2 次失败、第 3 次成功。"""
import ts_anomaly_td.connector as mod
# 缩短退避以加速测试
mod._MAX_BACKOFF = 0.01
try:
mock_connect.side_effect = [
Exception("connection refused"),
Exception("connection refused"),
MagicMock(execute=MagicMock(return_value=[[("3.3.6.0",)]])),
]
conn = TDConnection()
assert conn.is_healthy()
assert mock_connect.call_count == 3
finally:
mod._MAX_BACKOFF = 300
def test_retry_timeout(mock_connect):
"""持续失败超过 30 分钟应抛出 ConnectionError。"""
import ts_anomaly_td.connector as mod
mod._MAX_RETRY_DURATION = 0.1
mod._MAX_BACKOFF = 0.01
try:
mock_connect.side_effect = Exception("always fails")
with pytest.raises(ConnectionError, match="30"):
conn = TDConnection()
# 触发懒加载连接:访问 .conn 属性(非 is_healthy,后者会吞掉异常)
_ = conn.conn
finally:
mod._MAX_RETRY_DURATION = 1800
mod._MAX_BACKOFF = 300
def test_close(mock_connect):
"""close 关闭底层连接。"""
conn = TDConnection()
# 触发懒加载连接
conn.is_healthy()
conn.close()
mock_connect.return_value.close.assert_called_once()
def test_batch_insert(mock_connect):
"""batch_insert 拼接 INSERT SQL。"""
conn = TDConnection()
rows = [(1000, 1.0, 0), (2000, 2.0, 1)]
conn.batch_insert("test", rows)
mock_conn = mock_connect.return_value
# 验证 execute 被调用
calls = [str(c[0][0]) for c in mock_conn.execute.call_args_list]
assert any("INSERT INTO s_test VALUES" in c for c in calls)
+117
View File
@@ -0,0 +1,117 @@
"""TDengine 连接模块:taos-ws-py WebSocket 封装 + 指数退避重试。"""
import time
import logging
import taosws
logger = logging.getLogger(__name__)
_MAX_RETRY_DURATION = 1800 # 30 分钟
_MAX_BACKOFF = 300 # 5 分钟
_INSERT_BATCH_SIZE = 500
class ConnectionError(Exception):
"""TDengine 连接不可用时抛出。"""
pass
class TDConnection:
"""TDengine WebSocket 连接封装。
Args:
url: WebSocket 连接字符串,默认 ws://root:taosdata@localhost:6041
"""
def __init__(self, url: str = "ws://root:taosdata@localhost:6041"):
self._url = url
self._conn: taosws.Connection | None = None
@property
def conn(self) -> taosws.Connection:
if self._conn is None:
self._conn = self._connect_with_retry()
return self._conn
def _connect_with_retry(self) -> taosws.Connection:
"""指数退避重试连接 TDengine。
初始退避 1s,倍增因子 2,最大间隔 300s,最长持续 1800s。
每次重试前执行 SELECT SERVER_VERSION() 验证服务可用。
Raises:
ConnectionError: 30 分钟内仍不可用
"""
backoff = 1.0
started_at = time.monotonic()
last_exc = None
while True:
elapsed = time.monotonic() - started_at
if elapsed > _MAX_RETRY_DURATION:
raise ConnectionError(
f"TDengine 连接失败,已重试 {elapsed:.0f}s(超过 30 分钟上限)"
) from last_exc
try:
conn = taosws.connect(self._url)
conn.execute("SELECT SERVER_VERSION()")
logger.info("TDengine 连接成功 (url=%s)", self._url)
return conn
except Exception as e:
last_exc = e
logger.warning(
"TDengine 连接失败 (attempt, backoff=%.1fs, elapsed=%.0fs): %s",
backoff, elapsed, e,
)
time.sleep(backoff)
backoff = min(backoff * 2, _MAX_BACKOFF)
def execute(self, sql: str) -> list[tuple]:
"""执行查询 SQL,返回结果行列表。"""
logger.debug("execute: %s", sql[:120])
result = self.conn.execute(sql)
rows = []
if result is not None:
for row in result:
rows.append(tuple(row))
return rows
def execute_no_result(self, sql: str) -> None:
"""执行非查询 SQLINSERT / DDL)。"""
logger.debug("execute_no_result: %s", sql[:120])
self.conn.execute(sql)
def close(self):
"""关闭连接。"""
if self._conn is not None:
self._conn.close()
self._conn = None
def is_healthy(self) -> bool:
"""检查连接健康状态。"""
try:
self.execute("SELECT SERVER_VERSION()")
return True
except Exception:
return False
def batch_insert(self, stable: str, rows: list[tuple[int, float, int]]) -> None:
"""批量 INSERT 数据到 subtable。
每 500 条一个 batch,单次 execute_no_result()。
Args:
stable: supertable 名称(不含 ds_ 前缀)
rows: [(ts_ms, value, label), ...]
"""
total = len(rows)
for i in range(0, total, _INSERT_BATCH_SIZE):
batch = rows[i:i + _INSERT_BATCH_SIZE]
values_clauses = []
for ts, val, label in batch:
values_clauses.append(f"({ts}, {val}, {label})")
sql = f"INSERT INTO s_{stable} VALUES {' '.join(values_clauses)}"
self.execute_no_result(sql)
logger.debug("batch_insert: %d/%d rows", min(i + _INSERT_BATCH_SIZE, total), total)
logger.info("batch_insert complete: %d rows into s_%s", total, stable)
+60
View File
@@ -0,0 +1,60 @@
"""TDengine DDL 模块:supertable / subtable 创建。"""
import logging
logger = logging.getLogger(__name__)
def create_database(conn, db_name: str = "ts_anomaly") -> None:
"""创建数据库(不存在时创建)。"""
conn.execute_no_result(f"CREATE DATABASE IF NOT EXISTS {db_name} KEEP 36500")
logger.info("database %s ready", db_name)
def use_database(conn, db_name: str = "ts_anomaly") -> None:
"""切换到指定数据库。"""
conn.execute_no_result(f"USE {db_name}")
def create_supertable(conn, stable: str) -> None:
"""创建 supertable(不存在时创建)。
Args:
stable: 数据集名称,如 finance_001
"""
sql = (
f"CREATE STABLE IF NOT EXISTS ds_{stable} "
f"(ts TIMESTAMP, value DOUBLE, label INT) "
f"TAGS (series_id INT)"
)
conn.execute_no_result(sql)
logger.info("supertable ds_%s ready", stable)
def create_subtable(conn, stable: str, series_id: int = 1) -> None:
"""创建 subtable(不存在时创建)。
Args:
stable: 数据集名称
series_id: 时序 ID(单变量数据集为 1)
"""
sql = (
f"CREATE TABLE IF NOT EXISTS s_{stable} "
f"USING ds_{stable} TAGS ({series_id})"
)
conn.execute_no_result(sql)
logger.info("subtable s_%s ready (series_id=%d)", stable, series_id)
def setup_schema(conn, stable: str, db_name: str = "ts_anomaly") -> None:
"""一键建库建表。
Args:
conn: TDConnection 实例
stable: 数据集名称
db_name: 数据库名称
"""
create_database(conn, db_name)
use_database(conn, db_name)
create_supertable(conn, stable)
create_subtable(conn, stable)