Files
ProxyPool/app/repositories/proxy_repo.py

302 lines
10 KiB
Python

"""代理数据访问层 - 所有 SQL 操作收敛于此"""
import aiosqlite
from datetime import datetime, timedelta
from typing import List, Optional, Tuple, Union
from app.models.domain import Proxy
from app.core.log import logger
VALID_PROTOCOLS = ("http", "https", "socks4", "socks5")
def _to_datetime(value: Union[str, datetime, None]) -> Optional[datetime]:
if value is None:
return None
if isinstance(value, datetime):
return value
if isinstance(value, str):
for fmt in ("%Y-%m-%d %H:%M:%S", "%Y-%m-%d %H:%M:%S.%f", "%Y-%m-%dT%H:%M:%S", "%Y-%m-%dT%H:%M:%S.%f"):
try:
return datetime.strptime(value, fmt)
except ValueError:
continue
return None
def _row_to_proxy(row: Tuple) -> Proxy:
return Proxy(
ip=row[0],
port=row[1],
protocol=row[2],
score=row[3],
response_time_ms=row[4],
last_check=_to_datetime(row[5]),
created_at=_to_datetime(row[6]),
)
class ProxyRepository:
"""代理 Repository"""
@staticmethod
async def insert_or_update(
db: aiosqlite.Connection,
ip: str,
port: int,
protocol: str = "http",
score: int = 10,
) -> bool:
if protocol not in VALID_PROTOCOLS:
protocol = "http"
try:
await db.execute(
"""
INSERT INTO proxies (ip, port, protocol, score, last_check, created_at)
VALUES (?, ?, ?, ?, CURRENT_TIMESTAMP, CURRENT_TIMESTAMP)
ON CONFLICT(ip, port) DO UPDATE SET
protocol = excluded.protocol,
score = excluded.score,
last_check = CURRENT_TIMESTAMP
""",
(ip, port, protocol, score),
)
await db.commit()
return True
except Exception as e:
logger.error(f"insert_or_update proxy failed: {e}", exc_info=True)
return False
@staticmethod
async def update_score(
db: aiosqlite.Connection,
ip: str,
port: int,
delta: int,
min_score: int = 0,
max_score: int = 100,
) -> bool:
try:
# 原子更新:计算新分数并直接更新
await db.execute(
"""
UPDATE proxies
SET score = MAX(?, MIN(?, score + ?)),
last_check = CURRENT_TIMESTAMP
WHERE ip = ? AND port = ?
""",
(min_score, max_score, delta, ip, port),
)
# 删除分数已降至 0 及以下的代理
await db.execute(
"DELETE FROM proxies WHERE ip = ? AND port = ? AND score <= ?",
(ip, port, min_score),
)
await db.commit()
return db.total_changes > 0
except Exception as e:
logger.error(f"update_score failed: {e}", exc_info=True)
return False
@staticmethod
async def update_response_time(
db: aiosqlite.Connection,
ip: str,
port: int,
response_time_ms: float,
) -> bool:
try:
await db.execute(
"UPDATE proxies SET response_time_ms = ? WHERE ip = ? AND port = ?",
(response_time_ms, ip, port),
)
await db.commit()
return True
except Exception as e:
logger.error(f"update_response_time failed: {e}", exc_info=True)
return False
@staticmethod
async def delete(db: aiosqlite.Connection, ip: str, port: int) -> None:
await db.execute("DELETE FROM proxies WHERE ip = ? AND port = ?", (ip, port))
await db.commit()
@staticmethod
async def batch_delete(db: aiosqlite.Connection, proxies: List[Tuple[str, int]]) -> int:
if not proxies:
return 0
changes_before = db.total_changes
await db.executemany("DELETE FROM proxies WHERE ip = ? AND port = ?", proxies)
await db.commit()
return db.total_changes - changes_before
@staticmethod
async def get_by_ip_port(
db: aiosqlite.Connection, ip: str, port: int
) -> Optional[Proxy]:
async with db.execute(
"SELECT ip, port, protocol, score, response_time_ms, last_check, created_at FROM proxies WHERE ip = ? AND port = ?",
(ip, port),
) as cursor:
row = await cursor.fetchone()
if row:
return _row_to_proxy(row)
return None
@staticmethod
async def get_random(db: aiosqlite.Connection) -> Optional[Proxy]:
async with db.execute(
"SELECT ip, port, protocol, score, response_time_ms, last_check, created_at FROM proxies WHERE score > 0 ORDER BY RANDOM() LIMIT 1"
) as cursor:
row = await cursor.fetchone()
if row:
return _row_to_proxy(row)
return None
@staticmethod
async def list_all(
db: aiosqlite.Connection,
protocol: Optional[str] = None,
limit: int = 100000,
offset: int = 0,
) -> List[Proxy]:
query = "SELECT ip, port, protocol, score, response_time_ms, last_check, created_at FROM proxies"
params: List = []
if protocol:
query += " WHERE protocol = ?"
params.append(protocol.lower())
query += " LIMIT ? OFFSET ?"
params.extend([limit, offset])
async with db.execute(query, params) as cursor:
rows = await cursor.fetchall()
return [_row_to_proxy(row) for row in rows]
@staticmethod
async def iter_batches(
db: aiosqlite.Connection,
protocol: Optional[str] = None,
batch_size: int = 1000,
):
"""流式分批读取代理,避免一次性加载大量数据到内存"""
offset = 0
while True:
batch = await ProxyRepository.list_all(db, protocol, batch_size, offset)
if not batch:
break
yield batch
offset += batch_size
@staticmethod
async def list_paginated(
db: aiosqlite.Connection,
page: int = 1,
page_size: int = 20,
protocol: Optional[str] = None,
min_score: int = 0,
max_score: Optional[int] = None,
sort_by: str = "last_check",
sort_order: str = "DESC",
) -> Tuple[List[Proxy], int]:
conditions = ["score >= ?"]
params: List = [min_score]
if protocol:
conditions.append("protocol = ?")
params.append(protocol)
if max_score is not None:
conditions.append("score <= ?")
params.append(max_score)
where_clause = " AND ".join(conditions)
allowed_sort_by = {"ip", "port", "protocol", "score", "last_check"}
allowed_sort_order = {"ASC", "DESC"}
if sort_by not in allowed_sort_by or sort_order.upper() not in allowed_sort_order:
order_clause = "last_check DESC"
else:
order_clause = f"{sort_by} {sort_order.upper()}"
offset = (page - 1) * page_size
count_query = f"SELECT COUNT(*) FROM proxies WHERE {where_clause}"
async with db.execute(count_query, list(params)) as cursor:
row = await cursor.fetchone()
total = row[0] if row else 0
data_query = f"""
SELECT ip, port, protocol, score, response_time_ms, last_check, created_at
FROM proxies
WHERE {where_clause}
ORDER BY {order_clause}
LIMIT ? OFFSET ?
"""
params.extend([page_size, offset])
async with db.execute(data_query, params) as cursor:
rows = await cursor.fetchall()
proxies = [_row_to_proxy(row) for row in rows]
return proxies, total
@staticmethod
async def get_stats(db: aiosqlite.Connection) -> dict:
query = """
SELECT
COUNT(*) as total,
COUNT(CASE WHEN score > 0 THEN 1 END) as available,
AVG(score) as avg_score,
COUNT(CASE WHEN protocol = 'http' THEN 1 END) as http_count,
COUNT(CASE WHEN protocol = 'https' THEN 1 END) as https_count,
COUNT(CASE WHEN protocol = 'socks4' THEN 1 END) as socks4_count,
COUNT(CASE WHEN protocol = 'socks5' THEN 1 END) as socks5_count
FROM proxies
"""
async with db.execute(query) as cursor:
row = await cursor.fetchone()
if row:
return {
"total": row[0] or 0,
"available": row[1] or 0,
"avg_score": round(row[2], 2) if row[2] else 0,
"http_count": row[3] or 0,
"https_count": row[4] or 0,
"socks4_count": row[5] or 0,
"socks5_count": row[6] or 0,
}
return {
"total": 0,
"available": 0,
"avg_score": 0,
"http_count": 0,
"https_count": 0,
"socks4_count": 0,
"socks5_count": 0,
}
@staticmethod
async def get_today_new_count(db: aiosqlite.Connection) -> int:
try:
async with db.execute(
"SELECT COUNT(*) FROM proxies WHERE DATE(created_at) = DATE('now', 'localtime')"
) as cursor:
row = await cursor.fetchone()
return row[0] if row else 0
except Exception as e:
logger.error(f"get_today_new_count failed: {e}", exc_info=True)
return 0
@staticmethod
async def clean_invalid(db: aiosqlite.Connection) -> int:
await db.execute("DELETE FROM proxies WHERE score <= 0")
await db.commit()
return db.total_changes
@staticmethod
async def clean_expired(db: aiosqlite.Connection, days: int) -> int:
try:
await db.execute(
"DELETE FROM proxies WHERE last_check < datetime('now', '-' || ? || ' days')",
(days,),
)
await db.commit()
return db.total_changes
except Exception as e:
logger.error(f"clean_expired failed: {e}", exc_info=True)
return 0